PDA

View Full Version : Sample PshopTest Results



kaye
01-14-2005, 09:46 AM
Last two tests run according to 1.0b4 docs. All tests calculated according to 1.0b4 PshopTestScore doc. I may have to edit this if I find any errors. It looks daunting but after a couple of tests, I got the hang of it and is straightforward. I have some more advanced tests that Boots wants me to run.

Ran Pshop 7.0.1 for TEST1 and TEST2, and Pshop CS demo 8.0 (8.0x119) for TEST4, TEST5, and TEST6. Installing the demo is not recommended unless you plan to purchase CS or the CS upgrade from 7.0.

Note: the only differences between these tests:
1. In Step 10, Relative Colorimetric (default 7.0.1), Perceptual (default CS 8.0).
2. TEST1, TEST2, and TEST4 with single DiamondMax10-16MB 300GB PATA drive with Apple driver, capacity 279.35GB, available 279.29GB (empty).
3. TEST5 with 2x DiamondMax10-16MB 300GB PATA drives with Apple RAID, single volume, available 558GB (empty).
4. TEST6 with 2x DiamondMax10-16MB 300GB PATA drives with SoftRAID 3.1.1 RAID, two volumes, first volume capacity 100GB, available 99.96GB, with 64KB stripe unit size (empty). Second volume about 458GB and not used for the Pshop tests.
5. All tests, DMax drives on Acard 6885M, separate channels for each drive where RAID used.

All tests GigaMDD

TEST1(7.0.1)**TEST2(7.0.1)**TEST4(CS 8.0 demo)**TEST5(CS 8.0 demo)**TEST6(CS 8.0 demo)
Activity Monitor Update Frequency all tests 2s
Cache levels all tests 4
Available RAM TEST1-1879MB (7.0.1), TEST2-1881MB (7.0.1), TEST4 thru 6-1879MB (CS)
Max used by Pshop TEST1-50%=939MB, TEST2-70%=1317MB, TEST4 thru 6-70%=1315MB (CS)
Test file size 512MB - 2183.5 (for all tests)

1. Open=14.3s**14.5s**8.4s**8.5s**8.5s
2. Rotate Canvas 180=46.9s**14.2s**4.6s**4.6s**4.6s
3. Apply Image=14.0s/29.2s**7.3s/8.4s**6.8s/7.7s**5.5s/6.5s**5.4s/6.3s
4. Save=4.7s**5.0s**4.6s**4.4s**4.2s
5. Canvas Size=10.3s**2.6s**2.1s**2.1s**2.1s
6. Auto Color=3.0s/5.7s**1.8s/2.7s**1.9s/2.7s**2.0s/2.8s**1.9s/2.8s
7. Gaussian Blur=9.3s/17.2s**6.2s/7.1s**6.8s/7.6s**6.8s/7.6s**6.8s/7.6s
8. Sharpen=10.5s/26.7s**3.3s/4.2s**3.1s/3.9s**3.1s/3.9s**3.1s/3.9s
9. Save=5.8s**5.8s**5.3s**5.1s**4.4s
10.Convert to Profile=20.0s/53.1s**15.8s/16.0s**15.1s/15.1s**15.1s/15.2s**15.1s/15.1s
11.Save=11.4s**11.6s**10.7s**9.9s**9.2s
12.Convert to RGB=27.1s/36.0s**21.6s/21.6s**20.7s/20.7s**20.9s/20.9s**20.9s/20.9s
13.Save=9.3s**9.3s**9.0s**8.5s**7.7s

LineA. Operations Total
Pre-draw=TEST1-83.9s**TEST2-56.0s**TEST4-54.4s**TEST5-53.4s**TEST6-53.2s

Re-draw Total (not average)=TEST1-84.0s**TEST2-4.0s**TEST4-3.3s**TEST5-3.5s**TEST6-3.4s

Operations Total=TEST1-225.1s**TEST2-76.8s**TEST4-64.4s**TEST5-63.6s**TEST6-63.3s

LineB. Save Total=TEST1-31.2s**TEST2-31.7s**TEST4-29.6s**TEST5-27.9s**TEST6-25.5s

LineC. Open and Save Total=TEST1-45.5s**TEST2-46.2s**TEST4-38.0s**TEST5-36.4s**TEST6-34.0s

LineD. Test Total=TEST1-270.6s**TEST2-123.0s**TEST4-102.4s**TEST5-100.0s**TEST6-97.3s

GigaMDD 1.42GHz DP dual boot, 10.3.7, GF4ti, 2.0GB RAM. Pshop 7.0.1 (TEST1, TEST2), Pshop CS 8.0 demo (TEST4, TEST5, TEST6).

TEST1-Pshop 7.0.1 50% memory assigned. Boot and Pshop drive on 120GB 180GXP-8MB (dual boot volume) ATA/100 bus, capacity 115.02GB, available 70.37GB. File and scratch volume: DiamondMax10-16MB 300GB PATA drive on Acard 6885M (only drive on the 4-channel card), capacity 279.35GB, available 279.29GB (empty).

TEST2-Same as above except 70% memory assigned.

TEST4 thru TEST6, Pshop 8.0 CS demo. Boot and Pshop drive on 120GB 180GXP-8MB (dual boot volume) ATA/100 bus, capacity 115.02GB, available 70.04GB (installation of CS). File and scratch volume: explained at the top of this post.

Observations:
LineD is the bottom line total for results. The biggest difference for me was upping the Pshop RAM from 50% to 70%. CS also yields a large difference. Further honing at 70% RAM in CS, as you can see, not a lot of difference for these tests in striping the 2 drives with Apple RAID, just too large a volume and unable to partition. SoftRAID does allow partitioning and showed a better total decrease and could be important if larger file sizes are used. One single fast file and scratch drive and CS however are sufficient for my use of Pshop. k

A.D.
01-15-2005, 02:10 PM
First off, Boots thanks for all the effort in creating a very thorough and useful test. I ran it twice to compare using a FW drive to hold a work file vs using an internal Raptor 36GB to hold a work file.
Some specs: G5 2.5, OS 1.3.7, stock video card, 3GB RAM (memory allocation 100%), scratch is RAID 0x4 (Raptor36GB on Sonnet 4x4PCI card). DW showed all drives at 1% files out of order or less before running the test.
Results of 512MB test file:
Operations total(FW) 32.9
Operations total(Raptor) 33
Open+Save total(FW) 67.4
Open+Save total(Raptor) 32.8
TOTAL SCORE(FW) 100.3
TOTAL SCORE(Raptor) 65.8

Predraw scores were identical in both setups so the score difference reflects the slower open and saves on the FW drive. Compared with Kaye's results the G5 was at least twice as fast on all steps except for opens and saves, where the Raptor was within a second or less either way and the FW drive was considerably slower. I'm pleased with the results and later will try a memory allocation of 70% and see if that alters anything.

Boots
01-16-2005, 06:31 AM
Nice A.D.
Thank you for using the test.

If you can, do try the 192MB test file size at 70% and 100%.

My tests on G5 show at smaller file sizes, 100% allocation slows things down.
It seems to be that above certain file sizes 100% is faster-
But don't assume with CS that 100% allocation is fastest all across the board.

I'm planning more tests on this to get to the bottom of this particularity- I'm not yet sure of what's up with that..

daveseeley
01-16-2005, 02:15 PM
Rumors are rumors....but I heard that Adobe is likely to release CS2 in the next few months. I'm waiting for the move from 7.0

D

A.D.
01-18-2005, 07:28 AM
Boots,
I ran Pshop test with the 192MB file at both 70% (1.261MB) and 100% (1.803MB) RAM. Any speed differences favored the 70% allocation but the overall test scores were only 1 second apart. Auto color, Blur, and Convert were anywhere from .1s to .4s faster at 70%. It is interesting that allocating less memory would be measurably faster and if you get an answer out of Adobe I'd love to hear it. But given that both tests came in just over 2x as fast as the 512MB test, I'm content to stick with 100% and actually get some work done ;)
192MB results:
Operations total (70%)=14
Operations total (100%)=14.9
TOTAL SCORE (70%)=28
TOTAL SCORE (100%)=29

Boots
01-18-2005, 07:41 AM
A.D. Thanks for running those tests...Confirms my results.

You might also want to check out this thread also for an interesting FYI:
http://www.macgurus.com/forums/showthread.php?p=79625#post79625

kaye
01-18-2005, 11:48 AM
Boots sez:
If you're game for more testing Kaye, I'd recommend the following with Photoshop CS only:

I assume you have processor setting in energy saver at 'highest'-

1. Using the 192MB test file size, and your fastest hardware setup-
a. test with a 70% memory allocation (Advanced TEST1a = AT1a)
b. then again with a 100% memory allocation (Advanced TEST1b = AT1b)

Ran Pshop CS demo 8.0 (8.0x119) for AT1a and AT1b.

1. Both tests with 2x DiamondMax10-16MB 300GB PATA drives with SoftRAID 3.1.2 RAID, two volumes, first volume capacity 100GB, available 99.96GB, with 64KB stripe unit size (empty). Second volume about 458GB and not used for the Pshop tests.
2. Both tests, DMax drives on Acard 6885M, separate channels for each drive.

Both tests GigaMDD

AT1a(CS 8.0 demo)**AT1b(CS 8.0 demo)
Activity Monitor Update Frequency 2s
Cache levels 4
Available RAM AT1a-1878MB, AT1b-1879MB
Max used by Pshop AT1a-70%=1314MB, AT1b-100%=1879MB
Test file size 192MB

1. Open=4.3s**4.3s
2. Rotate Canvas 180=1.8s**1.7s
3. Apply Image=2.4s/2.4s**1.8s/2.4s
4. Save=2.5s**2.5s
5. Canvas Size=1.6s**1.6s
6. Auto Color=0.9s/1.3s**1.2s/1.8s
7. Gaussian Blur=2.5s/2.9s**3.4s/5.0s
8. Sharpen=1.3s/1.6s**4.2s/6.1s
9. Save=1.9s**2.8s
10.Convert to Profile=6.9s/6.9s**18.0s/18.0s
11.Save=3.0s**4.9s
12.Convert to RGB=8.4s/8.4s**14.0s/14.0s
13.Save=2.5s**3.7s

LineA. Operations Total
Pre-draw=AT1a-22.4s**AT1b-43.0s

Re-draw total (not average)=AT1a-1.1s**AT1b-4.7s

LineA. Operations Total=AT1a-26.9s**AT1b-51.0s

LineB. Save Total=AT1a-9.9s**AT1b-13.9s

LineC. Open and Save Total=AT1a-14.2s**AT1b-18.2s

LineD. Test Total=AT1a-41.1s**AT1b-69.2s

GigaMDD 1.42GHz DP dual boot, 10.3.7, GF4ti, 2.0GB RAM, Pshop CS 8.0 demo.

Tests AT1a andAT1b, Pshop 8.0 CS demo. Boot and Pshop drive on 120GB 180GXP-8MB (dual boot volume) ATA/100 bus, capacity 115.02GB, available 70.04GB. File and scratch volume: explained above.

Observations:
Limitation of 2GB of real RAM in GigaMDD at 100% memory assignment in Pshop CS really hurts. Up thru Test 5 the results are about the same. Beginning with Test 6 (Auto Color) the results start going downhill and rapidly get worse thru the last step, Step 13. I think I will leave Advanced Tests 2 to those with more real RAM. At least you get the idea of what a G4 can and cannot do. k

Boots
01-18-2005, 02:53 PM
This is awesome Kaye, thank you.

These results are starting to give us a real good idea about what to do for allocation in PS memory prefs with particular RAM conditions.

A.D.
01-18-2005, 04:50 PM
Re: Kaye's test numbers I can understand the significant dropoff with only 2GB RAM installed.The OS (and whatever else) are being starved for memory, far offsetting any gain in devoting more RAM to PS. With 3GB of memory on my G5, the dropoff was only 1 second in total, presumably because of the extra memory available to the rest of the system. So it would appear that the 70% rule is must for 2GB setups but virtually a non-issue when 3GB+ are installed.

Boots
01-19-2005, 04:35 AM
Yes, and it's this memory starvation- when does it begin, what's the relationship to rprvt in top or Activity Monitor....Because I think we are led to think X needs only 250-300MB to run and yes it does if you want to be frugal, but it is significant how much more she really wants to really run good. I want to see if the 70%/100%-bigfile/littlefile anomoly is eliminated with more than 4GB RAM in machine-

But I don't think it will be. I think Photoshop is so strong in it's memory management, how much it can control the OS, kinda brilliant cause it seems like it has a wide tolerance.

I have a feeling- and I say this with all due respect; that perhaps Adobe has not really tested this to the point where they can say 'this is the exact what's up with memory & Photoshop'. If you read what they say in this (http://www.adobe.com/support/techdocs/322829.html) and this (http://www.adobe.com/support/techdocs/317280.html), it seems vague and a bit contradictory, and out of date...
I venture to say, we are working it out for them a bit on the Mac. I'd love to re-write portions of those docs for them!

Chris Billington
01-19-2005, 08:56 AM
Yes, and it's this memory starvation- when does it begin, what's the relationship to rprvt in top or Activity Monitor....Because I think we are led to think X needs only 250-300MB to run and yes it does if you want to be frugal, but it is significant how much more she really wants to really run good. I want to see if the 70%/100%-bigfile/littlefile anomoly is eliminated with more than 4GB RAM in machine-

But I don't think it will be. I think Photoshop is so strong in it's memory management, how much it can control the OS, kinda brilliant cause it seems like it has a wide tolerance.



Just a quick question. I know what Activity Monitor is, but what does 'rprvt in top' refer too? Sorry if this is really dumb!

CB

Boots
01-19-2005, 10:11 AM
Chris, it's the amount of actual RAM an app is using just for itself and not any shared components (Resident private address space size), under the column "RPRVT" in 'top' output in a Terminal window: open Terminal, type 'top' without the quotes and hit enter, then the memory stats for all running apps are displayed.
Part of how much memory is being used at any one time..

Chris Billington
01-19-2005, 10:17 AM
Always usefull!!

CB :dance: :dance: :dance:

kaye
01-20-2005, 06:54 PM
Tested my QS/PF with 1.5GB RAM in Pshop 7.0.1. Not to compare with the GigaMDD but to test various Memory Allocations. This computer has everything but the kitchen sink and no CS demo. PCI slots full. No empty scratch disks. All tests the same except for memory assigned to Pshop.

Test file 512MB - 2183.5

50% test, 1394MB, 697MB assigned
60% test, 1393MB, 837MB assigned
65% test, 1395MB, 907MB assigned
70% test, 1394MB, 975MB assigned

Memory Assigned 50%test**60%test**65%test**70%test
LineD Test Total**609.4s**476.3s**358.6s**256.4s

70% allocation works very well with max RAM of 1.5GB. k

Nicolas
01-24-2005, 01:00 AM
Hello Kaye,

great info!

Regards

Nicolas

kaye
05-07-2005, 07:46 AM
Yesterday, I ran the Pshop Test version 1.0 on my GigaMDD, 2GB RAM, this time with a UL3D-033 and Burly with 4x 18GB X15.3 split into 2 channels, SoftRAID 3.1.3 striped.

Three test results, tho I actually did more, 10.3.9 and CS, 10.3.9 and CS2, 10.4 and CS2. More on those tests later when I figure out how to put it together in detail. NOTE: the Pshop Test docs clearly state that System Requirements are Mac OS 10.3.x. Not yet designed or modified for 10.4.

Installation of CS2, there was good and bad. First, it seems that CS2 does not like installation 2 times even on the same computer but different drives. I installed CS2 on my main 10.3.9 boot drive and on my test 10.4 boot drive while booted from the 10.3.9 drive. When booted in 10.4, CS2 would not launch, gave me two warning messages about some items missing and quit. So I had to redo the installation in 10.4. And the security now of preventing multiple installations is a pain. I will have to do a Transfer Activation of CS2 with Adobe to put it on a G5, rendering CS2 on my GigaMDD inoperative. Then I gotta figure out how to uninstall CS2 on the MDD. The good news is that in 10.3.9 it did not wipe my previous CS installation.

Couple of notes about CS2. Step 10 Convert to Profile is no longer Image menu>Mode>Convert to Profile. It is now Edit>Convert to Profile. Also in CS2, the same Step 10, the default is Relative Colorimetric (like Pshop 7) but I changed it to Perceptual to match CS results.

Last, in CS2, default cache levels is 6 whereas it was 4 in CS. So I did a couple of full tests in CS2 with 6 and 4 just to compare and there was no difference. So I left it at 6 for subsequent tests. Ultimately I will have CS on the GigaMDD and CS2 on a G5. More later. k

kaye
05-07-2005, 11:08 AM
Ran Pshop CS for TEST1 in 10.3.9, CS2 for TEST2 in 10.3.9, and CS2 for TEST3 in 10.4.

Note: the only other differences between these tests:
1. In Step 10, Perceptual (default CS), Relative Colorimetric (default CS2) changed to Perceptual.
2. Cache levels 4 (default CS), 6 (default CS2).

All tests GigaMDD, 2GB RAM, GF4ti, UL3D-033, Burly 4x 18GB X15.3 split into 2 channels, SoftRAID 3.1.3 striped, 64KB stripe units, Workstation.

TEST1(CS-10.3.9)**TEST2(CS2-10.3.9)**TEST3(CS2-10.4)
Activity Monitor Update Frequency all tests 2s
Cache levels 4 (CS), 6 (CS2)
Available RAM TEST1=1860MB (CS), TEST2=1963MB (CS2), TEST3=1963MB (CS2)
Max used by Pshop TEST1-70%=1306MB (CS), TEST2-70%=1374MB (CS2), TEST3-70%=1374MB (CS2)
Test file size 512MB - 2183.5 (all tests)

1. Open=8.8s**10.8s**11.1s
2. Rotate Canvas 180=4.0s**10.0s**10.4s
3. Apply Image=4.5s/5.5s**5.9s/8.1s**5.9s/7.1s
4. Save=4.3s**4.5s**4.4s
5. Canvas Size=2.1s**2.6s**2.5s
6. Auto Color=1.9s/2.8s**1.8s/2.8s**1.8s/2.8s
7. Gaussian Blur=7.0s/7.8s**6.5s/7.6s**6.3s/7.3s
8. Sharpen=3.1s/4.0s**3.0s/4.1s**3.1s/4.1s (See NOTE below)
9. Save=4.3s**4.4s**na
10.Convert to Profile=15.1s/15.2s**14.9s/16.3s**na
11.Save=8.2s**8.1s**na
12.Convert to RGB=20.8s/20.8s**20.7s/20.7s**na
13.Save=6.9s**6.3s**na

LineA. Operations Total
Pre-draw=TEST1-52.4s**TEST2-52.8s

Operations Total=TEST1-62.2s**TEST2-72.2s

LineB. Save Total=TEST1-23.7s**TEST2-23.3s

LineC. Open and Save Total=TEST1-32.5s**TEST2-34.1s

LineD. Test Total=TEST1-94.7s**TEST2-106.3s

Pshop CS2 certainly seems optimized for a G5 with lots more RAM capability.

NOTE: Sharpen in 10.4 hung at the completion of this test showing 99+% CPU usage in Activity Monitor tho no drive activity and confirmed no activity by Burly LEDs and Granite external LED indicator. Unable to complete a successful Forced Quit and had to hard shutdown. Not brave enough to try that again. k

Boots
05-07-2005, 03:50 PM
Interesting numbers Kaye. Thank you for testing and posting the results.

The new activation requirement certainly inspires careful planning, but at least Adobe still allows installation on two computers..

Is the open & save location for the test file on the striped volume or the startup disk?
The designated scratch disk I take it is the 4x15.3k?

Thank you for the observations on the change to "Edit>Convert to Profile". I undertsand also from other informants the Print with Preview dialog has changed- as well as the default memory allocation to 70% (up from 50%).


