DH Photoshop Benchmark Script Database ONLINE!

***Zardon*** · Aug 12, 2005, 05:00 PM

As most of you know, we have our own custom Driverheaven Photoshop script which tests real world system performance. If you have never seen this a tutorial for installing and running along with download information is available over here

Due to the fact our huge forum thread was turning into a bit of a mess we now have opened an online database so everyone can enter and compare their scores. The new database system automatically reads your test results and gives you the final score time.

The fastest scores go to the top of the database so while we all want you to compare your scores, please enter genuine figures so everyone can get a good indication of relative system performance across the board.

Once you submit your figures, the database will calculate your total score and position you accordingly within the table, clicking on each user name gives a full system specification for that entry along with the version of photoshop used.

So please be sure to help us fill this database and head over to here and if you have any questions please discuss in this thread.

***Zardon*** · Aug 12, 2005, 05:00 PM

The_Neon_Cowboy · Aug 13, 2005, 10:50 AM

Brug about by reading carefully I noticed something I know others will notice too asked
or not. In regards to the section for the users with 512 MB to chage the applications
memory settings. Got me thinking...For example I have a gig of ram and by defalt CS2
is set to use 55% of my memory or 499MB . Cache levels setting is defaulted to 6.

My question is this:
I havn't tested yet but won't the amount of memory the application allow it to use alter the enviroment andthe end result? Shouldn't we come up with a value for everyone to use? Or should you add a info box so that people should even the amount thier PS program is set to?

***Zardon*** · Aug 13, 2005, 11:05 AM

Quote:

Originally Posted by The_Neon_Cowboy

Brug about by reading carefully I noticed something I know others will notice too asked
or not. In regards to the section for the users with 512 MB to chage the applications
memory settings. Got me thinking...For example I have a gig of ram and by defalt CS2
is set to use 55% of my memory or 499MB . Cache levels setting is defaulted to 6.

My question is this:
I havn't tested yet but won't the amount of memory the application allow it to use alter the enviroment andthe end result? Shouldn't we come up with a value for everyone to use? Or should you add a info box so that people should even the amount thier PS program is set to?

We dont need to "come up" with a figure for everyone to use... If photoshop has 300 meg of ram available while benchmarking it wont enter the paging file during testing therefore it wont cloud the end results. Its been carefully measured to ensure anyone with 512 meg of ram can effectively run this test. anyone under 512 system memory wont be able to run this test accurately.

As for those over 512 system memory, like yourself Mike you wont need to alter a thing as Photoshop will default to settings in excess of minimum requirements. People with 2gig or 4 gig of ram wont gain any benefits by having more memory for this specific test suite but they would during general photoshop use as they could open larger files and work with more layers etc than those with lower ram installed. That is outside this testing environment however as ive specifically set this benchmark so the file suite used doesnt expand over 300 meg memory usage. Its all explained on the documentation page.

Of course its also worth mentioning that those with better memory and tighter (faster) latency settings will get higher results than those with inexpensive looser memory .....

The_Neon_Cowboy · Aug 13, 2005, 11:13 AM

Cool, thanks for clearing that up. I could of figured it out myself but I figured
I'd ask if not for my self then for off chance others might be haveing have the
same thought.

I'll have my results up after bit

***Zardon*** · Aug 13, 2005, 01:20 PM

decent bench for your system mike.

joeb · Aug 30, 2005, 12:36 PM

Quote:

Originally Posted by Zardon

As for those over 512 system memory, like yourself Mike you wont need to alter a thing as Photoshop will default to settings in excess of minimum requirements.

I would like to see the effects of multiple disk systems. If you have your os/apps on your PATA system and data on the SATA system does that effect the results? If you had a benchmark that included paging what would the effects of seperating the scratch drive to a third subsystem? It may be that the current benchmark is not effected by these issues. KISS (keep it simple stupid) is the best policy. I may be asking for too many variables.

I was suprised to see CS2 with better speed results. My impression was that CS2 was slower than CS. One reason I have not upgraded was this speed issue. Do the actions in the benchmark benifit from the CS2 upgrade more than most?

