View Full Version : Benchmarking for OSX...straight from the horse's mouth

03-13-2002, 09:11 PM
Here is an Apple release of some OSX benchmarking tools.

Be forewarned that the drive test has some 'optimistic' results on my machine.


[This message has been edited by chrismenke (edited 13 March 2002).]

03-14-2002, 11:12 AM
Thanks! those are interesting and useful. If you play with the settings and check just the write for 8MB it seems in line with what I would expect (or less even) than ExpressPro results. But there is a whole suite of tools there.

I'm trying to understand MONster but Disk Basher is much more flexible in configuring tests.

I think we've got a replacement for system, video, cache, and disk tests.


03-14-2002, 01:45 PM
It may be that it is having problems with my differentiating my volumes. My single 9.1GB Atlas V 9.2.2 Classic Drive on an Atto UL3D gives the same results as a 10 X 18.1GB Atlas V RAID, which gives the same results as the single IDE factory 7200 RPM Seagate boot drive.

03-14-2002, 02:38 PM
I had to select which volume it uses from the menu. I missed the note on www.xlr8yourmac.com (http://www.xlr8yourmac.com) Monday (3/11) mentioning CHUD (3 lines easy to miss and no links etc) but there is a more complete review etc over there now.

os_x_benchmark_tools (http://www.xlr8yourmac.com/OSX/os_x_benchmark_tools.html#storytop)


05-06-2002, 09:40 AM
Just noticed that over the weekend CHUD Performance Tools 2.0b was posted by Apple:

CHUD contain any update to Disk Basher 1.1.1.

[This message has been edited by Gregory (edited 06 May 2002).]

05-06-2002, 11:43 AM

We'll grab er up. Stick a copy on the ftp site.


06-05-2002, 04:47 PM
CHUD was just updated June 5th, probalby to work better with 10.1.5.

07-23-2002, 01:03 PM
CHUD 2.0 was posted by Apple.

Apple CHUD Tools 2.0 - Computer Hardware Understanding Development kit

Product Description:
The Computer Hardware Understanding Development kit (CHUD) is designed to simplify performance studies of PowerPC Macintosh systems running MacOS X. Hardware and software developers interested in tuning their products for the Macintosh platform can either take advantage of existing performance tools written with CHUD orroll their ow using the same set of building blocks.

* MONster: Samples hardware performance counters over time or in relation to other performance events.
* Shikari: Statistically shows where time is being spent (process or system wide) and correlates performance events with program code.
* Reggie: View the contents of PowerPC special purpose registers.
* bashpack.dmg: Tools for measuring memory, graphics, and disk throughput, in OS 9 or OS X (Carbon).

11-08-2002, 12:39 PM
CHUD 2.5 was posted recently (11/7/02):

CHUD 2.5 (ftp://ftp.apple.com/developer/Tool_Chest/Testing_-_Debugging/Performance_tools/CHUD_2.5.0.dmg.bin) direct download, or Performance Tools (ftp://ftp.apple.com/developer/Tool_Chest/Testing_-_Debugging/Performance_tools/) directory.

[This message has been edited by Gregory (edited 08 November 2002).]

11-19-2002, 01:16 PM
CHUD 2.5.1 (http://www.versiontracker.com/dyn/moreinfo/macosx/14358) has been posted.

DiskBasher was removed. Would have liked to rename this thread to "Apple CHUD Performance Tool"

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>
What's new in this version:

* CHUDProf.kext - fixed kernel panic on G3 machines when sampling
* CHUDProf.kext - fixed performance bug with OS counters on MacOS X 10.2 (now lower overhead)
* CHUDProf.kext - overhauled remote access and sampling facility management to account for poorly behaved clients
* Reggie SE - added CPU SPR control panel (disallows overclock of 750FX)
* Shikari - fixed potential crash if invalid column index selected
* Shikari - fixed bug in remote start/stop for > sample limit
* CHUD.framework - fixed problem with multiple remote access semaphore creation
* DiskBasher - removed from CHUD package due to unreliable results under certain filesystems
* Flurry Demo - crash/panic on DP systems fixed

11-19-2002, 03:03 PM
Thanks Gregory,

?So far I have yet to be comfortable with any of the results from DiskBasher. The numbers just don't seem to make sense with what I thought I knew. Generally the Average results are lower than I expect and the Peaks are much higher than I thought possible with the PCI bus limitations.

?I am still in the hunt for a test utility that shows an increase when testing a faster drive with all else being equal. DiskBasher will give very similar results with two very dissimilar drives. I keep wondering if it's the settings I am using. Or just ignorance on my part.


11-19-2002, 04:48 PM
I'm sure Apple would like to have a reliable test too http://macgurus.com/infopop/emoticons/icon_smile.gif

HDST QuickBench X seems okay. The new ATTO ExpressPCI tool is a firmware, nvram settings only tool, no benchmarking. Maybe they will add that later when it is possible? As MOSR stated, performance is iffy and Apple has a lot of work to do in the next months and years.