NOTE: the Pshop Test docs clearly state that System Requirements are Mac OS 10.3.x. Not yet designed or modified for 10.4.

I plan to purchase CS2 in the near future, and then I will update PshopTest accordingly once I've investigated.

-----

My guess is you might be better served dialing back the memory allocation a bit in CS2 for the MDD, since it shows slower results at 70% than CS1.

The results for Step #2 Rotate Canvas really stick out as a possible difference in the way CS2 implements that operation...or- could be system memory starvation. Another CS2 PshopTest tester on a G5 Dual 2GHz with 4GB installed RAM got results very similar to CS1 for that step- around 3.5 sec.

kaye
05-07-2005, 05:47 PM
Is the open & save location for the test file on the striped volume or the startup disk? The designated scratch disk I take it is the 4x15.3k?
Open & save location for the test file is on the striped volume, 4x15.3k, which is the designated scratch disk. I didn't even look for the Print with Preview dialog but my recollection of the default memory allocation is 70%, tho I'm not positive.


My guess is you might be better served dialing back the memory allocation a bit in CS2 for the MDD, since it shows slower results at 70% than CS1.

The results for Step #2 Rotate Canvas really stick out as a possible difference in the way CS2 implements that operation...or- could be system memory starvation. Another CS2 PshopTest tester on a G5 Dual 2GHz with 4GB installed RAM got results very similar to CS1 for that step- around 3.5 sec.
It looked to me that dialing back the memory allocation a bit would work better for my MDD and you confirm that. I didn't test that because I decided that CS2 would go on a G5 and CS would be for the MDD. I could not come to any conclusion about my poor result for Step #2 in CS2 but you are undoubtedly right. Steps 10 thru 13, CS2 does better than CS for me. k

kaye
05-08-2005, 11:42 AM
Worked on finding the sweet spot with Pshop CS2 on my GigaMDD. All Memory assignments below 70% (65-69%) were slower. So I tried some above (71-73%). I won't bore you with all of the details but there is a sweet spot for this MDD with 2GB RAM, 10.3.9. Some representative results with best CS2 and CS boldfaced:

LineD. Test Total=TEST1/CS2-66%-117.0s**TEST2/CS2-67%-112.8s**TEST3/CS2-68%-109.8s

LineD. Test Total=TEST4/CS2-70%-106.3s**TEST5/CS2-71%-105.0s**TEST6/CS2-72%-102.3s**TEST7/CS2-73%-107.2s

Still not faster than my CS score of 94.7s k

Boots
05-08-2005, 12:00 PM
Very interesting Kaye.
So it would appear CS2 is really more optimized for G5- at least in comparison to a GigaMDD Dual...


The results for Step #2 Rotate Canvas really stick out as a possible difference in the way CS2 implements that operation..

Were your results for step #2 still about 10 seconds on those "sweet spot" tests?

kaye
05-08-2005, 12:45 PM
Boots,

The one and only test where Step #2 in CS2 was below about 10s was this one TEST6/CS2-72%-102.3s. It was 7.9s. k

kaye
05-18-2005, 03:48 PM
Ran several Pshop Test CS2 tests on my G5-2.7DP, set High Performance, 10.4.1, 6GB RAM, Radeon X800 XT ME. Not yet trying for fastest speed, just trying my boot and backup boot internal mobo drives. So there will be a progression as I migrate to a SCSI RAID for scratch. All tests in CS2 are default, including Step 10 Relative Colorimetric, except for RAM assignments as I hunt for an approximate sweet spot for each setup. Test file size 512MB - 2183.5 (all tests).

TEST1-Primary boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, Users, File Save, Scratch Drive. RAM assigned 70%=2150MB, available 3072MB.
TEST2-Only difference RAM assigned 100%=3072MB, available 3072MB.
TEST3-RAM assigned 80%=2457MB, available 3072MB.
TEST4-RAM assigned 75%=2304MB, available 3072MB.

With this setup, the sweet spot appears to be between 70-75% RAM assigned.

TEST5-RAM assigned 70%=2150MB, available 3072MB. Differences from above are File Save and Scratch Drive my backup boot drive, a 250GB WDC WD2500JD-41HBC0-8MB cache drive.

I had planned 4 tests in this config but the TEST5 results indicate that the WD backup boot drive is too slow to help.

Progression of TESTs:
1. Open=9.2s**8.9s**9.2s**9.1s**10.3s
2. Rotate Canvas 180=6.1s**6.1s**6.2s**6.1s **6.0s
3. Apply Image=3.1s/3.6s**4.5s/6.6s**4.8s/5.5s**4.1s/4.7s**2.8s/3.3s
4. Save=4.1s**4.1s**4.2s**4.3s**4.1s
5. Canvas Size=1.1s**4.2s**1.1s**1.1s**1.1s
6. Auto Color=0.8s/1.2s**0.8s/1.3s**0.8s/1.2s**0.8s/1.2s**0.8s/1.2s
7. Gaussian Blur=2.8s/3.2s**3.6s/4.2s**2.8s/3.3s**2.8s/3.2s**2.8s/3.3s
8. Sharpen=1.4s/1.9s**1.5s/1.9s**1.4s/1.9s**1.9s/1.9s**1.4s/1.9s
9. Save=3.3s**3.5s**3.4s**3.4s**4.1s
10.Convert to Profile=15.9s/15.9s**15.9s/15.9s**15.9s/15.9s**15.9s/15.9s**15.9s/15.9s
11.Save=7.3s**7.6s**7.5s**7.3s**8.6s
12.Convert to RGB=21.3s/21.3s**24.3s/24.3s**21.5s/21.5s**21.3s/21.3s**21.3s/21.3s
13.Save=6.1s**6.4s**6.0s**6.0s **6.7s

LineA. Operations Total
Pre-draw=45.3s**50.6s**47.2s**46.8s**45.0s

Operations Total=54.3s**64.5s**56.6s**55.4s **54.0s

LineB. Save Total=20.8s**21.6s**21.1s**21.0s **23.5s

LineC. Open and Save Total=30.0s**30.5s**30.3s**30.1s**33.8s

LineD. Test Total=84.3s**95.0s**86.9s**85.5s **87.8s

The sweet spot for the initial config of everything on the boot drive appears to be between 70-75% available RAM and closer to 70%. k

A.D.
05-18-2005, 06:11 PM
Kaye,
Have you activated the bigger tiles plugin? See http://www.adobe.com/support/techdocs/331372.html for info. On my 2.5 dual with 3GB RAM, this plugin shaved significant time on filter and rotate operations. I haven't gone through PStest with CS2 yet, but thought this might help.

kaye
05-18-2005, 11:10 PM
AD,

I wasn't aware of that plugin nor do I understand most of the others mentioned on the Pshop Test. So I am running default. But I will try that one and thanks. k

kaye
05-19-2005, 07:30 AM
Comparing previous TEST1 with Pshop CS2 Bigger Tiles plug-in disabled and TEST6 with it enabled.

Pshop Test CS2 tests on G5-2.7DP, set High Performance, 10.4.1, 6GB RAM, Radeon X800 XT ME. Not yet trying for fastest speed, just trying my boot and backup boot internal mobo drives. All tests in CS2 are default, including Step 10 Relative Colorimetric. Test file size 512MB - 2183.5 (all tests).

TEST1-Primary boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, Users, File Save, Scratch Drive. RAM assigned 70%=2150MB, available 3072MB.
TEST6-Only difference Bigger Tiles plug-in enabled.

TEST1**TEST6:
1. Open=9.2s**7.4s
2. Rotate Canvas 180=6.1s**2.9s
3. Apply Image=3.1s/3.6s**1.3s/1.7s
4. Save=4.1s**4.2s
5. Canvas Size=1.1s**1.0s
6. Auto Color=0.8s/1.2s**0.8s/1.2s
7. Gaussian Blur=2.8s/3.2s**3.6s/3.9s
8. Sharpen=1.4s/1.9s**1.6s/1.9s
9. Save=3.3s**3.4s
10.Convert to Profile=15.9s/15.9s**15.3s/15.3s
11.Save=7.3s**7.5s
12.Convert to RGB=21.3s/21.3s**21.0s/21.0s
13.Save=6.1s**6.1s

LineA. Operations Total
Pre-draw=45.3s**43.6s

Operations Total=54.3s**48.9s

LineB. Save Total=20.8s**21.2s

LineC. Open and Save Total=30.0s**28.6s

LineD. Test Total=84.3s**77.5s

Bigger Tiles definitely helps with the Pshop Test in my configuration. Thanks AD. k

A.D.
05-19-2005, 11:40 AM
Glad to be of help, Kaye.

kaye
06-20-2005, 08:00 AM
Pshop Test CS2 v9.0 tests on G5-2.7DP, set Highest Performance, 10.4.1, 6.5GB RAM, Radeon X800 XT ME. All tests in CS2 are default, Available RAM 3072MB and 70% = 2150MB, plus Bigger Tiles plug-in enabled. Default History States 20, default Cache Levels 6. Test file size 512MB - 2183.5. Boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, and Users. Ethernet, USB printer, and FW disconnected. Spotlight and Dashboard disabled. 2x UL4D, the one on the 100MHz bus installed but not connected. No other PCI/PCI-X card installed.

TEST1-File Save Drive and Scratch Disk Drive 4x 18GB Cheetah X15.3, SoftRAID 3.2, 128KB stripe units, 68.33GB, not journaled, on UL4D 133MHz bus, 2-channels.
TEST2-File Save Drive and Scratch Disk Drive 4x 36GB Atlas 15K II, SoftRAID 3.2, 128KB stripe units, partitioned for 68.97GB, not journaled, on UL4D 133MHz bus, 2-channels.
TEST3-Same as TEST2 except not partitioned, 136.97GB, on UL4D 133MHz bus, 2-channels.

TEST1**TEST2**TEST3:
1. Open=6.4s**6.0s**6.0s
2. Rotate Canvas 180=2.9s**2.9s**2.9s
3. Apply Image=1.3s/1.8s**1.3s/1.8s**1.3s/1.8s
4. Save=3.3s**3.2s**3.2s
5. Canvas Size=1.0s**1.0s**1.0s
6. Auto Color=0.8s/1.2s**0.8s/1.2s**0.8s/1.2s
7. Gaussian Blur=3.6s/3.9s**3.5s/3.8s**3.5s/3.8s
8. Sharpen=1.6s/1.9s**1.5s/1.9s**1.5s/1.9s
9. Save=2.7s**2.8s**2.8s
10.Convert to Profile=15.3s/15.3s**15.4s/15.4s**15.4s/15.4s
11.Save=4.9s**4.6s**4.6s
12.Convert to RGB=20.9s/20.9s**20.9s/20.9s**20.9s/20.9s
13.Save=3.9s**3.7s**3.7s

LineA. Operations Total
Pre-draw=43.5s**43.4s**43.4s

Operations Total=48.9s**48.9s**48.9s

LineB. Save Total=14.8s**14.3s**14.3s

LineC. Open and Save Total=21.2s**20.3s**20.3s

LineD. Test Total=70.1s**69.2s**69.2s

To specifically address the Step 2, 10, and 12 results in Pshop CS2 v9.0, Bigger Tiles plug-in enabled based on a previous test I did:
Step 2-Bigger Tiles more than halves the time, 6.1s to 2.9s.
Step 10-Slight decrease with Bigger Tiles, 15.9s to 15.3s.
Step 12-Slight decrease with Bigger Tiles, 21.3s to 20.9s.

More than half of the Test Totals for both RAIDs is Steps 10 and 12 in Pshop CS2.
X15.3 RAID Test Total 70.1s, Steps 10 and 12 15.3s+20.9s=36.2s
15KII RAID Test Total 69.2s, Steps 10 and 12 15.3s+20.9s=36.2s

One last observation, when I shut down for the last time yesterday with the Atlas RAID running, after the G5 was completely shut down, one of the Burly activity LEDs was still lit and did not go out, tho the Granite external LED indicator showed no LEDs on. Don't know what that was all about. Never seen that before with the Atlas drives. The completed Pshop test file was still on the Atlas RAID. Forgot to trash it. Something to investigate later. For me the Atlas drives do need a minute to warm up before starting the G5. k

Boots
06-21-2005, 04:05 AM
Thank you very much Kaye. :xphat:

I think your results show with a 512MB test file, a pretty small advantage and gain with 4 x Atlas 15k II as opposed to 4 x Cheetah x15.3k. My assumption is the gain would be more substantial if the test file was larger.

I will try to get some info from Adobe about the apparent slower execution of Convert to CMYK and RGB, and Rotate Canvas. Even if it turns out these functions need better optimization, the field reports generally indicate PS CS2 to be faster overall than PS CS. And the new memory management and addressable capability must certainly be a boon to those routinely working with larger file sizes.

My Atlases have arrived. Here's hoping their install into my current Burly to replace 4 x x15.3 will not show that same slight write deficiency...which would mean I've got a problem with the cables or UL4.

I also have 8 GB PC3200 25330 RAM due to arrive; looking forward to testing with larger files. :)

kaye
06-21-2005, 07:30 AM
Boots,

So what typical file sizes do you work with? And what typical max file sizes? I look forward to a larger test file size. Or more Saves with the existing Pshop Test.

Very curious to learn what Adobe sez. I looked thru their website yesterday, tons of complaints about Pshop CS2 but could not find any specifically relating to Steps 10 and 12 slowness. As you said however, overall Pshop CS2 is faster judging by the answers from users and forum Techs.

You have your work cut out for you installing the Atlases and RAM. Good luck. I have to get to the APC UPS installation and go to the gum surgeon today for a followup look see of his handiwork. Hate going. k

Boots
06-21-2005, 08:19 AM
When I was first working out the details of PshopTest, I was including a chart of test file sizes and their respective resolutions for the "Image Size" dialog. I'll send this along to you. For the next iteration of PshopTest, I will probably suggest 4 test file sizes, which will include a large-format (.psb) test file size. This should allow for definitive tests with large amounts of installed RAM, and more clearly show the benefits and differences of our faster scratch disks, and useful comparisons between SATA and SCSI scratch disks.

My own personal work varies widely in file size. I sometimes scan transparencies which initially result in files over 1GB. Some typical art projects can balloon to 3 or 400MB before flattening. Then there is a bunch of photo work which ususally stays under 100MB. I'm thinking that in some ways, the standard PshopTest 512MB test file actually approximates working at much smaller sizes since their are no additional layers in the test file, and .psd or .tif files with layers take substantially longer to save than flattened files.

Re steps #10 & 12: The first order of business for me is establishing a contact within Adobe who is willing to answer these and other more technical questions; I'm working on that one...

Hope all goes well at the doctor's. :)

kaye
06-21-2005, 09:28 AM
Boots,

Got the chart. Sweet. I chose the biggest blast, 1.50GB-3780 Resolution. Got the test file size created and will rerun the X15.3 and 15KII test comparison with Pshop CS2. Take a couple of days probably depending on what the knife surgeon does to me today. Plus I have a lot to do tomorrow.

One note in CS2, Image Size, there is an additional checkbox besides Constrain Proportions and Resample Image (Bicubic is the resampling method). New checkbox is Scale Styles. All three checkboxes are checked by default. I did not uncheck Scale Styles but have no idea what that does, not that it matters. I just keep track of the seconds. k

kaye
06-23-2005, 07:20 AM
Except for test file size, these tests were run the same way. Reminds me of my first Pshop test with the GigaMDD, everything working very hard.

Pshop Test CS2 v9.0 tests on G5-2.7DP, set Highest Performance, 10.4.1, 6.5GB RAM, Radeon X800 XT ME. All tests in CS2 are default, Available RAM 3072MB and 70% = 2150MB, plus Bigger Tiles plug-in enabled. Default History States 20, default Cache Levels 6. TEST FILE SIZE 1.5GB - 3780. Boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, and Users. Ethernet, USB printer, and FW disconnected. Spotlight and Dashboard disabled. 2x UL4D, the one on the 100MHz bus installed but not connected. No other PCI/PCI-X card installed.

TEST4-File Save Drive and Scratch Disk Drive 4x 18GB Cheetah X15.3, SoftRAID 3.2, 128KB stripe units, 68.33GB, not journaled, on UL4D 133MHz bus, 2-channels.
TEST5-File Save Drive and Scratch Disk Drive 4x 36GB Atlas 15K II, SoftRAID 3.2, 128KB stripe units, partitioned for 68.97GB, not journaled, on UL4D 133MHz bus, 2-channels.
TEST6-Same as TEST5 except not partitioned, 136.97GB, on UL4D 133MHz bus, 2-channels.

TEST4**TEST5**TEST6:
1. Open=17.6s**16.8s**16.9s
2. Rotate Canvas 180=8.7s**10.3s**8.5s
3. Apply Image=6.5s/12.5s**6.5s/12.6s**6.5s/12.6s
4. Save=7.1s**6.8s**6.9s
5. Canvas Size=7.5s**7.6s**7.6s
6. Auto Color=5.3s/11.3s**5.3s/10.1s**5.3s/10.1s
7. Gaussian Blur=12.3s/18.8s**14.0s/20.0s**13.7s/19.2s
8. Sharpen=7.6s/12.5s**9.3s/14.7s**7.6s/12.4s
9. Save=7.0s**6.6s**6.5s
10.Convert to Profile=43.6s/51.6s**43.6s/49.7s**44.1s/50.4s
11.Save=17.9s**16.3s**16.2s
12.Convert to RGB=64.4s/66.7s**64.4s/66.8s**64.4s/66.6s
13.Save=12.3s**11.3s**11.6s

LineA. Operations Total
Pre-draw=139.7s**143.1s**141.6s

Operations Total=189.6s**191.8s**187.4s

LineB. Save Total=44.3s**41.0s**41.2s

LineC. Open and Save Total=61.9s**57.8s**58.1s

LineD. Test Total=251.5s**249.6s**245.5s

NOTE that for every step, the Burly LEDs were operating. RAID doesn't start immediately in all steps but obviously when real RAM is full, the drives do the work. More RAM would help. Temp file size reaches an approximate max of 18.19GB. Not seen the RAID operate for other than open and save before with the 512MB-2183.5 file.

Once again when I shut down for the last time today with the Atlas RAID running, after the G5 was completely shut down, one of the Burly activity LEDs was still lit and did not go out, tho the Granite external LED indicator showed no LEDs on. The completed Pshop test file was trashed on the Atlas RAID this time. However the previous file left on the Atlas RAID was still good tho big and trashed before installing a new file. k

kaye
07-28-2005, 04:47 PM
This Pshop Test CS2 v9.0 tests on G5-2.7DP, set Highest Performance, 10.4.2, 6.5GB RAM, Radeon X800 XT ME. Both tests in CS2 are default, Available RAM 3072MB and 70% = 2150MB, plus Bigger Tiles plug-in enabled. Default History States 20, default Cache Levels 6. TEST FILE SIZE 1.5GB - 3780. Boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, and Users. Ethernet, USB printer, and FW disconnected. Spotlight and Dashboard disabled. 2x UL4D. No other PCI/PCI-X card installed.

Comparing my previous fastest test, TEST6 in 10.4.1 with my latest config and 10.4.2.

TEST6-File Save Drive and Scratch Disk Drive 4x 36GB Atlas 15K II, SoftRAID 3.2, 128KB stripe units, 136.97GB, not journaled, on UL4D 133MHz bus, 2-channels, 10.4.1.

TEST7-File Save Drive and Scratch Disk Drive 4x 36GB Atlas 15K II on UL4D 133MHz slot and 4x 18GB Cheetah X15.3 on UL4D 100MHz slot, SoftRAID 3.2, 128KB stripe units, 136.61GB, not journaled, 4-channels, 10.4.2.

