Page 8 of 17 FirstFirst 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 LastLast
Results 141 to 160 of 332

Thread: G5 2.7GHz

  1. #141
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I have changed my PCI-X card configuration for a couple of reasons. First, I wanted to remove a UL4D from the 100MHz bus to save it as a spare and second, I wanted to replace it with another SeriTek/1VE4 to run a WDC WD1500ADFD-00NLR0 Raptor1500-16MB cache.

    G5-2.7GHz DP, ATI X800XT ME, OS 10.4.4, UL4D in 133MHz slot, nothing connected, 2x SeriTek/1VE4 in 100MHz slots, one connected to 1x Raptor1500, 8GB RAM. The Raptor1500 is in an external case and I have been exercising it for the last couple of days to loosen it up.

    No way I can duplicate Boot's Quad and Raptor1500 results but not totally out of the ballpark. QB 2.1-100MB.
    Boots:
    100 MegaBytes, Read- 73.014 MB/sec, Write- 96.640 MB/sec

    Mine:
    Over the last few days, 100MB Reads as high as 77.455 MB/sec, Writes as high as 89.561 MB/sec, but on average this is what I get

    The drive does haul and I'm pleased. Next to SoftRAID it with 4x Raptor74s.

    Will get Boots and my results on the DB. k

  2. #142
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Yesterday older Raptor1500:
    WD Raptor1500sr
    Extended Test Size: 100 MB Read: 72.680 MB/sec Write: 89.427 MB/sec

    Yesterday, new Raptor1500:
    WD Raptor1500sr2
    Extended Test Size: 20 MB Read: 78.994 MB/sec Write: 121.833 MB/sec
    Extended Test Size: 30 MB Read: 65.110 MB/sec Write: 121.951 MB/sec
    Extended Test Size: 40 MB Read: 64.821 MB/sec Write: 116.995 MB/sec
    Extended Test Size: 50 MB Read: 64.520 MB/sec Write: 109.774 MB/sec
    Extended Test Size: 60 MB Read: 67.030 MB/sec Write: 102.252 MB/sec
    Extended Test Size: 70 MB Read: 69.007 MB/sec Write: 100.565 MB/sec
    Extended Test Size: 80 MB Read: 70.508 MB/sec Write: 97.353 MB/sec
    Extended Test Size: 90 MB Read: 71.835 MB/sec Write: 95.507 MB/sec
    Extended Test Size: 100 MB Read: 72.758 MB/sec Write: 91.925 MB/sec


    Both Raptors in SeriTek/1EN2 External Two-Bay Serial ATA Enclosure running off of same SeriTek/1VE4 100MHz bus. I chose the 2-bay so that I could stack my SATA 4-bay Burly under a SCSI 4-bay Burly with the 2-bay on top and still get them under my desk and have adequate foot room. G5, SoftRAID 3.3, Workstation 64KB stripe units.

    Yesterday, 2xRaptor1500sr:

    Extended Test Size: 100 MB Read: 128.798 MB/sec Write: 211.002 MB/sec
    Today:
    Extended Test Size: 100 MB Read: 127.805 MB/sec Write: 217.528 MB/sec

    Why is the 2x greater than the two 1x results added together, even with RAID overhead? Because of the 128KB SU of the 2x. Going to give it one more day, another test tomorrow morning and post a screenshot of whatever is best. k

  3. #143
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Screenshot of best 1xRaptor1500 results:

    Ran four more 2xRaptor1500 tests this morning, 100MB reads 128-129MB/s, writes 209-216MB/s so best result still:

    Going to experiment with 6xRaptors today (2xRaptor1500 and 4xRaptor74). k

  4. #144
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Best 6xRaptors (2xRaptor1500 and 4xRaptor74):
    Extended Test Size: 100 MB Read: 331.617 MB/sec Write: 472.023 MB/sec