My primary use is to convert raw images and edit the images. The current benchmark addresses the editing step. As the camera raw plug-in works outside of the rest of photoshop it would need a seperate benchmark.

The_Neon_Cowboy · Aug 30, 2005, 09:51 PM

Quote:

Originally Posted by Zardon

decent bench for your system mike.

Quote:

Originally Posted by joeb

I would like to see the effects of multiple disk systems. If you have your os/apps on your PATA system and data on the SATA system does that effect the results? If you had a benchmark that included paging what would the effects of seperating the scratch drive to a third subsystem? It may be that the current benchmark is not effected by these issues. KISS (keep it simple stupid) is the best policy. I may be asking for too many variables.

Well If you have enough memory, nothing is paged out, on a very low memory
system yes it'd have probubly an effect. Obviously have the scratch on anther
drive would be benifical. Especally since sich a system wuld be useing the system
page file freqently.

Quote:

Originally Posted by joeb

I was suprised to see CS2 with better speed results. My impression was that CS2 was slower than CS. One reason I have not upgraded was this speed issue. Do the actions in the benchmark benifit from the CS2 upgrade more than most?

For one I understand CS2 supports multi-threading and takes advatage
of some things CS1 doesn't , but then again I'm a little fuzzy on the topic

pr0digal jenius · Sep 1, 2005, 11:11 PM

just how much does RAM affect this test....cuz I have 4x512 DDR333......would DDR400, or even 2x1024 improve it a lot you think?

The_Neon_Cowboy · Sep 2, 2005, 12:20 AM

Quote:

Originally Posted by pr0digal jenius

just how much does RAM affect this test....cuz I have 4x512 DDR333......would DDR400, or even 2x1024 improve it a lot you think?

BANDWITH AND TIMEINGS SHURELY WOULD HAVE AN EFFECT

Scargo · Oct 17, 2005, 11:21 AM

I was hoping to find discussions or information about why the differences in scores. Which ones test what part of your system? Can I tweak anything, using what I have. For instance, my Underpainting (12) score is poor @ 30.1 and so are 5 and 6. Why these scores? Even some AMD processors with significantly lower scores didn't do as good (or much better) than I did with the extrude and smart blur (10 & 11).
I am running my 15K RPM SCSI drive solely for the scratch disc.In PS I'm at 75%memory allocation. I have 2 Gig allocated in Windows for page filing and have the minimum page file allocation size Windows allows me to have (in the normal page file allocation area, on the slower main drive/OS). I read that you do not want page filing to be zero on the main drive.
On the other hand, I read in this thread that the benchmark does not get into pagefiling and doesn't even use much of the 2 (max usable by PS) Gig of RAM. I assume there is some use of the scratch disc during the benchmark?
Any comments/explanations?
Thanks

Follow-up:After further tweaking today I found that going from a cache level of 4 to 2 significantly improved my overall score. A cache level of 1decreased performance somewhat. How can the slight boost in my overall CPU speed from 2645 to 2657 explain a 57% decrease in my texturizer times...if not for the cache change??

Asus A8N-E nForce4 Ultra Athlon 64 (FX) Skt939 DDR ATX, w/Audio, Gigabit LAN, RAID/Serial ATA
AMD Athlon 64 4000+ cooled by CPUCOOL ZALMAN CNPS7000B-CU LED
OCZ4002048ELDCPE-K _DDRAM 2x1028MB (“Platinum”, matched pair-low latency 2-3-2-5) OCZ4001024ELDCPER2-K _DDRAM 2x512MB (“Platinum”, matched pair-low latency 2-2-2-5)
VGA MSI NX6600-TD256E RT
Western Digital Caviar RE WD2500SD 250GB Serial ATA 7200RPM w/8MB Buffer
Fujitsu MAU Series MAU3036NP 36.7GB 68pin U320-SCSI 15,000RPM Hard Drive w/8MB Buffer
ADAPTEC|29320A-R SCSI RAID CARD
OS: Windows XP Professional, SP2+ and Photoshop CS2
Running @ 2.65Ghz on air---very cool!

altimas · Apr 18, 2006, 12:04 AM

Any idea how to improve my scores?
Some of them seem real good but others stink.
Setup is an A64 winchester@ 242x4- 2420 mhz and 1GB of ram- 2x 512.. at cas 3,3,3,8 ram at 201 mhz
Is my memory timings hurting some of the scores a lot?
heres all my scores..