TEST6**TEST7:
1. Open=16.9s**16.8s
2. Rotate Canvas 180=8.5s**8.4s
3. Apply Image=6.5s/12.6s**6.8s/13.5s
4. Save=6.9s **6.9s
5. Canvas Size=7.6s**7.5s
6. Auto Color=5.3s/10.1s**5.6s/10.6s
7. Gaussian Blur=13.7s/19.2s**12.2s/18.2s
8. Sharpen=7.6s/12.4s**7.7s/12.7s
9. Save=6.5s**6.5s
10.Convert to Profile=44.1s/50.4s**43.2s/47.1s
11.Save=16.2s**16.0s
12.Convert to RGB=64.4s/66.6s**63.5s/65.9s
13.Save=11.6s**11.0s

LineA. Operations Total
Pre-draw=141.6s**139.0s

Operations Total=187.4s**183.9s

LineB. Save Total=41.2s**40.4s

LineC. Open and Save Total=58.1s**57.2s

LineD. Test Total=245.5s**241.1s

NOTE that for every step, the Burly LEDs were operating. RAID doesn't start immediately in all steps but obviously when real RAM is full, the drives do the work. More RAM would help. Temp file size reaches an approximate max of 18.19GB. Not seen the RAID operate for other than open and save before with the 512MB-2183.5 file.

Previous problem with one of the Atlas RAID Burly LEDs remaining on after the G5 was completely shut down no longer occurs in 10.4.2. k

kaye
08-01-2005, 07:58 AM
This Pshop Test CS2 v9.0 tests on G5-2.7DP, set Highest Performance, 10.4.2, 8.0GB RAM, Radeon X800 XT ME. All tests in CS2 are default, Available RAM 3072MB and 70% = 2150MB, plus Bigger Tiles plug-in enabled. Default History States 20, default Cache Levels 6. TEST FILE SIZE 1.5GB - 3780. Boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, and Users. Ethernet, USB printer, and FW disconnected. Spotlight and Dashboard disabled. 2x UL4D. No other PCI/PCI-X card installed.

Comparing my previous fastest test, TEST6 in 10.4.1, TEST7 in 10.4.2, with my latest config, TEST8 and TEST9, and 10.4.2.

TEST6-File Save Drive and Scratch Disk Drive 4x 36GB Atlas 15K II, SoftRAID 3.2, 128KB stripe units, 136.97GB, not journaled, on UL4D 133MHz bus, 2-channels, 10.4.1, 6.5GB RAM.

TEST7-File Save Drive and Scratch Disk Drive 4x 36GB Atlas 15K II on UL4D 133MHz slot and 4x 18GB Cheetah X15.3 on UL4D 100MHz slot, SoftRAID 3.2, 128KB stripe units, 136.61GB, not journaled, 4-channels, 10.4.2, 6.5GB RAM.

TEST8-Same as TEST7 except 8.0GB RAM.

TEST9-Same as TEST8 except TEST FILE SIZE 512MB - 2183.5.

TEST6**TEST7**TEST8**TEST9:
1. Open=16.9s**16.8s**16.5s**6.0s
2. Rotate Canvas 180=8.5s**8.4s**8.4s**2.8s
3. Apply Image=6.5s/12.6s**6.8s/13.5s**6.9s/13.9s**1.4s/1.9s
4. Save=6.9s **6.9s**7.0s**3.2s
5. Canvas Size=7.6s**7.5s**6.9s**1.0s
6. Auto Color=5.3s/10.1s**5.6s/10.6s**5.4s/10.2s**0.8s/1.2s
7. Gaussian Blur=13.7s/19.2s**12.2s/18.2s**12.0s/17.5s**3.1s/3.4s
8. Sharpen=7.6s/12.4s**7.7s/12.7s**7.7s/12.7s**1.6s/1.9s
9. Save=6.5s**6.5s**6.5s**2.6s
10.Convert to Profile=44.1s/50.4s**43.2s/47.1s**43.2s/47.0s**15.4s/15.4s
11.Save=16.2s**16.0s**15.7s**4.7s
12.Convert to RGB=64.4s/66.6s**63.5s/65.9s**63.3s/65.6s**21.0s/21.0s
13.Save=11.6s**11.0s**10.9s**3.7s

LineA. Operations Total
Pre-draw=141.6s**139.0s**138.5s**43.3s

Operations Total=187.4s**183.9s**182.2s**48.6s

LineB. Save Total=41.2s**40.4s**40.1s**14.2s

LineC. Open and Save Total=58.1s**57.2s**56.6s**20.2s

LineD. Test Total=245.5s**241.1s**238.8s**68.8s

NOTES/Conclusions:
1. Comparing LineD. Test Totals, the addition of 4 more drives to TEST7-9 and 8GB RAM in TEST8-9 from 6.5GB may not be worth the savings in time depending on the typical file sizes you use.
2. TEST9 in particular with the smaller file size showed almost no decrease from my previous TEST2 and TEST3 of 4x Atlas drives and 6.5GB RAM LineD. Test Total of 69.2, a drop of only 0.4s with twice the drives and 1.5GB more RAM.
3. Spotlight and Dashboard disabled does make a difference. Not sure if it is both or just Spotlight but after running TEST8, I ran another test with both enabled just to see if there was any difference and there was, 241.0s or 2.2s slower, negating the benefits of the additional 1.5GB of RAM I added.
4. Another experiment, since I started working with the big TEST FILE SIZE 1.5GB - 3780, initially I was curious about where this file was located, outer tracks or where. Using Drive Genius I took a look and it was not close to the outer tracks. So each time I have optimized the RAID for the big file. The file is never in the same place after copying to the RAID. I can't see a pattern. But if the file is not optimized before running the Pshop test there is a time penalty. For the experiment I ran, the penalty was 1.7s additional time compared to LineD TEST8. With the smaller file, there is no measurable penalty.
5. With CS2, half or nearly so of the LineD. Test Total is Steps10 and 12.
k

TZ
08-05-2005, 01:05 PM
Barefeats took a look at some dual core, dual cpu, Intel systems against a couple of G5s.
http://www.barefeats.com/dualcore.html

Makes you wonder about the wisdom or notion that IBM can't deliver decent enough systems :D
but for now, the G5 holds its own in some tests, slightly lags in CS2 (maybe they should have tried with 16GB of RAM ;)

Boots
08-06-2005, 06:29 AM
Thank you for all the recent PshopTest Kaye.
Sorry I haven't been around...
I've been tied up with moving- actually moved to a larger house just up the street, more room to play. :D

My fall projects include migrating to Tiger, resolving my strange Atlas/Burly/UL4 "lockup on startup" condition, and updating PshopTest.

One thing I will say based on all of the recent testing- Adobe has done a great job at making Photoshop CS2 RAM-centric: with a fast dual G5, Tiger, and lots of RAM, a user can now get peak performance with just 2 hard drives; a really fast 4 or 6 drive scratch array now has less overall impact on performance when enough RAM is installed. :p

kaye
09-23-2005, 10:25 AM
This Pshop Test CS2 v9.0 tests on G5-2.7DP, set Highest Performance, 10.4.2, 8.0GB RAM, Radeon X800 XT ME. All tests in CS2 are default, Available RAM 3072MB and 70% = 2150MB, plus Bigger Tiles plug-in enabled. Default History States 20, default Cache Levels 6. TEST FILE SIZE 1.5GB - 3780. Boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, and Users. Ethernet, USB printer, and FW disconnected. Spotlight and Dashboard disabled. SeriTek/1VE4 (ROM revision 5.1.2.1VE4) in 100MHz slot, 4-channels connected to SATA Burly, 3x Raptor74s, 1x Raptor36. UL4D in 100MHz slot, UL4D in 133MHz slot, nothing connected to either UL4D. Gurus one meter external SATA data cables.

Comparing with my previous fastest test, TEST8.

TEST8-File Save Drive and Scratch Disk Drive 4x 36GB Atlas 15K II on UL4D 133MHz slot and 4x 18GB Cheetah X15.3 on UL4D 100MHz slot, SoftRAID 3.2, 128KB stripe units, 136.61GB, not journaled, 4-channels, 10.4.2, 8.0GB RAM.

TEST10- SeriTek/1VE4, 2x Raptor74, SoftRAID 3.2.1, 128KB stripe units RAID0, 138.43GB, not journaled, 2-channels, 10.4.2, 8.0GB RAM.

TEST11-Same as TEST10 except 3x Raptor74, 207.67GB, and 3-channels.

TEST12-Same as TEST11 except 3x Raptor74, 1x Raptor36, 137.82GB, and 4-channels.

TEST8**TEST10**TEST11**TEST12:
1. Open=16.5s**18.9s**17.4s**16.6s
2. Rotate Canvas 180=8.4s**11.6s**10.0s**9.4s
3. Apply Image=6.9s/13.9s**7.0s/16.2s**6.8s/14.8s**6.8s/14.4s
4. Save=7.0s**7.3s**7.6s**6.8s
5. Canvas Size=6.9s**8.6s**7.8s**7.6s
6. Auto Color=5.4s/10.2s**5.7s/10.8s**6.3s/11.0s**5.5s/10.4s
7. Gaussian Blur=12.0s/17.5s**12.4s/17.2s**12.2s/16.7s**12.5s/18.2s
8. Sharpen=7.7s/12.7s**7.9s/13.4s**7.8s/12.6s**7.8s/12.9s
9. Save=6.5s**7.6s**6.5s**6.4s
10.Convert to Profile=43.2s/47.0s**43.9s/48.7s**45.3s/49.4s**44.6s/48.6s
11.Save=15.7s**18.7s**16.4s**15.7s
12.Convert to RGB=63.3s/65.6s**66.1s/68.6s**63.5s/65.9s**63.9s/65.8s
13.Save=10.9s**13.4s**11.7s**10.8s

LineA. Operations Total
Pre-draw=138.5s**143.0s**141.9s**141.1s

Operations Total=182.2s**195.1s**188.2s**187.3s

LineB. Save Total=40.1s**47.0s**42.2s**39.7s

LineC. Open and Save Total=56.6s**65.9s**59.6s**56.3s

LineD. Test Total=238.8s**261.0s**247.8s**243.6s

NOTES:
1. Love the SeriTek/1VE4, SR3.2.1, the SATA Burly, and the ease of the external cabling. Bulletproof.
2. Cannot optimize with Drive Genius v1.1.5. Sees the drives but will not see the volume for optimizing. Worked previously with SCSI drives and SoftRAID. So I created 2x Raptor74s with AppleRAID and the same 128KB SU and Drive Genius will optimize the volume. Created 1x Raptor74 with SoftRAID and again Drive Genius does not see the volume. Will send a report to SoftRAID. Both of my boot internal SATA drives, Tiger Disk Utility, do also optimize with Drive Genius.
3. TEST12 would of course been faster with a fourth Raptor74 instead of the Raptor36.
k

kaye
09-23-2005, 09:33 PM
Mark James of SoftRAID quickly answered my email about SoftRAID initialized SATA volumes not working with Drive Genius.

I reminded them at MacWorld, but need to again. It is their issue, to recognize the SoftRAID partition map. It should not be able to repartition, or any of those things, but simple "repairs" is no issue for them. I do send them SoftRAID, so it is obvious that the engineers are not getting it.

thanks for the alert. ;-)

The box said 'Requires Windows 95, or better.' So I bought a Macintosh.

Mark James
SoftRAID, LLC
Looks like I need to get on Prosoft Engineering's case. k

Boots
09-25-2005, 06:40 PM
Is Photoshop CS2 faster than Photoshop CS?

You betch your sweet registers it is- and by quite a bit.

To prove this, I took a break from the artwork and ran 9 PshopTests over the last two days; three with the 512MB test file, and six with the 1536MB (1.5GB) test file. The results show a dramatic speed improvement with CS2 over CS- provided substantial installed RAM is available to Photoshop and the OS.

The test machine is an original G5 Dual 2GHz running Mac OS10.3.9 with 2 x 73GB WD740 Raptor internal drives formatted as a single Apple RAID 0 volume of 138.5GB, with 109.7GB available, and 28.8GB used. Installed RAM is 8GB of PC3200U-25330 400MHz DDR SDRAM.
The AGP card is an ATI Radeon 9800 Pro Special Edition 256MB VRAM.
An ATTO UL4D running driver version 3.4.0 is installed in PCI slot #4 (133MHz).
Attached to the card- two drives per channel- in an external Burly enclosure, are four 37GB Atlas 15K II SCSI drives formatted with SoftRAID3.2 as a single RAID 0 (128k stripe-size) volume of 137GB, with136.3GB available, and 649.5MB used.
This volume was used as the designated Work Disk/Scratch Disk in tests #8 and #9.

PshopTest #1: 512MB test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.5GB, 109.7GB available, 28.8GB used.

Photoshop CS with default preferences except Image Cache at 6 and Maximum (memory) Used by Photoshop at 70% (1169MB of 1670MB Available RAM):

Operations Total: 46.5s
Save Total: 18.5s
Open & Save Total: 28.8s
Test Total: 75.3s
Re-draw Average: 0.483s

PshopTest #2: 512MB test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.5GB, 109.7GB available, 28.8GB used.

Photoshop CS2 with default preferences and Bigger Tiles Plug-in enabled and Maximum (memory) Used by Photoshop at 70% (2150MB of 3072MB Available RAM):

Operations Total: 42.4s
Save Total: 18.7s
Open & Save Total: 28.9s
Test Total: 71.3s
Re-draw Average: 0.233s

PshopTest #3: 512MB test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.5GB, 109.7GB available, 28.8GB used.

Photoshop CS with default preferences except Image Cache at 6 and Maximum (memory) Used by Photoshop at 100% (1669MB of 1669MB Available RAM):

Operations Total: 40.8s
Save Total: 18.8s
Open & Save Total: 29.3s
Test Total: 70.1s
Re-draw Average: 0.266s

PshopTest #4: 1536MB test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.5GB, 109.7GB available, 28.8GB used.

Photoshop CS with default preferences except Image Cache at 6 and Maximum (memory) Used by Photoshop at 70% (1168MB of 1669MB Available RAM):

Operations Total: 566.9s
Save Total: 181.3s
Open & Save Total: 218.1s
Test Total: 785.0s
Re-draw Average: 23.76s

PshopTest #5: 1536MB test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.5GB, 109.7GB available, 28.8GB used.

Photoshop CS2 with default preferences and Bigger Tiles Plug-in enabled and Maximum (memory) Used by Photoshop at 70% (2150MB of 3072MB Available RAM):

Operations Total: 213.0s
Save Total: 57.5s
Open & Save Total: 84.8s
Test Total: 297.8s
Re-draw Average: 6.516s

PshopTest #6: 1536MB test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.5GB, 109.7GB available, 28.8GB used.

Photoshop CS with default preferences except Image Cache at 6 and Maximum (memory) Used by Photoshop at 100% (1670MB of 1670MB Available RAM):

Operations Total: 490.7s
Save Total: 96.1s
Open & Save Total: 127.7s
Test Total: 618.4s
Re-draw Average: 21.75s

PshopTest #7: 1536MB test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.5GB, 109.7GB available, 28.8GB used.

Photoshop CS2 with default preferences and Bigger Tiles Plug-in enabled and Maximum (memory) Used by Photoshop at 100% (3072MB of 3072MB Available RAM):

Operations Total: 159.4s
Save Total: 52.2s
Open & Save Total: 78.2s
Test Total: 237.6s
Re-draw Average: 3.416s

PshopTest #8: 1536MB test file size. System and Users on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.5GB, 109.7GB available, 28.8GB used. Test file and designated scratch disk is the 4 x Atlas 15K II striped RAID 0 volume of 137GB, with136.3GB available, and 649.5MB used.

Photoshop CS with default preferences except Image Cache at 6 and Maximum (memory) Used by Photoshop at 100% (1667MB of 1667MB Available RAM):

Operations Total: 293.7s
Save Total: 68.8s
Open & Save Total: 91.6s
Test Total: 385.3s
Re-draw Average: 11.6s

PshopTest #9: 1536MB test file size. System and Users on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.5GB, 109.7GB available, 28.8GB used. Test file and designated scratch disk is the 4 x Atlas 15K II striped RAID 0 volume of 137GB, with136.3GB available, and 649.5MB used.

Photoshop CS2 with default preferences and Bigger Tiles Plug-in enabled and Maximum (memory) Used by Photoshop at 100% (3072MB of 3072MB Available RAM):

Operations Total: 156.9s
Save Total: 52.4s
Open & Save Total: 73.1s
Test Total: 230.0s
Re-draw Average: 3.05s

----------

Some conclusions and thoughts....

1. Photoshop CS2 whips Photoshop CS in every test, and really pulls way ahead in larger test file sizes. Clearly, the much greater RAM access is the key factor.

2. The last two tests show the big advantage of a really fast Work Disk/Scratch Disk for both versions, particularly CS.

3. if you've got plenty of RAM for the OS, max out Photoshop at 100% RAM allocation. Yes, there is a slight penalty in Open and Save times, but an overall speed increase is the result.

kaye
09-26-2005, 05:00 AM
That is really sweet Boots and obviously a lot of work. At some point I will have to retry Pshop CS2 at 100% in Tiger. Crystal clear that 100% is Da Bomb and shows how little I know. k

Boots
09-26-2005, 06:34 AM
Once I get Tiger installed, I'll be testing again. I expect an even better score on the 1.5GB test file because of the Tiger-enabled PSCS2 RAM caching routine....

A.D.
09-26-2005, 06:50 AM
Boots, Thanks for posting the test results. AND for putting to bed the myth that you need Tiger along with CS2 to utilize the extra RAM. Because of the barefeats postings months back there has been lots of misinformation on the web indicating you must upgrade to 10.4 to reap the RAM benefits in CS2. Adobe's engineers have indeed stated it is CS2 that is finally making use of OS caching with unused RAM, but time and again people attribute this benefit to 10.4. Now, off to buy some more RAM...

kaye
09-26-2005, 10:28 AM
Holy cow what a difference with Pshop at 100%.

This Pshop Test CS2 v9.0 tests on G5-2.7DP, set Highest Performance, 10.4.2, 8.0GB RAM, Radeon X800 XT ME. All tests in CS2 are default, Available RAM 3072MB and 100% = 3072MB, plus Bigger Tiles plug-in enabled. Default History States 20, default Cache Levels 6. TEST FILE SIZE 1.5GB - 3780. Boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, and Users. Ethernet, USB printer, and FW disconnected. Spotlight and Dashboard disabled. SeriTek/1VE4 (ROM revision 5.1.2.1VE4) in 100MHz slot, 4-channels connected to SATA Burly, 3x Raptor74s, 1x Raptor36. UL4D in 100MHz slot, UL4D in 133MHz slot, nothing connected to either UL4D. Gurus one meter external SATA data cables.

TEST13- File Save Drive and Scratch Disk Drive 3x Raptor74, 1x Raptor36, 137.82GB, and 4-channels, SoftRAID 3.2.1, 128KB stripe units RAID0, not journaled, 4-channels, 10.4.2, 8.0GB RAM.

TEST13:
1. Open=16.3s
2. Rotate Canvas 180=8.8s
3. Apply Image=5.0s/8.3s
4. Save=7.6s
5. Canvas Size=5.4s
6. Auto Color=3.1s/5.2s
7. Gaussian Blur=10.1s/12.7s
8. Sharpen=5.5s/7.9s
9. Save=6.4s
10.Convert to Profile=40.5s/43.3s
11.Save=13.5s
12.Convert to RGB=60.9s/61.9s
13.Save=10.8s

LineA. Operations Total
Pre-draw=125.1s
Operations Total=153.5s
LineB. Save Total=38.3s
LineC. Open and Save Total=54.6s
LineD. Test Total=208.1s
Re-draw Average=2.37s
k

Boots
09-27-2005, 05:58 AM
AND for putting to bed the myth that you need Tiger along with CS2 to utilize the extra RAM.

Well A.D., I certainly was one of those meekly following along with that edict until I did these tests.... :o

I'll be installing Tiger later today and once I get organized, we can see what difference it makes with the large test file size...
I'll have more or less the same setup; striped Raptor boot and about the same amount of files on disk, so it should b a reasonably close comparison.