    Now for the latest FrankenRAID, QB2.1, 6xRaptors (2xRaptor1500 and 4xRaptor74) and 4x 36GB Maxtor Atlas 15K II:
    Extended Test Size: 20 MB Read: 455.539 MB/sec Write: 828.638 MB/sec
    Extended Test Size: 30 MB Read: 505.348 MB/sec Write: 640.260 MB/sec
    Extended Test Size: 40 MB Read: 525.424 MB/sec Write: 778.210 MB/sec
    Extended Test Size: 50 MB Read: 524.940 MB/sec Write: 761.220 MB/sec
    Extended Test Size: 60 MB Read: 513.079 MB/sec Write: 767.038 MB/sec
    Extended Test Size: 70 MB Read: 512.070 MB/sec Write: 777.070 MB/sec
    Extended Test Size: 80 MB Read: 525.196 MB/sec Write: 841.450 MB/sec
    Extended Test Size: 90 MB Read: 526.070 MB/sec Write: 695.400 MB/sec
    Extended Test Size: 100 MB Read: 513.363 MB/sec Write: 816.587 MB/sec

    That is the fastest 100MB writes I have ever seen. As I test each configuration of drives, I am running Boot's Pshop tests. Will be awhile before putting it all together for posting to his test results. k

  5. #145
    Join Date
    Nov 2004
    Location
    Germany
    Posts
    2,352

    Talking Next step for ya: 1GB's








    Best regards

    Nicolas
    Custom Configurations! Rad Hacks and Mods!

  6. #146
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default Re: Next step for ya: 1GB's

    Well you got me thinking just for kicks. Way back when, Rick showed me how to add two Granite external shorty cables, in effect to add in my 4x X15.3 Burly to the Atlas Burly, two more drives per channel, and the UL4D in the 133MHz slot would be using 4 drives per channel. Have to think about it for later.

    For whatever reason I can't get Intech ZoneBench to see the new 150GB Raptors which means I can't test any of the configurations in ZoneBench. I have a question up on their forum. k

  7. #147
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    ZoneBench works. Operator error. k

  8. #148
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Yesterday I added the SCSI Burly with 4x Cheetah X15.3 drives to the previous 10x Drives. First I had to open the Burly up and change the ID of one of the drives. Either channel there was going to be an ID conflict with one of the Atlas drives. So did that. Then fired up just the Cheetah Burly, checked the IDs, and initialized all four drives without installing the SoftRAID driver. Shorty cables worked just fine. You must remove the two channel terminators on the Atlas Burly that is first on the chain coming from the UL4D (and therefore in my case also lose the remote LEDs for that Burly), then install the shorties in their place, and hook the other ends to the Cheetah Burly. Now the terminators on the Cheetah Burly become the termination for the entire 2-channels and 8 drives. Thanks Rick, I remembered.

    Shutdown and fired up the remaining cases, Burlys and FirmTek. Started the G5. Watched the watts, lower 600+W during startup to desktop, but then only the 10xDrives were stressing the watts since the Cheetahs were initialized with no driver. I created the 14xDrives RAID in SoftRAID. Rebooted and watched the watts. 699W highest during startup. Now I'm sweating because the APC BACK-UPS RS 1500VA is rated for 865W. Started a test and the watts blew thru 750W and climbing. Stopped the test. No way I'm going to stress the APC UPS that much.

    I did notice that tho the writes were going up, I only got thru 30MB when I stopped, the reads were slow, about 250MB/s. The Atlas and Cheetah drives on the same UL4D don't coexist well. In the past they were always on separate UL4Ds. Considered other options and decided to forget it. Too many things not right. In the future if I do add more Raptors or Atlas drives, they will have to be on a separate UPS. k

  9. #149
    Join Date
    Jan 2001
    Location
    Mobius Strip
    Posts
    13,045

    Lightbulb

    Do you have the Cinema on battery? When I asked APC, they didn't feel that hitting 865W would be a burden. When I put a TV on the UPS, that would try to pull a lot (and TVs are 'always on' which is a problem).

    Also, even putting the KAW in, may not allow the APC to get a good reading of how much power is required. I know the APC tech sort of 'flinched' when I mentioned I used one, and he didn't want to offer up more. I said I wanted to monitor usage (Belkin Bulldog does a good job via software, but APCs Powerchute won't, their business-class software does, but is not available for OS X, though he didn't seem to realize that). Maybe there is 3rd party software for monitoring an APC RS?