Code:

1 Texturiser Test (1)
2.0 (1)
2 CYMK Colour Conversion
5.3
3 RGB Colour Conversion
6
4 Dust and Scratches
4.5
5 Water Color
17.9
6: Texturiser Test (2)
2.9
7: Stained Glass
8.6
8: Lighting Effect
10.9
9: Mosiac Tiles
17.9
10: Extrude
58.9
11: Smart Blur
54.4
12. Underpainting
21.7

Scargo · Apr 19, 2006, 07:33 AM

RAM is the #1 priority for you. #2 would be a dedicated fast, small drive for PS cache memory.
I have 4Gb and run in "3Gb bridge" mode when doing intense PS. Microsoft has instructions about this mod.
Normally Windows on utilizes 2Gb of RAM for cache (PS calls its cache memory "scratch disk"). Both want you to have this temporary storage allocation be on a seperate drive. There are tweaks to be made in PS preferences and in PS and Windows cache and pagefile appropriations. Read up about that.

Judas · Apr 19, 2006, 11:30 AM

you know, perhaps setting timing mode and marking down the time it takes to load that image should be incorporated, as it's indeed part of the real world test. I mean graphic artists are constantly moving around things, opening and transfering and copying images and portions from one place to the next.

so enabling timing mode and then including the time it takes to open the document.... should be included?

altimas · Apr 19, 2006, 04:31 PM

Well I did happen to come upon 100 dollers (that can be used for whatever I want) but I'm not too sure I want to spend money on DDR1 when I should be switching to DDR2 with conroe

I do play elder scrolls- oblivion which needs uses about a gig of ram total at the settings i play at.
stutters...
On the postive side I could include that memory when I sell off my athlonxp 250+ mobo and 6600GT combo.

***Zardon*** · Apr 20, 2006, 02:24 AM

Quote:

Originally Posted by Judas

you know, perhaps setting timing mode and marking down the time it takes to load that image should be incorporated, as it's indeed part of the real world test. I mean graphic artists are constantly moving around things, opening and transfering and copying images and portions from one place to the next.

so enabling timing mode and then including the time it takes to open the document.... should be included?

No, then hard drive comes into play. something intentionally avoided in this test and noted on the photoshop page information text. If i had wanted to incorporate the hard drive into this performance test the image would have also been larger to use the PS scratch disk but this will cloud the results the test is aimed at - CPU, Memory/bandwidth etc

Judas · Apr 20, 2006, 10:55 AM

oh ok. Been awhile since i read though everything on there since it was first posted, was checking out a few machines that i had ran it on and wonder if perhaps the preformance messurement would include HD at all...

understand now.

*ocztony* · Jun 6, 2006, 08:03 PM

Hey guys i just wanted to show what the new Conroe is like at the DH bench for CS2, seem you have a new leader at the top

This with a 2.16 overclocked to 3143mhz all on air.

***Zardon*** · Jun 23, 2006, 07:13 AM

Quote:

Originally Posted by ocztony

Hey guys i just wanted to show what the new Conroe is like at the DH bench for CS2, seem you have a new leader at the top

This with a 2.16 overclocked to 3143mhz all on air.

yeah they pretty much smoke everything else........

swimtech · Jul 30, 2006, 12:07 AM

...and a Macbook Pro running Windows ties a Dual 2.5Ghz G5 PowerMac running OSX.

tastyweat · Aug 16, 2006, 05:07 PM

Hehe... first run with new processor and I get 9th on the leaderboard with 110.3 seconds

Happy

tastyweat · Aug 16, 2006, 05:22 PM

Fresh Restart and I jumped to 8th with 106 seconds.... 0.5s off 7th place!!! lol