And- as one member has suggested- does CS and CS2 run slower under Tiger than Panther? That will be an important question to answer.


Holy cow what a difference with Pshop at 100%.

That is one kick-ass PshopTest score Kaye. I can only imagine what you'd get with the 8x SCSI drives used as a WorkDisk/ScratchDisk
:kickass:

kaye
09-27-2005, 06:22 AM
Boots, the SCSI tests will be awhile. But I will get to them. I did notice from your post that there is a new driver version 3.4.0 for UL4D. So I updated the UL4Ds to that last night. k

kaye
09-28-2005, 10:50 AM
This Pshop Test CS2 v9.0 tests on G5-2.7DP, set Highest Performance, 10.4.2, 8.0GB RAM, Radeon X800 XT ME. All tests in CS2 are default, Available RAM 3072MB and 100% = 3072MB, plus Bigger Tiles plug-in enabled. Default History States 20, default Cache Levels 6. TEST FILE SIZE 1.5GB - 3780. Boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, and Users. Ethernet, USB printer, and FW disconnected. Spotlight and Dashboard disabled.

TEST13- (previous test for comparison) File Save Drive and Scratch Disk Drive 3x Raptor74, 1x Raptor36, 137.82GB, and 4-channels, SoftRAID 3.2.1, 128KB stripe units RAID0, not journaled, 4-channels, 10.4.2, 8.0GB RAM. SeriTek/1VE4 (ROM revision 5.1.2.1VE4) in 100MHz slot. Gurus one meter external SATA data cables. UL4D in 100MHz slot, UL4D in 133MHz slot, nothing connected to either UL4D.

TEST14- All same as above except File Save Drive and Scratch Disk Drive 4x 36GB Atlas 15K II, partitioned 67.97GB, on UL4D 133MHz slot, 2-channels. UL4D and SeriTek/1VE4 in 100MHz slots nothing connected. In the past, partitioning these drives proved to be slower in the big file Pshop Test but I wanted to try again, but still slower.

TEST15- Same as above except not partitioned, 136.97GB.

TEST16- Same as above except Apple RAID, now closer in performance to SoftRAID, 136.93GB.

TEST13**TEST14**TEST15**TEST16:
1. Open=16.3s**16.0s**16.2s**16.2s
2. Rotate Canvas 180=8.8s**8.1s**8.0s**8.1s
3. Apply Image=5.0s/8.3s**6.0s/9.1s**5.1s/8.4s**5.1s/8.5s
4. Save=7.6s**7.0s**7.0s**7.0s
5. Canvas Size=5.4s**6.2s**5.1s**5.1s
6. Auto Color=3.1s/5.2s**3.1s/5.1s**3.1s/5.2s**3.1s/5.3s
7. Gaussian Blur=10.1s/12.7s**10.3s/12.9s**10.0s/12.6s**10.1s/13.6s
8. Sharpen=5.5s/7.9s**5.5s/8.0s**5.5s/8.0s**5.6s/8.0s
9. Save=6.4s**6.5s**6.5s**6.5s
10.Convert to Profile=40.5s/43.3s**40.9s/43.6s**40.7s/43.5s**40.4s/43.2s
11.Save=13.5s**13.6s**13.6s**13.6s
12.Convert to RGB=60.9s/61.9s**61.1s/62.0s**61.2s/62.2s**61.0s/61.9s
13.Save=10.8s**10.9s**11.0s**11.0s

LineA. Operations Total
Pre-draw=125.1s**126.9s**125.6s**125.3s
Operations Total=153.5s**155.0s**153.0s**153.7s
LineB. Save Total=38.3s**38.0s**38.1s**38.1s
LineC. Open and Save Total=54.6s**54.0s**54.3s**54.3s
LineD. Test Total=208.1s**209.0s**207.3s**208.0s

Re-draw Average=2.37s**2.30s**2.37s**2.53s

My 8x SCSI test will have to wait. k

Boots
09-29-2005, 02:36 PM
Does Photoshop run faster or slower under Tiger? Part I

On the surface- using PshopTest in its present form- it would appear it runs slower. However, when the test details are considered, the opposite is probably true.

To test this, I tried to closely match the parameters of my tests 1 through 9 which compared Photoshop CS and Photoshop CS2 above.

I was able to match most things including disk space pretty closely, but I quickly discovered Tiger allows Photoshop CS to 'see' a larger amount of "Available RAM" in Photoshop menu>Preferences>Memory & Image Cache...
Under Panther, on this test machine with 8GB RAM installed, CS 'sees' about 1670MB "Available RAM". Under Tiger, CS 'sees' about 1735MB.

This time, I ran 5 PshopTests; two with the 512MB test file, and three with the 1536MB (1.5GB) test file. The results show both CS and CS2 when running under Tiger are at least as fast if not faster then when they are both running under Panther for all test steps except for #10 and #12, which not only run slower in CS2 than in CS, but are also slower when Photoshop CS and CS2 are used with Tiger instead of Panther.

Why are these two test steps slower? Good question. I need to investigate whether it has something to do with how the colors of the prepared test image are interpreted by CS2; are its colors "seen" by PSCS2 differently than in PSCS? Is this also a system-level ColorSync issue?
It may be- and I'm hoping- something that can be adjusted in the preferences prior to running PshopTest, perhaps bringing those particular test step results inline. Or it may simply be a case of Photoshop and the OS doing a more thorough or higher-quality job of converting the test image colors in those two test steps.
Stay tuned.

Meanwhile, here are the test results. I was able to fairly closely match the amount of disk space used by the startup disk and the Atlas RAID to those parameters in my previous tests with Panther (tests 1 thru 9). When the totals are looked at independent of the individual steps, Tiger fares poorly next to Panther. If one takes test steps #10 and #12 out of the equation, and looks at the individual test step results, the results are quite different.

For these tests, I did not disable Dashboard, but I did disable Spotlight using Spotlight preferences. Because my testing routine requires erasing the external Atlas RAID when it is to be used as Photoshop's WorkDisk/ScratchDisk, I was careful to first erase, then bring the volume into Spotlight preferences Privacy tab window before dragging a copy of the prepared test file to the volume. As long as a volume is not erased, Spotlight privacy settings are retained through restarts.

The specs for the test machine in tests 10 thru 14 are almost exactly the same as in tests 1 thru 9:

The test machine is an original G5 Dual 2GHz running Mac OS10.4.2 with 2 x 73GB WD740 Raptor internal drives formatted as a single Apple RAID 0 volume of 138.2GB, with 108.7GB available, and 29.5GB used. Installed RAM is 8GB of PC3200U-25330 400MHz DDR SDRAM.
AGP card is Radeon 9800 Pro Special Edition 256MB VRAM.ATI
An ATTO UL4D running driver version 3.4.0 is installed in PCI slot #4 (133MHz).
Attached to the card- two drives per channel- in an external Burly enclosure, are four 37GB Atlas 15K II SCSI drives formatted with SoftRAID3.2.1 as a single RAID 0 (128k stripe-size) volume of 137GB, with136.3GB available, and 649.8MB used.
This volume was used as the designated Work Disk/Scratch Disk in test #14.

PshopTest #10: 512MB test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.2GB, 108.7GB available, 29.5GB used.

Photoshop CS with default preferences except Image Cache at 6 and Maximum (memory) Used by Photoshop at 100% (1734MB of 1734MB Available RAM):

Operations Total: 61.3s
Save Total: 18.0s
Open & Save Total: 27.8s
Test Total: 89.1s
Re-draw Average: 0.3s

PshopTest #11: 512MB test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.2GB, 108.7GB available, 29.5GB used.

Photoshop CS2 with default preferences and Bigger Tiles Plug-in enabled and Maximum (memory) Used by Photoshop at 100% (3072MB of 3072MB Available RAM):

Operations Total: 66.7s
Save Total: 18.6s
Open & Save Total: 28.5s
Test Total: 95.2s
Re-draw Average: 0.383s


PshopTest #12: 1536MB (1.5GB) test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.2GB, 108.4GB available, 29.8GB used.

Photoshop CS with default preferences except Image Cache at 6 and Maximum (memory) Used by Photoshop at 100% (1736MB of 1736MB Available RAM):

Operations Total: 509.7s
Save Total: 93.8s
Open & Save Total: 123.0s
Test Total: 632.7s
Re-draw Average: 19.983s


PshopTest #13: 1536MB test file size. System, Users, File Save and Scratch Disk all on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.2GB, 108.4GB available, 29.8GB used.

Photoshop CS2 with default preferences and Bigger Tiles Plug-in enabled and Maximum (memory) Used by Photoshop at 100% (3072MB of 3072MB Available RAM):

Operations Total: 200.8s
Save Total: 50.3s
Open & Save Total: 75.4s
Test Total: 276.2s
Re-draw Average: 2.25s


PshopTest #14: 1536MB test file size. System and Users on internal 2 x 73GB WD740 Raptors Apple RAID 0 volume 138.2GB, 108.4GB available, 29.8GB used.
Test file and designated scratch disk is the 4 x Atlas 15K II striped RAID 0 volume of 137GB, with136.3GB available, and 649.8MB used.

Photoshop CS2 with default preferences and Bigger Tiles Plug-in enabled and Maximum (memory) Used by Photoshop at 100% (3072MB of 3072MB Available RAM):

Operations Total: 195.0s
Save Total: 49.9s
Open & Save Total: 70.0s
Test Total: 265.0s
Re-draw Average: 2.183s

Boots
09-29-2005, 02:40 PM
Does Photoshop run faster or slower under Tiger?

On the surface- using PshopTest in its present form- it would appear it runs slower. However, when the test details are considered, the opposite is probably true.

If we compare the test details from the fastest tests in Panther & Tiger:

First, using the 512MB test file....

Test #3: Panther, PSCS, 512MB test file, System, Users, File Save, and Scratch on same volume, 100% memory (1669MB).
Test #10: Tiger, PSCS, 512MB test file, System, Users, File Save, and Scratch on same volume. 100% memory (1734MB).

Test3**Test10**

Test step #1 Open=10.5s**9.8s
Test step #2 Rotate Canvas=3.8s**3.4s
Test step #3 Apply image=2.1s**2.0s
Test step #4 Save=4.3s**4.1s
Test step #5 Canvas size=1.2s**1.2s
Test step #6 Auto color=1.2s**1.3s
Test step #7 Guassian Blur=4.0s**4.0s
Test step #8 Sharpen=2.1s**2.2s
Test step #9 Save=3.5s**3.3s
Test step #10 Convert to Profile=11.1s**18.4s
Test step #11 Save=6.1s**5.9s
Test step #12 RGB Color=15.3s**28.8s
Test step #13 Save=4.9s**4.7s

Pre-draw Total=34.5s**55.2s
Operations Total=40.8s**61.3s
Save Total=18.8s**18.0s
Open & Save Total=29.3s**27.8s
Test Total=70.1s**89.1s

Re-draw average=0.266s**0.300s

Comparing test 3 and test 10, we see that indeed, test steps #10 and #12 show a major difference. Those steps execute much more slowly in CS under Tiger than in CS under Panther. If we take those test steps out of the equation, CS under Tiger is faster than CS under Panther:

Test 3: CS/Panther test total minus steps #10 & #12: 43.7s
Test 10: CS/Tiger test total minus steps #10 & #12: 41.9s

Next, using the 1536MB (1.5GB) test file.....

Test #6: Panther, PSCS, 1.5GBMB test file, System, Users, File Save, and Scratch on same volume, 100% memory (1670MB).
Test #12: Tiger, PSCS, 1.5GB test file, System, Users, File Save, and Scratch on same volume. 100% memory (1736MB).
Test #7: Panther, PSCS2, 1.5GBMB test file, System, Users, File Save, and Scratch on same volume, 100% memory (3072MB).
Test #13: Tiger, PSCS2, 1.5GB test file, System, Users, File Save, and Scratch on same volume. 100% memory (3072MB).

Test6**Test12**Test7**Test13:

Test step #1 Open=31.6s**29.2s**26.0s**25.1s
Test step #2 Rotate Canvas=48.6s**45.4s**13.5s**12.6s
Test step #3 Apply image=66.6s**62.1s**16.5s**10.2s
Test step #4 Save=9.3s**9.1s**9.1s**9.1s
Test step #5 Canvas size=38.6s**34.7s**8.2s**6.9s
Test step #6 Auto color=53.6s**48.8s**7.5s**6.2s
Test step #7 Guassian Blur=69.8s**62.5s**18.1s**16.2s
Test step #8 Sharpen=75.7s**71.5s**10.6s**9.2s
Test step #9 Save=8.9s**8.6s**8.9s**8.6s
Test step #10 Convert to Profile=57.5s**72.4s**35.1s**56.0s
Test step #11 Save=63.3s**61.8s**18.7s**18.0s
Test step #12 RGB Color=80.3s**112.3s**49.9s**83.5s
Test step #13 Save=14.6s**14.3s**15.5s**14.6s

Pre-draw Total=273.0s**309.7s**117.2s**167.8s
Operations Total=490.7s**509.7s**159.4s**200.8s
Save Total=96.1s**93.8s**52.2s**50.3s
Open & Save Total=127.7s**123.0s**78.2s**75.4s
Test Total=618.4s**632.7s**237.6s**276.2s

Re-draw average=21.75s**19.983s**3.416s**2.25s

Comparing test 6, 12, 7, and 13- the 1.5GB test file- we see again the same pattern; test steps #10 & #12 are slower using CS2/Panther and CS2/Tiger- with Tiger being slower than Panther. Those same steps using CS/Panther are faster than using CS/Tiger. And again, when we take those test steps out of the equation, CS/Tiger is faster than CS/Panther, and CS2/Tiger is faster than CS2/Panther:

Test 6: CS/Panther test total minus steps #10 & #12: 480.6s
Test 12: CS/Tiger test total minus steps #10 & #12: 448.0s
Test 7: CS2/Panther test total minus steps #10 & #12: 152.6s
Test 13: CS2/Tiger test total minus steps #10 & #12: 136.7s

Finally, using the 1536MB (1.5GB) test file with the WorkDisk/ScratchDisk method....

Test #9: Panther, PSCS2, 1.5GBMB test file, System and Users on startup, File Save and Scratch on Atlas RAID, 100% memory (3072MB).
Test #14: Tiger, PSCS2, 1.5GBMB test file, System and Users on startup, File Save and Scratch on Atlas RAID, 100% memory (3072MB).

Test9**Test14:

Test step #1 Open=20.7s**20.1s
Test step #2 Rotate Canvas=11.3s**9.2s
Test step #3 Apply image=13.0s**9.7s
Test step #4 Save=9.1s**8.9s
Test step #5 Canvas size=8.3s**5.9s
Test step #6 Auto color=7.8s**6.2s
Test step #7 Guassian Blur=18.2s**16.2s
Test step #8 Sharpen=11.9s**9.2s
Test step #9 Save=9.1s**8.6s
Test step #10 Convert to Profile=35.2s**54.4s
Test step #11 Save=18.9s**17.9s
Test step #12 RGB Color=51.2s**84.2s
Test step #13 Save=15.3s**14.5s

Pre-draw Total=119.0s**166.8s
Operations Total=156.9s**195.0s
Save Total=52.4s**49.9s
Open & Save Total=73.1s**70.0s
Test Total=230.0s**265.0s

Re-draw average=3.05s**2.183s

Test 9: CS2/Panther test total minus steps #10 & #12: 143.6s
Test 14: CS2/Tiger test total minus steps #10 & #12: 126.4s

So- with PshopTest steps #10 & #12 taken out of the test results, Photoshop CS2 is clearly faster than CS, and both of them run faster under Tiger than under Panther.....

Now it's off to investigate this thing with those two test steps, Convert to Profile and RGB Color. ?????

kaye
11-29-2005, 08:53 AM
With the help of Boots, TZ, and other Moderators, I have been experimenting with larger file sizes in the Pshop Test. There is a lot to note about this, so bear with me. 3GB-5350 resolution and 6GB-7565 resolution sizes were the target but I only tested the 6GB size. Using Pshop CS2 v9.0.

Notes:
1. Both the 3GB-5350 resolution and 6GB-7565 resolution sizes created have to be saved in Large Document Format (.psb).
2. 3GB file saves 845.5MB and 6GB file saves 1.19GB and at end of test (after last Save Step13), the 6GB file is 4.03GB.
3. Pshop timer is definitely off, sometimes way off on the low side. One thing I cannot measure is when the Pshop timer is actually done, before I can see the time or as I see the time. Can only stop the stopwatch as I see the Pshop timer displayed.
4. Getting used to using the stopwatch and not stopping it on the Pre-draw caused me one false start and a reboot. Forgetting to zero the stopwatch after a step caused another false start and another reboot.
5. Watching the Activity Monitor showed the SCSI/SATA RAID at mostly 40-60MB/sec, same during Save, sometimes during some steps up to 200MB/sec write spikes but never that high for reads nor during Save.
6. Progress bars very slow. At least one step early on bar actually backed up a couple of times. Some other steps do the same for me. Another test early on got a Histogram progress bar before the regular progress bar which I have never see before.
7. Forget about Pre-draw times.
8. So you need a stopwatch for these larger file sizes. I ended up purchasing a large digital at a sporting goods store, last year's model, for $15.
9. The 6GB file size starts at 1.19GB. By the end of Step1-10GB, after Step6-40GB, Step10-65GB, reaches a peak of 79+GB during Step13 and 75GB on completion of the step. After completion of the test and quitting Pshop CS2, it is 4.03GB.
10. For the current published smaller file sizes Pshop Test, an 8x or 12x drive RAID may be very slightly slower or faster than a 4x drive RAID, insignificant at about 1s difference. But if you do work with large document sizes, more drives is definitely better. Rambling on...

These Pshop Test CS2 v9.0 tests on G5-2.7DP, set Highest Performance, 8.0GB RAM, Radeon X800 XT ME. All tests in CS2 are default, Available RAM 3072MB and 100% = 3072MB, plus Bigger Tiles plug-in enabled. Default History States 20, default Cache Levels 6. TEST FILE SIZE 6GB-7565. Boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, and Users. Ethernet, USB printer, and FW disconnected. Spotlight and Dashboard disabled.

TEST1- 10.4.2, SoftRAID 3.2.1 RAID0 Workstation 128KB SU, using stopwatch. File Save Drive and Scratch Disk Drive 4x Raptor74 and 4x Atlas 36GB 15K II, partitioned to 144GB total, 6-channels, not journaled. Using SeriTek/1VE4 (ROM revision 5.1.2.1VE4) in 100MHz slot, UL4D 133MHz slot. Gurus one meter external SATA data cables. Other UL4D in 100MHz slot nothing connected.

TEST2- Same as TEST1 except AppleRAID 256KB SU and not partitioned, 273GB.
TEST3- Same as TEST1 except not partitioned, 274GB.
TEST4- Same as TEST1 except 12x FrankenRAID with 4x 18GB X15.3 added to UL4D in 100MHz slot, 8-channels, not partitioned, 205GB.
TEST5- Differences,10.4.3, Apple driver. Boot drive, 300GB MaxLineIII-16MB cache used as boot, CS2 app, Users, File Save, and Scratch Disk Drive, everything. Journaled, 279.36GB Capacity, 38.67GB Used, 13.9%, 240.68GB Available.