    One channel of UL4D would not really support four drives. Always subtract 10-15% for "overhead" to what you can expect, or about 280MB/sec. If you did choose to add additional 15K IIs, I would try 3 per channel.

    I had trouble with mixing drives on UL3D /after/ updating to 10.4.4. I have no idea why mixing SCSI drives would. I think you asked ATTO at one point, maybe Mark or someone at SR would offer some ideas.

    Are you pushing your 2.7DP to its limits already? Wishing for 32GB RAM even? I think it will be another two years before MacIntel Pro will be ready for the likes of heavy-lifting video editing.

  10. #150
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I have everything including the Cinema on either the APC Performance SurgeArrest 11 Outlet with Tel2/Splitter, Coax and Ethernet Protection, 12A max, PF11VNT3, or the Kensington MasterPiece Surge Remote (which is plugged into the SurgeArrest. Then from the SurgeArrest to the APC BACK-UPS RS 1500VA via its own single heavy duty AC cord. The Kensington has a circuit breaker and if I recall correctly some kind of aural warning. Neither has ever happened.

    The APC SurgeArrest has three LEDs.
    1. Site Wiring Fault (red), building wiring not properly installed. Never seen this LED on.
    2. Protection Working (green). I see this one continuously.
    3.Overcurrent Detection (yellow) when the unit is loaded above 12A. Never seen this LED on either.

    I am sure you are right about just adding 2 more 15K IIs. The other tip I expounded about way back when was from ATTO tech about how to read the ATTO Config Tool to see what features of the U320 spec the drives implement. My Cheetah X15.3 drives were purchased immediately when they became available. No doubt implement more of the spec now. The 15K IIs have more of the spec features but on the same channel they backtrack to the same as my older X15.3s. According to ATTO tech there is a way that might work by individually programming in Tools each drive but he did not, as I recall, think that would give better I/O results. 2x UL4Ds on separate buses worked just fine.

    I think my G5 is maxed out unless someone comes up with 2GB RAM sticks for it. But then I might be better off with a new latest G5 at some point. k

  11. #151
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Updated the diagram with changes:
    1. G5 info updated with RAM quantity and PCI-X cards
    2. Added iSight, mobo FW powered
    3. Added Belkin USB2 powered hub
    4. Photo of SATA Burly, top bay is a removable
    5. FirmTek SATA 2-bay
    6. Photo of 15KII Burly
    7. In order to plug in the USB2 hub, I used an AC cable splitter for both hubs

    k

  12. #152
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I have been aiming to test the mobo ports of my G5. It has one FW800 port on the rear and two FW400 ports, one front, one rear. There is one USB2 port on the front and two on the rear, backwards compatible with USB1.1. I tested only the rear FW800/FW400/USB2 ports using a Gurus Granite #20-1032 case, no longer sold, that has three FW800 and one USB2 ports on the rear. Power brick included as well as all necessary cables. Older technology to be sure but it works. Drive installed is an IBM Deskstar 120GB 180GXP-8MB drive formatted as one volume via SoftRAID 3.3, Workstation.

    FW800 (mobo, 800Mb/s = 100MB/s theoretical)
    QuickBenchª 2.1 Test Results File
    ©2000-2005 Intech Software Corp.
    Test file created on Tuesday, February 7, 2006 at 6:29:41 AM
    Test Volume name: ComboDrive

    Extended Test Size: 20 MB Read: 53.514 MB/sec Write: 51.642 MB/sec
    Extended Test Size: 30 MB Read: 53.886 MB/sec Write: 51.835 MB/sec
    Extended Test Size: 40 MB Read: 53.984 MB/sec Write: 51.791 MB/sec
    Extended Test Size: 50 MB Read: 54.297 MB/sec Write: 51.721 MB/sec
    Extended Test Size: 60 MB Read: 54.455 MB/sec Write: 51.783 MB/sec
    Extended Test Size: 70 MB Read: 53.776 MB/sec Write: 51.842 MB/sec
    Extended Test Size: 80 MB Read: 54.168 MB/sec Write: 51.653 MB/sec
    Extended Test Size: 90 MB Read: 53.892 MB/sec Write: 51.675 MB/sec
    Extended Test Size: 100 MB Read: 54.537 MB/sec Write: 51.801 MB/sec