TEST1**TEST2**TEST3**TEST4**TEST5:
1. Open=63.58s**63.95s**63.00s**62.60s**203.09s
2. Rotate Canvas 180=133.76s**136.35s**133.00s**120.90s**469.36s
3. Apply Image=124.81**138.54s**122.60s**121.40s**302.90s
4. Save=83.66s**88.52s**84.25s**82.51s**136.60s
5. Canvas Size=91.70s**96.35s**89.10s**86.10s**186.43s
6. Auto Color=119.03s**129.55s**120.10s**117.75s**230.23s
7. Gaussian Blur=147.24s**155.86s**144.30s**141.90s**277.89s
8. Sharpen=114.43s**xxx.xxs**111.40s**107.50s**217.67 s
(see Note1 below)
9. Save=97.62s**xx.xxs**96.51s**93.28s**159.90s
10.Convert to Profile=279.26s**xxx.xxs**280.58s**271.20s**493.18 s
11.Save=226.08s**xxx.xxs**230.14s**221.90s**479.10 s
12.Convert to RGB=345.67s**xxx.xxs**343.78s**337.75s**481.86s
13.Save=135.83s**xxx.xxxs**133.59s**129.21s**315.3 3s

LineA. Operations Total
Operations Total=1355.90s**xxxx.xxs**1344.86s**1304.50s**2659 .52s
LineB. Save Total=543.19s**xxx.xxs**544.49s**526.90s**1090.93s
LineC. Open and Save Total=606.77s**xxx.xxs**607.49s**589.50s**1294.02s
LineD. Test Total=1962.67s**xxxx.xxs**1952.35s**1894.00s**3953 .54s

NOTE1: TEST2 Step 8 the volume unmounted and I got a Pshop warning as well as a MacOS Tiger warning and had to force quit Pshop. That is a first. And there was an invisible large file on the volume after I rebooted. I fixed that in SR 3.2.1. So much for AppleRAID and 256KB SU which was significantly slower up to this point anyway. SoftRAID forever on.
NOTE2: TEST3 with drives not partitioned was about 10s faster than TEST1 with drives partitioned, plus many more GBs.
NOTE3: TEST4 with the 12x FrankenRAID by far the fastest Test Total and all of the steps.
NOTE4: TEST5, because the files sizes increase to 79+GB during the test and the desire was to test a single drive, my RAID drives were too small for File Save and Scratch Disk Drive, 74GB Raptor, 18GB X15.3, and 36GB Atlas 15K II. SeriTek/1VE4 in 100MHz slot, UL4D 133MHz slot, and UL4D in 100MHz slot nothing connected to any of the cards. MaxLineIII is in the B (lower) drive position. Before rebooting I checked its temperature at 40.0C. After completion of the test, launched Hardware Monitor, still at 40.0C. Was curious because the drive went thru a humongous amount of work.
k

kaye
01-09-2006, 07:24 PM
Computer: Power Mac G5 Dual 2.7GHz Boot ROM vers: 5.2.4f1
OS: Mac OS 10.4.3
Graphics card: ATY,R420; X800XT ME
PCI cards: ATTO ExpressPCI UL4D 1.5.0 ROM, ATTOExpressPCI4 3.5.0 driver, ATTO Configuration Tool 3.04
Display: Apple Cinema HD Display 23" 1920 x 1200 resolution, Millions
Installed RAM: 8.0GB, 8x1GB DDR SDRAM PC3200U-20320, CL2 (5ns)

System settings:
Processor Performance: Highest
DSL and audio system connected and running. Inkjet printer connected, not running.
Dashboard running in background, Spotlight on but all running RAIDs on "privacy", no Finder windows open, Activity Monitor not running, no antivirus or disk protection software.
Journaling on all boot drives, not on RAIDs.

Photoshop settings:
Photoshop CS2 (9.0)
Memory: Available RAM 3072MB, Maximum Used by Photoshop: 100% or 3072MB.
Image Cache: Default 6 all tests
Bigger Tiles Plug-in installed & active.

Hard drives tested: see below

TEST1-512MB file (2183.7 res). Boot, Users/Data Volume MLIII-16MB, File Save and Scratch SATA Burly 4x Raptor74-8MB, SeriTek/1VE4, 100MHz slot, SoftRAID 3.3, Workstation, 128KB SU, 4-channels, 4 drives. 2x UL4D 133 and 100MHz slots, nothing connected.
TEST2-Same as TEST1 except File Save and Scratch SCSI/Atlas Burly 4x 36GB Atlas 15K II-8MB, UL4D, 133MHz slot, 2-channels, 4 drives. SATA Burly off.
TEST3-Same as TEST1 except File Save and Scratch both above SATA and SCSI/Atlas Burlys SoftRAIDed together, 6-channels, 8 drives. 1x UL4D 100MHz slot nothing connected.
TEST4-Same as TEST3 except File Save and Scratch SCSI/Cheetah Burly 4x 18GB X15.3 added to other Burlys, UL4D, 100MHz slot, 8-channels, 12 drives.

TEST1**TEST2**TEST3**TEST4
512M***512M***512M***512M
*78s****78s*****77s****77s

TEST5-1.5GB file (3780 res). Boot, Users/Data Volume MLIII-16MB, File Save and Scratch SATA Burly 4x Raptor74-8MB, SeriTek/1VE4, 100MHz slot, SoftRAID 3.3, Workstation, 128KB SU, 4-channels, 4 drives. 2x UL4D 133 and 100MHz slots, nothing connected.
TEST6-Same as TEST5 except File Save and Scratch SCSI/Atlas Burly 4x 36GB Atlas 15K II-8MB, UL4D, 133MHz slot, 2-channels, 4 drives. SATA Burly off.
TEST7-Same as TEST5 except File Save and Scratch both above SATA and SCSI/Atlas Burlys SoftRAIDed together, 6-channels, 8 drives. 1x UL4D 100MHz slot nothing connected.
TEST8-Same as TEST7 except File Save and Scratch SCSI/Cheetah Burly 4x 18GB X15.3 added to other Burlys, UL4D, 100MHz slot, 8-channels, 12 drives, FrankenRAID.

TEST5**TEST6**TEST7**TEST8
*1.5G***1.5G***1.5G***1.5G
*251s***244s***240s***240s

TEST9-3.0GB file (5350 res, .psb Large Document Format file). Boot, Users/Data Volume MLIII-16MB, File Save and Scratch SATA Burly 4x Raptor74-8MB, SeriTek/1VE4, 100MHz slot, plus SCSI/Atlas Burly 4x 36GB Atlas 15K II-8MB, UL4D, 133MHz slot, SoftRAID 3.3, Workstation, 128KB SU, 6-channels, 8 drives. 1x UL4D 100MHz slot, nothing connected.
TEST10-Same as TEST9 except File Save and Scratch SCSI/Cheetah Burly 4x 18GB X15.3 added to other Burlys, UL4D, 100MHz slot, 8-channels, 12 drives, FrankenRAID.
TEST11-6.0GB file (7562 res, .psb Large Document Format file). Otherwise same as TEST9.
TEST12-6.0GB file (7562 res, .psb Large Document Format file). Otherwise same as TEST10.

TEST9**TEST10**TEST11**TEST12
*3.0G****3.0G****6.0G****6.0G
*597s****594s***1864s***1786s

The best I can do with dual processors and the current drives I have. k

kaye
01-31-2006, 08:16 AM
PshopTest2 Results
Computer: Power Mac G5 Dual 2.7GHz Boot ROM vers: 5.2.4f1
OS: Mac OS 10.4.4
Graphics card: ATY,R420; X800XT ME
PCI cards:
1. ATTO ExpressPCI UL4D, 133MHz slot, 1.5.0 ROM, ATTOExpressPCI4 3.5.0 driver, ATTO Configuration Tool 3.04
2. SeriTek/1VE4, 100MHz slot
3. SeriTek/1VE4, 100MHz slot
Display: Apple Cinema HD Display 23" 1920 x 1200 resolution, Millions
Installed RAM: 8.0GB, 8x1GB DDR SDRAM PC3200U-20320, CL2 (5ns)
For all configurations below, Boot, Users/Data Volume MLIII-16MB

System settings:
Processor Performance: Highest
DSL and audio system connected and running. Inkjet printer connected, not running.
Dashboard running in background, Spotlight on but all running RAIDs on "privacy", no Finder windows open, Activity Monitor not running, no antivirus or disk protection software.
Journaling on all boot drives, not on RAIDs.

Photoshop settings:
Photoshop CS2 (9.0)
Memory: Available RAM 3072MB, Maximum Used by Photoshop: 100% or 3072MB.
Image Cache: Default 6 all tests
Bigger Tiles Plug-in installed & active.

Configurations:
Conf1- File Save and Scratch 2x 150GB Raptor1500, 2-channels, SoftRAID 3.3, Workstation, 128KB SU, 279GB available
Conf2- File Save and Scratch 6x Raptors (4x Raptor74, 2x Raptor1500), 6-channels, SoftRAID 3.3, Workstation, 128KB SU, 415GB available
Conf3- File Save and Scratch 6x Raptors and 4x 36GB Atlas 15K II, 8-channels, SoftRAID 3.3, Workstation, 128KB SU, 342GB available

TEST1**TEST2**TEST3**TEST4**TEST5**TEST6
512M***512M***512M**1.5GB***1.5G***1.5G
Conf1**Conf2***Conf3**Conf1***Conf2**Conf3
*83s***76s*****77s***282s****243s***239s

TEST7**TEST8**TEST9**TEST10
*3.0G***3.0G***6.0G****6.0G
Conf2***Conf3**Conf2***Conf3
*609s***593s**1930s***1859s

Comments:
1. For the 3 and 6GB file tests, I was able to monitor Peak Watts (PW) used.
Conf2, 6x Raptors, 592PW during the Convert Mode (To: RGB color mode)
Conf3, 10x Drives, 664PW during the Convert Mode. Almost the entire test, well above 600PW except during Saves where I could hear the internal fans slow down and mostly saw 530-550PW. 510PW in the Finder at idle, no tests running.
2. I made no attempt to compare these Configurations with previous PshopTest2 results except to say that while this Conf3 10x Drive RAID tests faster in QuickBench, the previous 12x drive FrankenRAID did test faster in the 6GB file test and 10.4.3. 10.4.4 may be slower, don't know. However SCSI can still be king in Pshop.
3. This ability to stack the cases however gives me much more foot room under the desk and much easier connection of cables to the G5.
4. PshopTest2 runs very well and SoftRAID 3.3 brings all of these dissimilar drives together in a bulletproof way. And checking SoftRAID after testing, I have never seen a report of I/O errors.
k

kaye
02-13-2006, 11:13 AM
Yesterday I ran PshopTest2 in CS2 to view what runs single processor (SP) and what runs dual (DP) via Activity Monitor. After the action steps in preparing the file, I chose the 1.5GB file size, there are 13 action steps to be timed. The percentages vary during each step so these are the max observed:
1. Open, 99.9%, SP
2. Rotate 180, 170%, DP
3. Apply Image, 170%, DP
4. Save, 100%, SP
5. Canvas Size, 175%, DP
6. Levels, 172%, DP
7. Gaussian Blur, 181%, DP
8. Sharpen, 179%, DP
9. Save, 99.9%, SP
10.Convert to Profile, 185%, DP
11.Save, 99%, SP (however during Prepare to Save, 136%, DP)
12.Convert Mode (to RGB), 199.5%, DP
13.Save, 100%, SP

So all of the Image and Filter action steps are DP aware, all of the Saves are SP except for #11 during the initial Prepare to Save where it is DP aware. k

Boots
02-13-2006, 07:03 PM
Thanks for that analysis Kaye. So PshopTest is a bit weighted in favor of multi-processors it seems.

Btw- when I was preparing those big-time .psb files to test the 150GB Raptors, I experienced the same phenomenom you did- progress bars going backwards before eventually going forwards. It gives a twist to that adage "two steps forwards, one step back". ;)

kaye
02-13-2006, 07:28 PM
I've got the ultimate backwards progress bar for you. After I created overnight that 8.9GB Finder file, I decided to open it. I forget that exact window name, Open or Prepare to Open, but anyway the bar excruciatingly slowly traveled to the right, stayed full right for awhile, then disappeared and another window of the same name the bar started empty then another similar slow progress to the right. Two pregress windows of the same name. Took over an hour to finish opening that 192+GB (?) file. k

kaye
10-21-2006, 07:43 AM
bump

Boots
11-27-2006, 08:09 AM
Since I purchased the two 150GB Raptors (WDC WD1500ADFD-00NLR0 rev. 19.06P19) back in January, I've run them in various configurations from the internal SATA bus. The fastest config for PshopTest turns out to be both drives striped together as a single 279.2GB startup volume containing all files.

Here are some comparisons using PshopTest2.1 (soon to be released):

512MB test file, one Raptor drive as a single standard startup volume containing all files including the test file, and also used as the scratch disk. Previous best is 65s

512MB test file, 2 x Raptor drives striped as a single standard startup volume containing all files including the test file, and also used as the scratch disk. New best results are 59s

----

512MB test file, one Raptor drive as a single standard startup volume containing all files including the test file but not the scratch file. 4 x striped Atlas 15k SCSI used as the scratch disk. Previous best is 63s

512MB test file, 2 x Raptor drives striped as a single standard startup volume containing all files including the test file but not the scratch file. 4 x striped Atlas 15k SCSI used as the scratch disk. New best results are 57s

----

512MB test file, one Raptor drive as a single standard startup volume containing all files except the test file and scratch file. 4 x striped Atlas 15k SCSI used as a Work Disk/Scratch Disk containing the test file and the scratch disk. Previous best is 56s

512MB test file, 2 x Raptor drives striped as a single standard startup volume containing all files except the test file and scratch file. 4 x striped Atlas 15k SCSI used as a Work Disk/Scratch Disk containing the test file and the scratch disk. New best results are 55s

----

I plan to run this striped config awhile. Of course I back everything up at least twice a day with Retrospect (sometimes more) and do a fresh CCC clone at least once a week.

ricks
11-27-2006, 11:42 AM
Thanks for the post Boots. We get asked on a daily basis what the best configuration is, all your posts really help.

If you have a spare moment, could you post your best thoughts on how a good setup would be using, say, 4 total SATA drives? A pair of Raptors and a pair of big 7200s. All 7200s. Just a reasonable thought on what our recommendations should be for reasonably large PS photo editing.

Again, you are remarkable. All the configurations you have tried and keeping track of the performance for each. Amazing!

Rick

Boots
11-29-2006, 06:36 AM
If you have a spare moment, could you post your best thoughts on how a good setup would be using, say, 4 total SATA drives? A pair of Raptors and a pair of big 7200s. All 7200s. Just a reasonable thought on what our recommendations should be for reasonably large PS photo editing.
Rick

A safe easy maintenance setup:

The MaXLine III's as I understand it are supposed to still be generally the best overall SATA performers, so one for Startup, one for backup, and Raptors striped as scratch disk. But TZ is the expert on the current state of which large SATA drives are the best choices. Anyhow,

A live dangerously approach with moderately easy maintenance:

Raptors striped as Startup, 2 SATA drives striped as 2 volumes, first as scratch, second for image files. External backups.

or

Another live dangerously approach with moderately easy maintenance:

2 x large SATA drives striped as Startup and all files, 2 x Raptors striped as scratch. External backups.


Maybe slower but safer and easy maintenance:

1 Raptor as Startup, 1 Raptor as scratch, 1 SATA as image file volume, 1 SATA as backup.

I don't know which of these would be the fastest Photoshop performer- I'd have to test them.

Boots
11-29-2006, 06:51 AM
If I had the $ and a Mac Pro, first thing I'd try would be 4 x 150G Raptors striped together containing all files and scratch, and relying on external backups.

Then I'd test 2 x stripe startup, 2 x stripe scratch- again relying on external backups.

kaye
11-29-2006, 04:13 PM
PshopTest2.1
Ran it twice on 512MB file size (2183.7 res, 273.8MB Finder starting size, 512.0MB Pshop size after opened), Finder size after test completion 388.8GB.
Test1
Mac Model: G5-2.7GHz DP
Mac OSX version: 10.4.8
Video Card(s): ATI Radeon X800XT ME 256MB VRAM
Display(s): 23" Apple Cinema HD Display, 1920x1200, Millions
PCI Card(s): 2x SeriTek/1VE4 in 100MHz slots, Fw800 card in 133MHz slot
Installed RAM: 8GB
System Settings notes: Activity Monitor running: N, Internet: Y
System Volume
Hard Drive info: 279.5GB SATA Maxtor MaxLine III, 7L300S0, Spotlight Y
Volume Name: 7LMIII-16MB, Partition Number: 1 of 1
Volume Formatted Size: 279.5GB, Available Space: 144.8GB , Used Space: 134.7GB
Disk Driver Software and Version: Apple DU 10.5.6, Journaling: Y
Users/Data Volume
Same as System Volume
File Save Volume
Same as System Volume
Scratch Volume
4x150GB WD1500ADFD, Firmware 2x 00NLR0, 2x 00NLR1
1xSeriTek/1VE4, Spotlight N
Volume Name: 4xRaptor150GBRAID, Partition Number: 1 of 1
Volume Formatted Size: 558GB, Available Space: 558GB , Used Space: 0
Disk Driver Software and Version: Apple RAID DU 10.5.6, Journaling: N, 128KB stripes
Test2
Only differences from Test1 are Data Volume and File Save Volume are the RaptorRAID. Scratch Volume still the RAID Volume
Test3
Same as Test2 except file sized 1.5GB
Test4
Same as Test2 except file sized 3.0GB
Test5
Same as Test2 except file sized 6.0GB

Photoshop Version: CS2 9.0.2
Adjusted Refresh Plug-in: N, Bigger Tiles Plug-in: Y, Image Cache: 6
Photoshop Memory Settings
Available RAM: 3072MB, Max Used by Pshop: 100% = 3072MB

*Test1****Test2***Test3****Test4****Test5
512MB***512MB***1.5GB***3.0GB****6.0GB
*80s******67s*****198s****581s****1944s
Observations:
1. I added the fact that for all of the tests I used RAID 128KB stripes.
2. Keeping the System Volume out of the test mix in Test2 does make a difference, especially since it is 48% full, and even more full during the test.
3. The test runs great Boots. Very nice and thoughtful work. No complaints at all.http://macgurus.com/forums/images/smilies/phat.gif
4. The 6.0GB file size test, 60s pauses, took somewhat longer than tested before, previous best 1859s, but I always used an 11 or 12 drive RAID. 85s more is a small price to pay for simplicity of a 4-drive RAID and more foot room.
5. For most of the 6.0GB tests, 60s pauses were more than ample, longest I saw was 40-45s of activity after the Progress window disappeared. That is until I got near the end of the tests. For instance, Convert to Profile and Convert Mode. But the 60s pause timer is smart. In Convert Mode:
Progress window closed, 67s later disk activity stopped, another 45s it went on to the next step. So 60s pauses were good for me because the start of the pause did not start when the Progress window closed.
6. My calculations for the 6GB test with 60s pauses:
Total test time 45m24s = 2724s, 13x60s pauses = 780s
2724s - 780s = 1944s
k

Boots
11-30-2006, 03:57 AM
Thank you Kaye!

As you can see, test results are faster tho' test takes longer to do. But now fractional differences between first timed open step and starting a stopwatch and 'test is complete' message and stopping a stopwatch will have even less impact on 'randomizing' a final score.

But the main thing is: with the built-in 25s pauses, the scratch disk is allowed to fully write/update/optimize between steps- at least up to a 3GB file size. I believe this in turn allows the RAM-caching of scratch data to also occur in an optimal fashion so there is less memory fragmentation during the auto part of the test. The net should be more consistent scores with large amounts of installed RAM.

I tried to keep the pause value as low as possible. Originally I started with 10s but it proved to be too little with larger file sizes. 20s was on the edge with the 3GB test file size so I finally boosted it to 25s. The upside of the pauses, it does allow one to sort of walk away a bit from the computer- instead of anxiously standing by. I also think it helps to have the 'test complete' message occuring after a pause after the last save step. It gives you time to get ready to stop the timer.

kaye
11-30-2006, 07:01 AM
Yes I see all that. And with the 3GB test I was able to watch the RAID's drive activity lights and did see activity stop before the start of the next step.

I was going to do the 6GB test yesterday but ran out of steam and decided to take a rest. I'm wondering, that test used to take me 31 minutes, 1859s was my best score with considerably more drives. Maybe I should set the pause for 60s and see how that goes. All I need is to see drive activity stop for several seconds to be sure the step completes, right? Or, since some of the later steps take considerably longer than earlier ones, maybe 75s? I don't know, what do you think? If I have this pause thing right, after a step completes, the timed pause starts once a step is complete but the drives/RAID will still be working to "maximizes scratch disk and memory optimization". So the pause may need to be much greater with a 4-drive RAID and my G5 DP as opposed to your Quad. Also tho I have 8GB RAM, it is much slower than yours. So many considerations. k

Boots
11-30-2006, 12:03 PM
Maybe I should set the pause for 60s and see how that goes. All I need is to see drive activity stop for several seconds to be sure the step completes, right? Or, since some of the later steps take considerably longer than earlier ones, maybe 75s?

With the 6GB test file, ya- to be on the safe side, probably 60s.



I don't know, what do you think? If I have this pause thing right, after a step completes, the timed pause starts once a step is complete but the drives/RAID will still be working to "maximizes scratch disk and memory optimization". So the pause may need to be much greater with a 4-drive RAID and my G5 DP as opposed to your Quad. Also tho I have 8GB RAM, it is much slower than yours. So many considerations. k

Yes, the pause starts after the actual operation completes; the scratch disk then has a chance to do its thing completely. You've probably noticed- after most of the scratch disk drive activity occurs, then there's a little pause, then a bit more drive activity. That little bit of more activity is PSCS2 cleaning up and optimizing the scratch disk data- which I think also includes the part where it updates the scratch data held in RAM.

If you want to check this out with Activity Monitor, there should a good 5s between when all drive activity ceases and the beginning of a next step- at least on several steps. I think the routine needs a least a few steps where the scratch optimization can 'catch up' so to speak, in order for full 'optimization' to occur. It may not have to occur on every step, but I could be wrong. I'm hoping 25s is a good figure for most testers whom I'm guessing are likely to stay under the 1.5GB test file size, but your question has prompted me to wonder whether I should provide 2 choices- say 30s (all G5s & Mac Pro) and 60s (all G4s and less), depending on computer...

On a G5 Quad, 10s is more than enough for the 512MB size.

kaye
11-30-2006, 05:03 PM
My previous test thread has been updated to include RAID stripe size, Test5, a 6.0GB test, and more comments about that test. Saw as high as 330MB/s disk activity in Activity Monitor. k

Boots
12-01-2006, 08:00 AM
Thanks for the observations and updated results Kaye. Sounds like 60s is good for 6GB test file, and 25 or 30s good for all others. I may include something on this in an updated version..

Boots
01-08-2007, 04:43 PM
Installed the CS3 beta.

PshopTest2.1 appears to run fine without any apparent need for revising. (he says while keeping fingers crossed!)

My first few tests with it:

G5 Quad 2.5GHz, 16GB RAM, Mac OS 10.4.8
PS CS3b set to default prefs except 100% RAM = 3072MB, cache level 6 (default)

1. 512MB test file, 2 x 150G Raptor SATA drives as a single 279GB volume (227GB available) stripe as startup system, users, file open & save, and scratch:

58s

2. 512MB test file, 2 x 150G Raptors SATA drives as a single 279GB volume (227GB available) stripe as startup system, users, and file open & save. Scratch on 4 x Atlas 15k II SCSI drives as a single otherwise empty 136.5GB volume stripe:

57s

3. 512MB test file, 2 x 150G Raptors SATA drives as a single 279GB volume (227GB available) stripe as startup and users. File open & save and scratch on 4 x Atlas 15k II SCSI drives as a single otherwise empty 136.5GB volume stripe:

57s

-----

So from first appearances (for PPC only), slightly faster than CS2 when using startup as scratch, but about the same or a tad slower when using external stripe as scratch or work disk/scratch disk.

It'll be interesting to compare results using larger test files.

There is no Bigger Tiles plugin. When I attempted to use a copy of CS2's Bigger Tiles plugin, CS3b said it could not load it and it wasn't compatible during startup of the app.

There is an implication from the scratch disk prefs in CS3b, that internal drives are preferable to external. Also notice the listing of the graphics card; apparently Adobe was planning on Photoshop using the graphics card for computing in some way but it didn't pan out yet for this beta (according to info on the CS3 forum):

Boots
01-12-2007, 04:58 PM
Using PshopTest2.1 (http://www.macgurus.com/forums/showthread.php?t=20218), I did some comparison speed testing between Photoshop CS2 (9.0.2) and the CS3 beta.

The test machine is a G5 Quad with 16 GB of RAM running Mac OS X 10.4.8. All files including the scratch file (except where noted), are on two matching 150GB Raptor internal hard drives striped together in RAID0 as a single 279GB volume with 227GB of available capacity.

On certain test runs, an external SCSI array was used as Photoshop's first scratch disk. The array consists of four 36GB Maxtor Atlas 15K II drives striped together as a single empty 136GB RAID0 volume used only for Photoshop's scratch file.

During the tests no other applications were running- but except where noted Activity Monitor was pre-loaded as a startup item. CS2 had the Bigger Tiles plugin installed and enabled, but not the DisableVMbuffering plugin.

Versions of these plugins for CS3 were not included with the download. In CS3, a copy of the CS2 version of the Bigger Tiles plugin would not load, returning an error message: "Could not load the Bigger Tiles module because it does not work with this version of Photoshop."

Except where the scratch disk and memory settings were altered, the tests used the default installation Photoshop preferences.

Memory settings for both CS2 and CS3 were exactly the same; when set to 70% this translated to 2150MB. When set to 100%, this became 3072MB.

The test results (lower scores are better):

Photoshop CS2 and the 512MB test file
Test #1: scratch on startup, memory at 70%: 59s
Test #2: scratch on RAID array, memory at 70%: 53s

Photoshop CS3 beta and the 512MB test file
Test #3: scratch on startup, memory at 70%: 54s
Test #4: scratch on RAID array, memory at 70%: 52s

Photoshop CS2 and the 1.5GB test file
Test #5: scratch on startup, memory at 100%: 180s (3 min.)
Test #6: scratch on RAID array, memory at 100%: 167s (2 min. 47 sec.)
Test #7: scratch on startup, memory at 70%, Activity Monitor running: 223s (3 min. 43 sec.)

Photoshop CS3 beta and the 1.5GB test file
Test #8: scratch on startup, memory at 100%: 785s (13 min. 5 sec.)
Test #9: scratch on RAID array, memory at 100%: 756s (12 min. 36 sec.)
Test #10: scratch on startup, memory at 70%, Activity Monitor running: 1504s (25 min. 4 sec.)

-----

During the 1.5GB tests where Activity Monitor was also running, I observed memory usage closely. CS2 (and the OS) used every bit of the 16GB of installed RAM- even paging a bit towards the end of the test.

In marked contrast, CS3 and the OS never used more than 4.5GB of RAM; scratch disk activity was pretty well constant.

Comparing these results, I would hazard to guess the buffering of scratch data in RAM is not yet implemented in CS3.

Boots
01-23-2007, 09:24 AM
According to Adobe, a Bigger Tiles plugin was not available at the time of the release of the CS3 public beta, and the VM buffering is not yet enabled.

I'm looking forward to eventually posting CS3 results where a Bigger Tiles plugin is installed and the caching of scratch disk data in RAM is enabled (VM Buffering).

My tests with the CS3 beta and the 512MB test file show it is already measurably faster on PPC, even without those two speed enhancements.

I am also finding that because of the Mac OSX virtual memory 'pause' bug, I need to run PshopTest2.1 2-4 x in each test configuration- both for CS2 and CS3 beta- to obtain the fastest possible results. This was not the case with CS or 7 where results could be reliably kept to within 1 or 2 seconds over a series of repeated tests with the same test configuration.

Individual test results can vary by as much as 5-10 seconds otherwise- despite using a scrupulous and exacting test preocedure where every little variable is kept the same ie., no internet, no auto scripts, no indexing, the exact same background processes and disk space between tests, waiting a full minute after desktop is loaded after restart, a full minute before Photoshop is started, a full minute before starting auto portion of test etc.

Boots
02-01-2007, 08:36 AM
My new best results with the 512MB test file. PS7 and CS were each tested twice, results exactly the same; CS2 was tested 15x with results varying up to 5 seconds but pretty consistently 54s, 55s):

Photoshop 7.01, 4GB installed RAM, 2x150G Raptor stripe as startup (279.2GB, 245.2GB available, 34.1GB used) with scratch and test file on startup, 70% memory allocation (1819MB available 1273MB used by Photoshop): 77s

Photoshop CS, 4GB installed RAM, 2x150G Raptor stripe as startup with scratch and test file on startup, 70% memory allocation (1819MB available 1273MB used by Photoshop): 54s

Photoshop CS2, 4GB installed RAM, 2x150G Raptor stripe as startup with scratch and test file on startup, 70% memory allocation (3072MB available 2150MB used by Photoshop), with BiggerTiles enabled: 54s

Then I re-installed all 16GB RAM, testing 7 and CS 3x each;results were exactly the same. I tested CS2 once more with the result below.

Photoshop 7.01, 16GB installed RAM, 2x150G Raptor stripe as startup with scratch and test file on startup, 70% memory allocation (1533MB available 1073MB used by Photoshop): 90s

Photoshop CS, 16GB installed RAM, 2x150G Raptor stripe as startup with scratch and test file on startup, 70% memory allocation (1533MB available 1073MB used by Photoshop): 65s

Photoshop CS2, 16GB installed RAM, 2x150G Raptor stripe as startup with scratch and test file on startup, 70% memory allocation (3072MB available 2150MB used by Photoshop), with BiggerTiles enabled: 53s

--------

Notice the huge difference in the amount of memory PS7 and PS CS sees when 16GB RAM is installed. This was also the case when I tested 4 or 8GB in my earlier G5 Dual 2GHz machine.

And the surprise for me too is how fast CS is relative to CS2- before this I had believed the speed difference between each of these 3 versions was roughly the same ie. CS was 20% faster than 7, CS2 was 20% faster than CS and so forth.
Instead my tests are showing CS to be just as fast as CS2 with smaller files. It's only when you get to larger amounts of installed RAM and bigger files that CS2 leaves CS in the dust.

Also- these results, and a bunch of others with CS3b- definitely confirm for me that a 70% memory allocation is much better when working with smaller files. I still think a 100% allocation is an option when working with large files. You get a faster overall result but file saves suffer because the OS is being starved, and other specialized plugins may suffer also.

In my literature I will likely back off my "100% allocation with 4 or more GB RAM installed" recommendation and instead recommend 70% with the 100% option as a specialized tactical strategy for big files under some circumstances...

kaye
03-14-2007, 11:24 AM
Pshop Test 2.1, Cache Levels 6, G5, 10.4.8, SoftRAID 3.6, 128KB SU. Photoshop CS2 v9.0.2.

512MB file, Memory 70% 2150MB, 30sec interval, 6xRaptor150, 64sec
512MB file, Memory 100% 3072MB, 30sec interval, 6xRaptor150, 68sec
512MB file, Memory 70% 2150MB, 30sec interval, 4xAtlas15K2, 65sec
512MB file, Memory 100% 3072MB, 30sec interval, 4xAtlas15K2, 67sec
Previous 2.1 test with 4xRaptor150, 67sec
------------
1.5GB file, Memory 70% 2150MB, 30sec interval, 6xRaptor150, 235sec
1.5GB file, Memory 100% 3072MB, 30sec interval, 6xRaptor150, 196sec
1.5GB file, Memory 100% 3072MB, 30sec interval, 4xAtlas15K2, 201sec
Previous 2.1 test with 4xRaptor150, 198sec
------------
3.0GB file, Memory 100% 3072MB, 60sec interval, 6xRaptor150, 568sec
3.0GB file, Memory 100% 3072MB, 60sec interval, 4xAtlas15K2, 582sec
3.0GB file, Memory 100% 3072MB, 60sec interval, both for 10 drives, 564sec
Previous 2.1 test with 4xRaptor150, 581sec
------------
6.0GB file, Memory 100% 3072MB, 60sec interval, 6xRaptor150, 1836sec
6.0GB file, Memory 100% 3072MB, 60sec interval, 4xAtlas15K2, 1917sec
6.0GB file, Memory 100% 3072MB, 60sec interval, both for 10 drives, 1780sec
Previous 2.1 test with 4xRaptor150, 1944sec
------------
Some Observations:
1. The 30sec intervals for 512MB file tests, I could have used less, but easier for me to calculate for subtracting the 13 intervals from the stopwatch total to yield the test result.
2. For my G5 and RAID the 1.5GB file tests, 30sec interval for some of the later steps comes pretty close to the 30sec limit. You want all drive operations completed, including cleanup and whatever else is going on which does not happen immediately after a particular step completes. Pshop does some background stuff.
3. For me with 8GB RAM, the 512MB file size, the results time difference between Memory 70% and 100% is not significant enough for me to reset Prefs to 70%. And the 1.5GB file test, 70% is much slower anyway.
4. 1.5GB file size, 4xRaptor150 is only 2sec more than 6xRaptor150. If that is the limit of your PS work, then 2 more Raptor150 drives is a waste.
5. The same can be said for the 3.0GB file size, less than 20sec time difference. I will say that even tho the stopwatch doesn't lie, 6xRaptor150 does feel during the test as if there is more breathing room.
6. If 6.0GB file file size (largest scratch file size at the end of the test is 71GB) is your typical, with Cache Levels 6, PS does benefit from as many large and fast drives as you can afford. k

Boots
03-14-2007, 04:20 PM
Nice results Kaye.

I would think if you could stripe two of the 150G Raptors as an internal startup disk with testfile open/save, then 4 x 150G Raptors striped as external scratch disk- you could get the 512MB test down to 60s and the 1.5G test into the 180's.

For your rig, I would set the 1.5G test interval to 45s to maximize the fastest score potential.

Your results pretty much jive with my feeling that a 4-drive scratch-dedicated stripe is about the best cost/benefit ratio- maybe 6 drives if one is constantly working files that grow to over 2G... :xphat:

kaye
03-15-2007, 10:23 AM
Thanks Boots, I'm in the process of doing that, tho I am using one of my external 2-bay cases with 2xRap150 drives, the 2 with the oldest drive firmware (the other 4 for the Pshop scratch RAID are all the same later firmware). Will be Apple RAID 64K blocks. Was torn between default 32K and 64K. What do you use for block size? k

Boots
03-15-2007, 12:51 PM
32k on the system....

....also- did some tests recently with different block sizes for a dedicated RAID stripe (SoftRAID 3.6); turned out 32k was slightly faster than 64k or 128k despite convential wisdom of 128k being the appropriate setting to use for Photoshop....

TZ
03-15-2007, 02:14 PM
Boots,

that is what I have been reading (and being told!) lately. Makes me think there have been changes under the hood.

Will want to see new QB4 test results. From what I read, Barefeats only mentioned using it for small file sizes, not whether 10-100MB file tests will change or not.

Boots
03-15-2007, 04:48 PM
I remember testing with a very early beta precursor of PshopTest (it wasn't even called that yet and the test routine was substantially different) and SoftRAID 3.0.3 back in September 2004. This test used a G5 Dual 2GHz (Original), Photoshop 7, and 879.9MB test file. Below is the chart of results I made at the time- and you can see why I concluded 128k was the fastest...

But looking back, I really don't fully trust my test results from that era because a) we discovered later the Photoshop timer didn't reflect the total time involved in an operation, and b) my testing procedure and routine was not nearly as refined as it is today.

But nevertheless, it is interesting to consider these results.

The newer tests I refered to in the post above this one used PshopTest2.1 and Photoshop CS (v8):

I wanted to test the conventional wisdom of using a larger stripe block size on a RAID0 array of multiple drives used only for Photoshop's scratch disk. I tested with Photoshop CS because of it's exact repeatability in test scores- the (Tiger) OSX vm buffering pause bug seems to have very little effect on PshopTest2.1 scores using Photoshop 7 or CS- obviously CS2 and CS3 are a different story.

I also used a large test file size to constantly work the scratch disk.

Same test bed as above: G5 Quad 2.5Ghz, 16GB RAM (8 x 2GB matched modules), nVidia 7800GTX 512MB GPU, 2 x 150GB WD Raptors striped as a single 279.2GB Mac OS 10.4.8 startup volume with 35.2GB used and 244.1GB available.

1.5GB testfile, 70% memory (1532-34MB available,1072- 1074MB used by Photoshop, cache 4, History states 20, scratch on 4 x Atlas 15K II RAID0 as a single 136.5GB volume with 47MB used and 136.4GB available. The volume was formatted with SoftRAID 3.6 as Mac OS Extended, non-journaled. Spotlight indexing was disabled on the volume. 2 tests were run for each configuration:

32k block size: 485s, 485s
64k block size: 486s, 486s
128k block size: 486s, 486s

Boots
03-15-2007, 05:11 PM
Kaye, if you are inclined, it'd be real interesting if you could test the different block sizes on the striped scratch disk. I don't really know whether to take my newest results as a definitive indication that the block size makes very little difference, with 32k being slightly faster or maybe because I'm testing SCSI drives- that makes it somehow different.

I remember in OS9, changing the mode pages on my SCSI drives using FWB Hard Disk Tools (?) and how that seemed to speed up Photoshop scratch, and there were even notes in the user's guide to that effect. Course, I didn't have any standard test to really see if it had an effect.

---

I sure think it would be good for me to own a PC- eee-ghadz- where I could turn the 3GB SATA II switch on for my E7K500 drives and futz with the mode pages of my SCSI drives and otherwise have access to the deep-level drive stuff that we Mac users are sorely lacking....

kaye
03-15-2007, 08:11 PM
Pshop Test 2.1, Cache Levels 6, G5, 10.4.9, boot volume 2xRap150 Apple RAID 32KB blocks or 64KB blocks, did not make even 1sec difference, PS scratch/file 4xRap150, SoftRAID 3.6, 128KB SU. Photoshop CS2 v9.0.2.

512MB file, Memory 70% 2150MB, 30sec interval, 4xRaptor150, 64sec
512MB file, Memory 100% 3072MB, 30sec interval, 4xRaptor150, 67sec
Previous 2.1 test with primary boot volumne and 6xRaptor150, 64sec
------------
1.5GB file, Memory 100% 3072MB, 45sec interval, 4xRaptor150, 198sec
Previous 2.1 test with primary boot volumne and 6xRaptor150, 196sec
------------
3.0GB file, Memory 100% 3072MB, 60sec interval, 4xRaptor150, 576sec
Previous 2.1 test with primary boot volumne and 6xRaptor150, 568sec
------------
No point in running the 6.0GB file test.
------------
Observations:
1. Seems as if this DP is at its max, however I configure it. Whatever gained by a faster boot drive (and this 2xRap150 RAID is about 2.5 times faster than my normal boot drive) is lost by the 2 less drives of the PS scratch. Appears that my horsepower is limiting. BTW, after cloning to the 2xRap150 RAID with SuperDuper, I trashed more than half of it and then used iDefrag to eliminate all the holes, one of them 60GB for X-Plane, pretty much clean and mean. Capacity 279GB, Used 73GB, 26%. Cloned to a FW RAID so that I could switch from 32KB to 64KB blocks with the 2xRap150 Apple RAID. Exercise in futility.

Nuts, I see you wanted the various sizes on the scratch disk. Maybe Saturday, got to be here the whole day since Da Boss is driving back from Thousand Oaks and I like to be here in case she has any trouble. k

Boots
03-15-2007, 10:02 PM
Well thank you Kaye- that's a lot of work..

Funny, I thought you might have been able to pick up some speed with that striped startup. Makes me think I should retest my setup to verify...

kaye
03-17-2007, 09:51 AM
Pshop Test 2.1, Cache Levels 6, G5, 10.4.9, boot volume 2xRap150 Apple RAID 64KB blocks, PS scratch/file 4xRap150 or 4x15K2, SoftRAID 3.6, 32KB, 64KB, or 128KB SU. Photoshop CS2 v9.0.2.

512MB file, Memory 70% 2150MB, 30sec, 4xRaptor150, 32KB SU, 64sec
512MB file, Memory 70% 2150MB, 30sec, 4xRaptor150, 64KB SU, 64sec
512MB file, Memory 70% 2150MB, 30sec, 4xRaptor150, 128KB SU 64sec
512MB file, Memory 70% 2150MB, 30sec, 4xAtlas15K2, 32KB SU, 64sec
512MB file, Memory 70% 2150MB, 30sec, 4xAtlas15K2, 64KB SU, 64sec
512MB file, Memory 70% 2150MB, 30sec, 4xAtlas15K2, 128KB SU 65sec
------------
1.5GB file, Memory 100% 3072MB, 45sec, 4xRaptor150, 32KB SU, 197sec
1.5GB file, Memory 100% 3072MB, 45sec, 4xRaptor150, 64KB SU, 197sec
1.5GB file, Memory 100% 3072MB, 45sec, 4xRaptor150, 128KB SU 198sec
1.5GB file, Memory 100% 3072MB, 45sec, 4xAtlas15K2, 32KB SU, 202sec
1.5GB file, Memory 100% 3072MB, 45sec, 4xAtlas15K2, 64KB SU, 201sec
1.5GB file, Memory 100% 3072MB, 45sec, 4xAtlas15K2, 128KB SU 201sec
------------
3.0GB file, Memory 100% 3072MB, 60sec, 4xRaptor150, 32KB SU, 577sec
3.0GB file, Memory 100% 3072MB, 60sec, 4xRaptor150, 64KB SU, 577sec
3.0GB file, Memory 100% 3072MB, 60sec, 4xRaptor150, 128KB SU 576sec
3.0GB file, Memory 100% 3072MB, 60sec, 4xAtlas15K2, 32KB SU, 588sec
3.0GB file, Memory 100% 3072MB, 60sec, 4xAtlas15K2, 64KB SU, 583sec
3.0GB file, Memory 100% 3072MB, 60sec, 4xAtlas15K2, 128KB SU 582sec

Disappointing that I can't get any kind of speed bump. Consistency achieved within each file size by defraging with iDefrag after installing the file on the scratch RAID then rebooting to run the test. The boot RAID does not seem to help me. k

TZ
03-17-2007, 10:07 AM
Looks like ammo for why you need a Mac Pro :)
(after Leopard and CS3 only!)

Boots
03-17-2007, 11:38 AM
Kaye, 3 items stick out:

1) that your 2 x 150G Raptor striped startup is not on the built-in internal SATA bus but should be- I think it will be faster that way, too much contention with everything on PCI.

2) the 64k block size for it; I think TZ tested and showed 64k was slower as a startup disk...32k was faster.

3) test file on scratch disk- the 512MB will be faster with test file on striped startup, the larger test files, I'm not sure.

The striped startup should be faster. I knocked off 6s with the test file on the internal SATA striped startup; scratch on external 4 x stripe.

TZ
03-17-2007, 11:46 AM
The Quad has better SATA support. Kaye can't use those Raptors on his internal bus.

Booting from and using 4-6 drive RAID for both startup AND scratch or something, with test file present, of course? :D

Boots
03-17-2007, 01:49 PM
Kaye can't use those Raptors on his internal bus.

I thought that was only a problem with humungous single video files. :confused:

kaye
03-17-2007, 04:21 PM
As i understand it, those Raptors on the internal bus may not operate correctly as TZ says. Most likely not a problem at first but later on. Don't want to take a chance with that. Plus with what I would want on the Raptor RAID would put me at 54% full. I found no difference between 32KB or 64KB on the 2xRap150 startup disk.

I can't imagine that 4 drives on one PCI-X card, 2 drives On the other PCI-X card, both on the 100MHz bus, would slow the system down. Should be way more bandwidth than 6 drives suck.

Its academic anyway because I get the same results with my normal 500GB boot drive and 6xRap150 RAID. k

TZ
03-17-2007, 04:26 PM
Its academic anyway because I get the same results with my normal 500GB boot drive and 6xRap150 RAID. k

Boots
03-18-2007, 10:25 AM
Kaye, thanks for going through all this- at least we know what's what with the G5 Dual 2.7GHz/PS CS2/stripe or non-stripe startup. Sorry it didn't yield any benefit.

------

I decided to verify my previous results regarding the startup disk as a striped volume or a standard volume with Photoshop CS2 (9.0.2).

First I broke the 2 x 150G Raptor striped startup volume apart, cloned (from backup drive) back all files- except an image library- to a single standard DU volume 150G Raptor, then put the image library on the second 150G Raptor as a single standard DU volume, then repaired and optimized everything:

Test platform: G5 Quad, 16GB RAM, nVidia 7800GTX 512MB GPU, Mac OS 10.4.9

Two internal drives on built-in SATA bus, both 150G Raptors and both formatted as single 139.6GB volumes with Disk Utility:

"Studio": startup with 118.4GB available, 21.26GB used.
"Images": 117GB available, 22GB used.

External RAID Array (ERA) scratch disk: 4 x Atlas 15K II SCSI drives formatted with SoftRAID 3.6 as a single 136.5GB RAID0 volume optimized as "Server" with 32k block size, 136.4GB available and 46.6MB used before use as scratch disk. These drives are connected to a UL5D in slot #4 using ATTO ExpressPCI4 4.0.0 driver and ATTO Configuration Tool 3.07 with default settings.

------

I tested with PshopTest 2.1 and used a 30s pause for the 512MB testfile and a 45s pause for the 1.5GB testfile. Photoshop CS2 had the Bigger Tiles plugin enabled.

9 test runs with the512MB testfile:

1. 512MB TF, scratch on startup, testfile on Desktop, 70% (2150MB) memory, cache 6: 69s
2. 512MB TF, scratch on startup, testfile on Desktop, 70% (2150MB) memory, cache 6: 75s
3. 512MB TF, scratch on startup, testfile on Desktop, 70% (2150MB) memory, cache 6: 58s
4. 512MB TF, scratch on startup, testfile on Desktop, 70% (2150MB) memory, cache 6: 81s
5. 512MB TF, scratch on startup, testfile on Desktop, 70% (2150MB) memory, cache 6: 79s
6. 512MB TF, scratch on startup, testfile on 2nd Raptor, 70% (2150MB) memory, cache 6: 58s
7. 512MB TF, scratch on ERA, testfile on 2nd Raptor, 70% (2150MB) memory, cache 6: 57s
8. 512MB TF, scratch on ERA, testfile on 2nd Raptor, 70% (2150MB) memory, cache 6: 58s
9. 512MB TF, scratch on ERA, testfile on startup, 70% (2150MB) memory, cache 6: 58s

2 test runs with the 1.5GB testfile:

10. 1.5GB TF, scratch on ERA, testfile on 2nd Raptor, 100% (3072MB) memory, cache 6: 177s
11. 1.5GB TF, scratch on ERA, testfile on 2nd Raptor, 100% (3072MB) memory, cache 6: 176s

------

This all jives with my previous scores with the same config almost exactly.

Here are my best scores accomplished with the 150G Raptors as a single 279GB striped volume. These were achieved in late January with 10.4.8:

512MB TF: scratch on ERA, testfile on startup, 70% (2150MB) memory, cache 6: 53s

1.5GB TF, scratch on ERA, testfile on startup, 100% (3072MB) memory, cache 6: 166s

So the striped startup gain for me is 4s with the 512MB testfile, and 10s with the 1.5GB testfile.

------

When I do eventually get an MP, I expect the fastest PS setup will be a multiple-drive striped startup coupled with a fast external array....

My current fantasy Photoshop dream machine: 3GHz Mac Pro Quad (8 cores), 32GB RAM, 4 x 150G SATA II Raptor stripe + 4x Fujitsu SAS (150MB p/s each) as scratch. Ha!

kaye
03-18-2007, 12:38 PM
Renewed energy Boots after fixing my UPS problem yesterday by downloading and running the 10.4.9 Combo update PPC. That was my last resort, no utility would fix it nor downloading and applying the APC software again.

Deleted the 10.4.9 boot 2xRap150 RAID with 64KB blocks then created a new Apple RAID with 32KB blocks. SuperDuper cloned my FW800 RAID with 10.4.9 to the 2xRap150 RAID with 32KB blocks. Checked it all out with utilities then shutdown so that I could shut off the FW800 RAID.

Pshop Test 2.1, Cache Levels 6, G5, 10.4.9, boot volume 2xRap150 Apple RAID 32KB blocks, PS scratch only 4xRap150, SoftRAID 3.6, 32KB or 64KB SU. Photoshop CS2 v9.0.2. File now on boot 2xRap150 RAID, not on scratch RAID.

512MB file, Memory 70% 2150MB, 30sec, 4xRaptor150, 32KB SU, 65sec
512MB file, Memory 70% 2150MB, 30sec, 4xRaptor150, 64KB SU, 65sec
Previous best 64sec
------------
1.5GB file, Memory 100% 3072MB, 45sec, 4xRaptor150, 32KB SU, 198sec
1.5GB file, Memory 100% 3072MB, 45sec, 4xRaptor150, 64KB SU, 200sec
Previous best 197sec
------------
3.0GB file, Memory 100% 3072MB, 60sec, 4xRaptor150, 32KB SU, 581sec
3.0GB file, Memory 100% 3072MB, 60sec, 4xRaptor150, 64KB SU, 584sec
Previous best 576sec

Still no speed bump. And still the boot RAID does not seem to help. k

Boots
03-19-2007, 07:44 AM
Thanks Kaye.
Well- the upside of all this anyway should be a healthy speed bump for PPC with PS CS3 when that arrives.

For me, with the above test specs (2 x Raptor stripe startup, 70% memory)-

the 512MB test file is 45s with CS3 Public Beta
the 512MB test file is 53s with CS2

For me, with the above test specs (2 x Raptor stripe startup, 100% memory)-

the 1.5GB test file is 146s with CS3 Public Beta
the 1.5GB test file is 166s with CS2 (best score)

About a 15% speed bump in terms of PshopTest2.1 anyway... :)

kaye
03-19-2007, 08:15 AM
A 15% speed bump in PshopTest2.1 would certainly be welcome. For instance, the 3.0GB test I did yesterday, to get 581sec result took me 22min 41sec. k

kaye
03-22-2007, 11:09 AM
Based on what I learned from the previous test plus Boots' previous test post and TZ's comments and links, I decided to dump the 2xRaptor150 boot RAID and go back to my Maxtor 500GB-16MB cache drive for boot and create once again a 6xRaptor150 striped RAID with 32KB SU for both file and scratch. This turned out to result in a new level of performance results for the Pshop Test. I also tested the 10 drive RAID.

Pshop Test 2.1, Cache Levels 6, G5, 10.4.9, 6xRaptor150 file and scratch, SoftRAID 3.6, 32KB SU. Photoshop CS2 v9.0.2.

512MB file, Memory 70% 2150MB, 30sec, 6xRaptor150, 32KB SU, 64sec
Previous 2.1 test with 4xRaptor150 and 2xRaptor boot, both 32KB SU, 65sec
------------
1.5GB file, Memory 100% 3072MB, 45sec, 6xRaptor150, 32KB SU, 195sec
1.5GB file, Memory 100% 3072MB, 45sec, 10 drives, 32KB SU, 196sec
Previous 2.1 test with 4xRaptor150 and 2xRaptor boot, both 32KB SU, 198sec
------------
3.0GB file, Memory 100% 3072MB, 60sec, 6xRaptor150, 32KB SU, 561sec
3.0GB file, Memory 100% 3072MB, 60sec, 10 drives, 32KB SU, 565sec
Previous 2.1 test with 4xRaptor150 and 2xRaptor boot, both 32KB SU, 581sec
------------
6.0GB file, Memory 100% 3072MB, 60sec, 6xRaptor150, 32KB SU, 1730sec
6.0GB file, Memory 100% 3072MB, 60sec, 10 drives, 32KB SU, 1709sec
Previous 2.1 best, 10 drives, 128KB SU, 1780sec

The 10 drive RAID does not provide a speed bump until somewhere between 3.0GB and 6.0GB file size but closer to 3.0GB. The 6xRaptor150 file and scratch really shows a bump for 3.0GB and 6.0 file sizes and this RAID is enough for me without the 10 drives.

Out of curiosity I ran the 10 drive RAID at 32KB SU in QB3.1 to compare it with 128KB SU previously tested.

100MB, 745MB/s read - 775 write - 128KB SU
100MB, 641MB/s read - 787 write - 32KB SU k

Boots
03-24-2007, 04:47 PM
Nice Kaye- looks like you've eeked out pretty much the max acceleration. :xphat:

I discovered something nice today I hadn't known about: relocating Bridge's cache and 'use centralized cache..' for both images and RAW to the Atlas RAID speeds up the building thumbnails process in a huge way, and it also speeds up the Bridge screen redraw of folders with a large number of images after the thumbs are built. Sweet. :)

kaye
03-27-2007, 06:39 AM
I didn't know that Boots but I see how it's done. Even tho Bridge is of course on my G5, I have not used it since it first came out and was very buggy. I can see where your work would benefit greatly but my real work with Photoshop CS2 and ImageReady CS2 don't call for using Bridge as far as I can tell, unless you feel differently about that or can tell me how I would benefit. Always open to your expertise. k

Boots
03-27-2007, 11:22 AM
I use Bridge as my primary "digital asset manager". It's pretty handy when you've got multiple versions of images, and need to flag stuff to compare and create collections of images for web etc.

Since my image archive is big (20G) and organized in numerous subfolders, Bridge makes it a snap to quickly find stuff.

For CS2, it's up to 1.0.4 and is pretty stable on PPC at this point; 10.4.9 gave it a speed boost too.

kaye
03-29-2007, 06:42 AM
OK, I'll give it a go next week after reading up on it Boots, thanks. Got my hands full this week with my contractor and a series of overdue house repairs and upgrades the wife wants. k

kaye
04-08-2007, 02:50 PM
More learning. Added 2 more Raptor150 drives. This turned out to result in a new level of performance results for the large file (3.0GB and 6.0GB) Pshop Test. And because of the volume size,1.090TB, once running out of real RAM (8GB), plenty more breathing room for virtual RAM and Cache Levels 6 on the 8xRaptor150's. Adding the 4x15K2 drives for a 12 drive RAID hurt by comparison because the volume size reduced to 409.4GB.

Pshop Test 2.1, Cache Levels 6, G5-2.7GHz DP, 8GB RAM, 10.4.9, 8xRaptor150 file and scratch, volume size 1.090 Terabytes, SoftRAID 3.6, 32KB SU. Photoshop CS2 v9.0.2 on boot Maxtor Pro 500GB-16MB cache drive.

512MB file, Memory 70% 2150MB, 30sec, 8xRaptor150, 32KB SU, 64sec
512MB file, Memory 100% 3072MB, 30sec, 8xRaptor150, 32KB SU, 66sec
512MB file, Memory 70% 2150MB, 30sec, 6xRaptor150, 32KB SU, 64sec
Previous 2.1 test with 4xRaptor150 and 2xRaptor boot, both 32KB SU, 65sec
------------
1.5GB file, Memory 100% 3072MB, 45sec, 8xRaptor150, 32KB SU, 195sec
1.5GB file, Memory 100% 3072MB, 45sec, 6xRaptor150, 32KB SU, 195sec
1.5GB file, Memory 100% 3072MB, 45sec, 10 drives, 32KB SU, 196sec
Previous 2.1 test with 4xRaptor150 and 2xRaptor boot, both 32KB SU, 198sec
------------
3.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 551sec
3.0GB file, Memory 100% 3072MB, 60sec, 6xRaptor150, 32KB SU, 561sec
3.0GB file, Memory 100% 3072MB, 60sec, 12 drives, 32KB SU, 565sec
3.0GB file, Memory 100% 3072MB, 60sec, 10 drives, 32KB SU, 565sec
Previous 2.1 test with 4xRaptor150 and 2xRaptor boot, both 32KB SU, 581sec
------------
6.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 1609sec
6.0GB file, Memory 100% 3072MB, 60sec, 6xRaptor150, 32KB SU, 1730sec
6.0GB file, Memory 100% 3072MB, 60sec, 12 drives, 32KB SU, 1708sec
6.0GB file, Memory 100% 3072MB, 60sec, 10 drives, 32KB SU, 1709sec
Previous 2.1 best, 10 drives, 128KB SU, 1780sec

k

kaye
04-19-2007, 02:42 PM
Hardware Overview:
System Version 10.4.9 (8P135)
Machine Name: Power Mac G4
Machine Model: PowerMac3,6
CPU Type: PowerPC G4 (3.3)
Number Of CPUs: 2
CPU Speed: 1.42 GHz
Version: 7455 (Apollo) v3.3
L1 Cache 32K (instruction), 32K (data)
L2 Cache 256K @ 1.42 GHz
L3 Cache 2048K
Memory: 2 GB (2048MB)
Bus Speed: 167 MHz
Boot ROM Version: 4.4.8f2

Pshop Test 2.1, Cache Levels 6, Giga G4-1.42GHz DP, 2GB RAM (2x MacGurus 1GB CL2.5 sticks in memory slots J21 and J22, J23 and J20 empty), 10.4.9, SeriTek/1VE4 PCI card drives external 4xRaptor150 file and scratch, volume size 558GB, SoftRAID 3.6, 32KB SU. Photoshop CS2 v9.0.2 on boot Maxtor DiamondMax 10 300GB-16MB cache PATA drive on Acard 6885M 4-channel PCI card. ATI Radeon 9800 Pro ME 256MB VRAM, 23" Cinema HD Display, 1920x1200, Millions. Default Photoshop CS2 Plugins plus Bigger Tiles enabled.

512MB file, Memory 70% 1374MB, 60sec, 4xRaptor150, 32KB SU, 125sec
------------
1.0GB file, Memory 70% 1374MB, 60sec, 4xRaptor150, 32KB SU, 502sec
------------
Frankly I was surprised at how responsive the MDD was. I could have easily used 30sec pauses for the 512MB file and 45sec for the 1.0GB file. k

kaye
04-25-2007, 10:12 AM
Learning again thanks to Boots plus added new file size tests. RAID now striped across a SeriTek/1VE4 in the 133MHz bus slot and a SeriTek/1VE4 in a 100MHz slot.

Pshop Test 2.1, Cache Levels 6, G5-2.7GHz DP, 8GB RAM, 10.4.9, 8xRaptor150 file and scratch, volume size 1.090 Terabytes, SoftRAID 3.6, 32KB SU. Photoshop CS2 v9.0.2 on boot Maxtor Pro 500GB-16MB cache drive.

512MB file, Memory 70% 2150MB, 30sec, 8xRaptor150, 32KB SU, 64sec
Previous best 64sec, no change for this file size
------------
1.0GB file, Memory 70% 2150MB, 45sec, 8xRaptor150, 32KB SU, 135sec
1.0GB file, Memory 100% 3072MB, 45sec, 8xRaptor150, 32KB SU, 117sec
New file size tested on this G5.
------------
1.5GB file, Memory 100% 3072MB, 45sec, 8xRaptor150, 32KB SU, 194sec
Previous best 195sec
------------
3.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 551sec
Previous best 551sec
------------
6.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 1595sec
Previous best 1609sec

Awaiting arrival of Photoshop CS3.
k

Boots
04-25-2007, 04:41 PM
Kaye, when you receive CS3- in addition to enabling the Bigger Tiles plugin, you'll want to test using perhaps two additional performance plugins.

I don't know whether Adobe has left VM buffering disabled or enabled it like with CS2. If disabled by default, there should be a plugin called something like "Enable VM Buffering". If not, it's probably already enabled by default. In any case you'll be able to tell pretty quickly when testing a large file size.

The other one should be especially useful with your monster fast scratch disk-
it's a plugin called "Disable Scratch Compression".