    FW400 (mobo, 400Mb/s = 50MB/s theoretical)
    Extended Test Size: 20 MB Read: 40.235 MB/sec Write: 32.261 MB/sec
    Extended Test Size: 30 MB Read: 40.819 MB/sec Write: 32.262 MB/sec
    Extended Test Size: 40 MB Read: 40.669 MB/sec Write: 32.281 MB/sec
    Extended Test Size: 50 MB Read: 40.841 MB/sec Write: 32.244 MB/sec
    Extended Test Size: 60 MB Read: 40.802 MB/sec Write: 32.242 MB/sec
    Extended Test Size: 70 MB Read: 40.737 MB/sec Write: 32.313 MB/sec
    Extended Test Size: 80 MB Read: 40.925 MB/sec Write: 32.247 MB/sec
    Extended Test Size: 90 MB Read: 40.781 MB/sec Write: 32.242 MB/sec
    Extended Test Size: 100 MB Read: 40.886 MB/sec Write: 32.281 MB/sec

    USB2 (mobo, 480Mb/s = 60MB/s theoretical)
    Extended Test Size: 20 MB Read: 17.177 MB/sec Write: 15.629 MB/sec
    Extended Test Size: 30 MB Read: 17.177 MB/sec Write: 15.602 MB/sec
    Extended Test Size: 40 MB Read: 17.237 MB/sec Write: 15.627 MB/sec
    Extended Test Size: 50 MB Read: 17.327 MB/sec Write: 15.627 MB/sec
    Extended Test Size: 60 MB Read: 17.363 MB/sec Write: 15.651 MB/sec
    Extended Test Size: 70 MB Read: 17.312 MB/sec Write: 15.623 MB/sec
    Extended Test Size: 80 MB Read: 17.302 MB/sec Write: 15.639 MB/sec
    Extended Test Size: 90 MB Read: 17.293 MB/sec Write: 15.646 MB/sec
    Extended Test Size: 100 MB Read: 17.405 MB/sec Write: 15.631 MB/sec
    Big Ouch for USB2. In the past several Moderators have tried for better USB2 thruput in various G4s and trying various PCI cards, all combo cards of various combinations. No luck tho some cards did do better than others. k

  13. #153
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I have had a dotMac account for quite awhile now, maybe 3+ years. Anyway, I have been publishing web pages for several subjects, most about Da Boss's hiking trips such as Machu Picchu Peru and Norway. Others such as our Grandkids and weddings. And there is a catch-all site. I have not yet purchased iLife 6 so am using iPhoto 5.0.4.

    From within iPhoto you publish thru dotMac to create a webpage choosing either HomePage or .Mac Slides. The Slides choice can only be used once. If you use it again, you wipe whatever was there previously. HomePage allows max 48 photos whereas Slides allows much more, don't know if there is a limit. I tried about 300 photos. The problem with that is it takes a very long time to load the webpage and thumbnails, remembering that there are those who don't have DSL or cable. So I replaced the 300 with 58 photos for one of my projects.

    dotMac has very good instructions for how to create your pages. And you can password protect a site which I have done for my GrandKids. All of the sites are read only. Works very nice. Only one criicism is that when uploading my handiwork, I frequently get, when done, a warning window that it failed to complete. I've learned that is never true. Everything uploaded OK and am hopeful that iPhoto 6 will fix that criticism as well as have more webpage designs.

    A sample, the catch-all web page:
    http://homepage.mac.com/kaye_yum/Menu1.html
    Clicking one of those thumbnails takes you to another page of thumbnails. When you see a page with Start Slideshow near the top, you can choose that to get full or nearly full size photos. k

  14. #154
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    With the dotMac update to 1024MB, my Pref pane looks like this:

    The remaining 50MB is for Mail. I only use about 8MB for Mail so I may change that to 25MB.