I'm not certain if the final release contains these plugins on the cd/download, or whether you'll need to download them separately.

kaye
04-25-2007, 05:57 PM
Boots, I'll look for all of that. k

kaye
04-30-2007, 11:34 PM
Boots,

I enabled Bigger Tiles and found Disable Scratch Compression on the DVD and installed it. Enable VM Buffering I could not find. Force VM Compression I found on the DVD but did NOT install it. Did I do right? k

Boots
05-01-2007, 03:45 AM
That's right; Bigger Tiles you want to enable, Disable Scratch Compression you'll want to enable when- and probably only when- using your wicked fast scratch disk, Force VM Compression if enabled will most likely slow you down.

I still don't know if VM buffering is enabled by default....if you find a Disable VM Buffering plugin then it is enabled by defualt...take a look in the Applications/Adobe Photoshop CS3/Extensions folder after installing; you might find it or one or the other of those there.

kaye
05-02-2007, 09:54 AM
I looked for and installed everything I could find at Adobe's website and on the DVD. No VM Buffering Enable or Disable. Plus found this:
http://www.adobe.com/cfusion/knowledgebase/index.cfm?id=kb401089

Decided to plunge in and try the Pshop Test. Default History States 20, Cache Levels 6. Results not good but I have a clue why.

8xRap150sr32KBsu and 64KBsu tested.
3.0GB, 60sec, 100%-3072MB, 893sec (64KBsu)
3.0GB, 60sec, 100%-3072MB, 895sec (32KBsu)
3.0GB, 60sec, 100%-3072MB, 895sec (tested again 32KBsu)

Previous best CS2, 551sec.

It appears that some of the steps (except Saves) don't start the 60sec pause when the redraw is complete. Instead they appear to start the 60sec pause after ALL drive/volume activity is complete.

That does add a lot of time to the test time. For instance, the Convert to Profile and Convert to Mode each add about 57sec of volume activity after the redraw before the 60sec pause. Some of the other Steps less, some 57sec. I don't have enough eyes and timers to see what is going on. Maybe you see a pattern Boots. If not, I will go back to default setup with only Bigger Tiles added and work up from there. k

TZ
05-02-2007, 10:20 AM
From your link:
Allocating Memory above 2 GB with 64-bit Processors

Photoshop can access up to 8 GB of RAM.
You can see the actual amount of RAM Photoshop can use in the Let Photoshop Use number when you set the Let Photoshop Use slider in the Performance preference to 100%.

The RAM above the 100% used by Photoshop, which is from approximately 3 GB to 3.7 GB, can be used directly by Photoshop plug-ins (some plug-ins need large chunks of contiguous RAM), filters, and actions.

If you have more than 4 GB (to 8 GB), the RAM above 4 GB is used by the operating system as a cache for the Photoshop scratch disk data. Data that previously was written directly to the hard disk by Photoshop is now cached in this high RAM before being written to the hard disk by the operating system.

Sounds like VM buffering is enabled by default.

kaye
05-02-2007, 11:31 AM
OK, thanks TZ. I was wondering about that. Apparently CS3 is built in whereas the beta was not. k

Boots
05-02-2007, 11:37 AM
Kaye, I'm not sure what to say about the pause thing. I didn't see any inconsistencies or differences from CS2 when I tested CS3 Public beta and various beta builds...
I don't yet own the PS CS3 final release so I can't test it yet.

TZ, I've already alerted the powers that be over at Adobe Photoshop about that paragraph you quoted above which is downright inaccurate. (There's no '8GB limit')
Yes, it would imply that buffering is enabled by default.

Kaye, can you verify viewing Activity Monitor while testing? "Inactive:" (memory) should bring your total used memory to 8GB when running the 1.5GB testfile..if buffering is enabled by default. Otherwise, ("Inactive:") will reach 4GB max.

kaye
05-02-2007, 12:36 PM
Boots, I will try that. Got to go get a haircut first. k

kaye
05-02-2007, 05:55 PM
Ran the 3.0GB test file again with Activity Monitor set to System Memory.

Used varied from 3.63GB up to 7.69GB.
Inactive varied from 400MB up to 2.66GB.

At end of test before quit:
Used 5.25GB
Inactive 2.43GB
So looks like VM buffering is enabled by default.

Now on to the other observation of mine, the Pauses (60sec set):

1. Completion of Open to start of Rotate - 63sec
2. Completion of Rotate to start of Apply Image - 59sec
3. Completion of Apply Image to start of Save - 60sec
4. Completion of Save to start of Canvas Size - 60sec
5. Completion of Canvas Size to start of Levels - 60sec
6. Completion of Levels to start of Gaussian Blur - 90sec
7. Completion of Gaussian Blur to start of Sharpen - 92sec
8. Completion of Sharpen to start of Save - 93sec
9. Completion of Save to start of Convert to Profile - 60sec
10.Completion of Convert to Profile to start of Save - 104sec
11.Completion of Save to start of Convert Mode - 60sec
12.Completion of Convert Mode to start of Save - 91sec
13.Completion of Save to start of Stop - 60sec

Looking at the above, pretty obvious that the steps are taking longer to complete because the Pauses are longer than 60sec.

Think I should trash Prefs and start over with just the default plugins plus Bigger Tiles and test from there. k

Boots
05-02-2007, 08:23 PM
Something's not right here. From your figures, I would say VM buffering is not enabled; "Inacative" should climb up to over 7GB during the test.

I think they forgot to put the plugins (Enable VM Buffering and Disable VM Buffering) on the installer dvd:

http://www.adobeforums.com/cgi-bin/webx?128@@.3bc3b4ed

Adam Jerugim is part of the Photoshop team at Adobe. I'll see if I can get some more info.

TZ
05-03-2007, 06:43 AM
ForceVMBuffering plug-in

It should be in the Extensions folder in the Optional Plug-ins folder. If it isn't, then we'll need to post it so people can download it. I'm checking various media now...

We're working on getting them available to users ASAP. Adam

Boots
05-03-2007, 08:59 AM
I checked with AJ this morning and confirmed that indeed, those plugins were not included on the installer dvd but should have been. They're working on making them available...

kaye
05-03-2007, 10:42 AM
Boots, that is a relief. I decided to wait on further CS3 testing until I heard more. Those plugs should be available anytime now.

I have been doing more CS2 testing in the meantime. One thing I noticed on the CS3 DVD is that the read me for the Disable Scratch Compression plugin is actually for CS2 (identifys it as CS2) so the location for that plug is not exactly correct. But it set me to thinking, there must be such a plugin on the CS2 DVD. Sure enough it is there. Installed it and ran some tests for 1.5GB, 3.0GB, and 6.0GB file sizes. All faster. Will post later, dog sitting today. k

kaye
05-03-2007, 05:19 PM
CS2 with default plugs plus Bigger Tiles and Disable Scratch Compression ENABLED.

Pshop Test 2.1, History States 20, Cache Levels 6, G5-2.7GHz DP, 8GB RAM, 10.4.9, 8xRaptor150 file and scratch, volume size 1.090 Terabytes, SoftRAID 3.6, 32KB SU. Photoshop CS2 v9.0.2 on boot Maxtor Pro 500GB-16MB cache drive.

------------
1.5GB file, Memory 100% 3072MB, 45sec, 8xRaptor150, 32KB SU, 189sec
Previous best 194sec
------------
3.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 481sec
Previous best 551sec
------------
6.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 1542sec
Previous best 1595sec
------------
Only change to these tests is Disable Scratch Compression plug enabled. k

Boots
05-04-2007, 05:31 AM
Nice Kaye.
I found on my setup- with DisableScratchCompression enabled- I had faster results when the testfile was on my striped boot, rather than on the scratch disk...

Pretty impressive plugin though.

kaye
05-09-2007, 03:24 PM
I joined that Adobe Forums thread today, previously linked to, and AJ said he has nothing to do with getting the download online but that it should be out soon. k

Boots
05-09-2007, 03:45 PM
Nice Kaye. Some of the regulars on that forum can be pretty nasty and argumentative; it seems they enjoy pushing people away...
I post there very infrequently, but do end up learning a lot in the long run.

kaye
05-09-2007, 05:01 PM
Exactly Boots. I gave it the old kill 'em with kindness routine annd stuck to my guns. Then AJ responded and they said not a word after that. k

kaye
05-14-2007, 09:16 AM
I opened an Adobe web based support case asking for the Photoshop CS3 ForceVMBuffering plugin sent via email. I should hear within 24 hours. k

kaye
05-21-2007, 03:17 PM
It has taken awhile but after going thru Adobe Customer Service then thru Tech Support then engineering I think, just got a phone call from Tech Support that the ForceVMBuffering plug is available online for PS CS3:
http://www.adobe.com/support/downloads/detail.jsp?ftpID=3623

Not sure if you have to be logged in or not. k

kaye
05-22-2007, 04:19 PM
Comparing CS3 to CS2.

CS2 plugs:
Default, plus DisableScratchCompress.plugin and Bigger Tiles.plugin.

CS3 plugs:
Default, plus DisableScratchCompress.plugin, Bigger Tiles.plugin, and ForceVMBuffering.plugin.

CS3 now flies.
Pshop Test 2.1, History States 20, Cache Levels 6, G5-2.7GHz DP, 8GB RAM, 10.4.9, 8xRaptor150 file and scratch, volume size 1.090 Terabytes, SoftRAID 3.6, 32KB SU. Photoshop CS2 v9.0.2 and CS3 v10.0 on boot Maxtor Pro 500GB-16MB cache drive.

------------
CS3, 1.5GB file, Memory 100% 3072MB, 45sec, 8xRaptor150, 32KB SU, 183sec
Previous best CS2 189sec
------------
CS3, 3.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 455sec
Previous best CS2 481sec
------------
CS3, 6.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 1477sec
Previous best CS2 1542sec
------------
k

kaye
09-11-2007, 05:29 PM
Been awhile. Several changes:
1. Pshop Test 2.2
2. OSX 10.4.10
3. SoftRAID 3.6.3
4. AGP ATI X800XT ME updates
5. Backup boot Maxtor drive larger size and same as boot drive

CS3 plugs:
Default, plus:
1. DisableScratchCompress.plugin
2. Bigger Tiles.plugin
3. ForceVMBuffering.plugin.

Pshop Test 2.2, History States 20, Cache Levels 6, G5-2.7GHz DP, 8GB RAM, OSX 10.4.10, 8xRaptor150 file and scratch, volume size 1.090 Terabytes, SoftRAID 3.6.3, 32KB SU. Photoshop CS3 v10.0 on boot Maxtor Pro 500GB-16MB cache drive. Backup Boot Maxtor Pro 500GB-16MB cache drive.
Scratch Disks order:
1. 8xRap150, 1.090 TB Free Space
2. Backup boot, 304.15GB Free Space
3. Boot, 303.48GB Free Space

------------
CS3, 512MB file, Memory 70% 2150MB, 30sec, 8xRaptor150, 32KB SU, 61sec
------------
CS3, 1.5GB file, Memory 100% 3072MB, 45sec, 8xRaptor150, 32KB SU, 186sec
------------
CS3, 3.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 463sec
------------
CS3, 6.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 1546sec
------------
k

kaye
09-18-2007, 06:55 PM
More changes:
1. SoftRAID 3.6.4
2. Previous tests post, forgot to remove Photoshop prefs to get default prefs, a big DUH!

Otherwise, everything the same as previous post.
CS3 plugs:
Default, plus:
1. DisableScratchCompress.plugin
2. Bigger Tiles.plugin
3. ForceVMBuffering.plugin.

Pshop Test 2.2, History States 20, Cache Levels 6, G5-2.7GHz DP, 8GB RAM, OSX 10.4.10, 8xRaptor150 file and scratch, volume size 1.090 Terabytes, SoftRAID 3.6.4, 32KB SU. Photoshop CS3 v10.0 on boot Maxtor Pro 500GB-16MB cache drive. Backup Boot Maxtor Pro 500GB-16MB cache drive.
Scratch Disks order:
1. 8xRap150, 1.090 TB Free Space
2. Backup boot, 304.15GB Free Space
3. Boot, 303.48GB Free Space

------------
CS3, 512MB file, Memory 70% 2150MB, 30sec, 8xRaptor150, 32KB SU, 59sec
2sec faster
------------
CS3, 1.5GB file, Memory 100% 3072MB, 45sec, 8xRaptor150, 32KB SU, 186sec
same as previous
------------
CS3, 3.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 461sec
2sec faster
------------
CS3, 6.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 1541sec
5sec faster
------------
k

Boots
09-19-2007, 07:13 AM
Nice to see SR 3.6.4 speeding up the scratch disk a bit. :)

I've been working a lot recently with RAW files using Bridge CS3 and the latest Adobe Camara Raw plugin.

That plugin requires a lot of processing power and dedicated RAM to run not slowly.

Bridge cache size and performance has developed as a not-inconsequential performance issue. You can direct the Bridge cache to be stored where you want and also control (somewhat) it's size- if you have a lot of images, a seriouly large cache can result. I have a "measly" 27GB or so of images; the cache is about 1.2GB.

Both of these are solid arguments for lotsof installed RAM and big fast drives + fast multiple processors.
Been doing a bunch of this work on my girl's iMacG5 single 2GHz, 2GB RAM, 250GB 16MB SATA Maxtor. It's everything we can do to tweak performance to get it to run faster dealing with 8 megapixel 30D RAW images.

kaye
09-19-2007, 07:33 AM
So how big are those 8 megapixel 30D RAW images that she works with Boots? Must bring her iMac G5 to a stand still. k

Boots
09-19-2007, 07:47 AM
They're not too bad- <10MB in raw form, some lighting adjustments made, then brought to Photoshop for a crop, then a small save for web (whose dialog also needs a lot of computing power).

I'll have to check to see what size if they are saved out as psd files without cropping...

I forgot to mention the Camera Raw plugin also has a cache feature which can be adjusted. Haven't fooled around with that yet- but I'm sure it'll make a difference one way or the other.

Also- getting Bridge to open faster is a question of turning off all or most of the various default startup scripts that activate when Bridge is launched: Bridge preferences>Startup scripts.

kaye
11-19-2007, 05:19 PM
Changes comparison:
Photoshop CS3 v10.0 on Backup Boot Maxtor Pro 500GB-16MB cache drive. Photoshop CS3 v10.0.1 on Boot Maxtor Pro 500GB-16MB cache drive. Apple Disk Utility v10.5.6 (Tiger 10.4.10), all drives

CS3 plugs both versions:
Default, plus:
1. DisableScratchCompress.plugin v10.0
2. Bigger Tiles.plugin v10.0
3. ForceVMBuffering.plugin v10.0

Pshop Test 2.2, History States 20, Cache Levels 6, G5-2.7GHz DP, 8GB RAM, OSX 10.4.10, 1xCuda-32MBcache file and scratch (MGBurly1HS LCD external case), volume size 931.39GB, Apple Disk Utility. Photoshop CS3 v10.0 on Backup Boot Maxtor Pro 500GB-16MB cache drive. CS3 v10.0.1 on Boot Maxtor Pro 500GB-16MB cache drive. Most Menu Items disabled, DSL connected/running.

Scratch Disks order v10.0.1 on Boot Drive:
1. 1xCuda-32MBcache(7200.11, 1TB), 930.74GB Free Space
2. Backup Boot, 299.90GB Free Space
3. Boot, 299.64GB Free Space

Scratch Disks order v10.0 on Backup Boot Drive:
1. 1xCuda-32MBcache(7200.11, 1TB), 930.74GB Free Space
2. Boot, 299.88GB Free Space
3. Backup Boot, 299.65GB Free Space

------------
CS3-10.0, 512MB file, Memory 70% 2150MB, 60sec, 1xCuda, 69sec
CS3-10.0, 512MB file, Memory 100% 2150MB, 60sec, 1xCuda, 72sec
CS3-10.0.1,512MB file, Memory 70% 2150MB, 60sec, 1xCuda, 67sec
CS3-10.0.1,512MB file, Memory 100% 2150MB,60sec, 1xCuda, 70sec
Previous best:
CS3-10.0, 512MB file, Memory 70% 2150MB, 30sec, 8xRaptor150, 32KB SU, 59sec
8sec faster
------------
CS3-10.0, 1.5GB file, Memory 100% 3072MB, 60sec, 1xCuda, 257sec
CS3-10.0.1,1.5GB file, Memory 100% 3072MB, 60sec, 1xCuda, 253sec
Previous best:
CS3-10.0, 1.5GB file, Memory 100% 3072MB, 45sec, 8xRaptor150, 32KB SU, 186sec
67sec faster
------------
CS3-10.0, 3.0GB file, Memory 100% 3072MB, 60sec, 1xCuda, 913sec
CS3-10.0.1,3.0GB file, Memory 100% 3072MB, 60sec, 1xCuda, 908sec
Previous best:
CS3-10.0, 3.0GB file, Memory 100% 3072MB, 60sec, 8xRaptor150, 32KB SU, 461sec
447sec or almost 2x faster
-------------------------
Conclusions:
1. Since I am limited to 8GB RAM with this G5, the size of the Cuda or 8xRaptor150 both at about 1TB is very helpful when RAM is full.
2. For file size at 1.5GB or less, the single Cuda does a fine job. 3.0GB or more it lags way behind the 8xRaptor150. 2xCudas in RAID0 would be killer. I have an extra LCD case but not the extra Cuda to go with it.

In the future will be Leopard tests, once more stable, will not test 10.4.11.
The Cuda inside the MGBurly1HS LCD external case never reached a temperature higher than 29.1C during these tests, superb. With drive idling, case power supply draws 18W AC. k

kaye
12-21-2007, 01:44 PM
My G5, I ran one Pshop 2.2 test in Leopard 10.5.1, the 6GB test. I left everything running, DSL, background apps such as menu items, etc. Booted to my Cuda1TB-32MB which is now 181.42GB used, 931.39GB capacity. 8xRap150 running via SoftRAID 3.6.4. Photoshop CS3 10.0.1 at default settings.

Scratch Disks order:
1. 8xRap150-16MB SoftRAID RAID0, 32KBsu, 1041.32GB Free Space
2. 1xCuda1TB, Apple DU, 749.96GB Free Space, 10.5.1, external boot drive
3. 7HMLIII-16MB 500GB, Apple DU, 296.77GB Free Space, 10.4.11, internal backup boot drive
4. 7HMLIII-16MB 500GB, Apple DU, 291.59GB Free Space, 10.4.11, internal boot drive

CS3-10.0.1, 6.0GB file, Memory 100% 3072MB, 60sec, 8xRap150, 1737sec
Previous best with Pshop 2.2 test in Tiger 1541sec
BareFeats' analysis is that Adobe has yet to optimize Pshop CS3 for Leopard and that it runs faster in Tiger. He ran 6 multi-processor actions. The Pshop test runs 13, tho several are Save which as I recall are single-proccessor. Can't remember for sure tho. Anyway, it does run slower in Leopard but it does run with no problems whatsoever, Pshop CS3, The test, and the SoftRAID stripe. k

Boots
12-21-2007, 02:36 PM
Thanks Kaye.
And yet, CS3 (and Leopard) supposedly optimized for Intel..hopefully it's just an adjustment for Adobe/Apple to gain back this apparent lost ground- at least on the PPC side of things.

kaye
10-23-2008, 09:25 AM
Macworld Review: Adobe Photoshop CS4 and CS4 Extended (http://www.macworld.com/article/136256/2008/10/pscs4.html?lsrc=rss_reviews) which is very interesting. Also a link on the page for someone still using v3.0. k

Mil6
09-17-2018, 01:09 PM
Thanks for posting the test results. AND for putting to bed the myth that you need Tiger along with CS2 to utilize the extra RAM.

ricks
09-17-2018, 08:12 PM
Thanks for posting the test results. AND for putting to bed the myth that you need Tiger along with CS2 to utilize the extra RAM.

I'm tempted to just delete this post. This is a 10 year old thread! It does not need to be resurrected with posts from Panama.