    Anyway, I have well over 2000 photos ranging from 512KB to 2048KB, on average probably 1024KB. Industrial strength compression going on at Apple.

    This morning I watched 2 Apple QT movies covering the features of iPhoto6 in iLife '06. Gone are Homepage and iMac Slides, replaced and not similar by iWeb and Photocast. Very impressive. So I went to Apple Discussions and the big problem, among others, is that HomePage cannot be simply updated/upgraded to iWeb. No simple answer to that one. Most are starting over, not a simple solution to my way of thinking (and a lot of others who are sticking with HomePage for now). Ouch! k

  15. #155
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Can't keep up with a Quad, but latest version:

    CINEBENCH 9.5
    **************************************************

    Tester : kyum

    Processor : G5-2.7DP
    MHz : 2700
    Number of CPUs : 2
    Operating System : 10.4.5

    Graphics Card : X800XT
    Resolution : < 1920x1200>
    Color Depth : < Millions>

    **************************************************

    Rendering (Single CPU): 414 CB-CPU
    Rendering (Multiple CPU): 753 CB-CPU

    Multiprocessor Speedup: 1.82

    Shading (CINEMA 4D) : 405 CB-GFX
    Shading (OpenGL Software Lighting) : 1177 CB-GFX
    Shading (OpenGL Hardware Lighting) : 2150 CB-GFX

    OpenGL Speedup: 5.31

    **************************************************
    k

  16. #156
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Like I did with my GigaMDD, I ran Xbench on the G5 with a CRT which showed improvements.

    Xbench 1.2 results 166.61 here and 1.1.3 413.46 here k

  17. #157
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I have two more Raptor 1500's. Where RAID, SoftRAID 3.3, 128KB SU, 10.4.6.

    QB2.1 100MB results rounded, MB/s:
    1xRap1500 - 73/94
    2xRap1500 - 128/218, same as on DB
    3xRap1500 - 187/346
    4xRap1500 - 246/456
    4xRap1500/4xRap74 - 453/556
    4xRap1500/4x15K II - 535/679 with ATTO software updated
    11xDrivesRAID - 480/847
    ---3xRap1500/4xRap74/4x15K II
    11xDrivesRAID - 517/856
    ---4xRap1500/3xRap74/4x15K II
    4x15K II - 328/401

    Right now 11 drives seems to be the max for me. I tried 12 drives. 8x Raptors and 4x 15K II. Terrible reads, around 228 max, writes also max 827. Going to revisit that config and see if I can figure out what is wrong. I did try rebuilding it in SoftRAID twice after the first build. No change.

    If there is no change I will post the best 11xDrivesRAID screenshot and ZoneBench. k

  18. #158
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    4x Atlas 15K II results added to previous post. The good news, I was able to get the 12x drives RAID out of its funk. When it runs, it runs beautifully.

    Not what I had hoped for in writes but much more than expected in reads. More balanced.

    The bad news, once out of its funk of bad reads (200+MB/s), there is no consistency of making it stick. Sometimes it will survive a reboot, sometimes it will survive a shutdown, sometimes it will survive a shutdown and drive cases power down. But more often than not, it will not survive. Tried a number of utlities to regain the read speed and usually one of them would recover the reads but no consistency of which one would do it, so I just started using them all and that would do it. But a pain in the butt. Finally broke up the RAID and decided to give up. I tested each and every one of the 12 drives with SR and QB2.1 just to confirm they are all performing up to par. No subpar performance.

    And before breaking up the RAID, SoftRAID said all of the drives no errors and the 12x volume i/o requests: 317,102, i/o errors: 0. Back to 11x drives for the G5 and on to the G4 with a smaller RAID. k

  19. #159
    Join Date
    Jan 2001
    Location
    Mobius Strip
    Posts
    13,045

    Lightbulb

    kaye,

    This afternoon, ZZF price $176 very tempting.

  20. #160
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    I've been watching that page TZ but missed that latest price. Thanks. k

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •