Finally. 9262 and 9800 GTX Crysis Benchys

Ok, Finally got results on my first test
Game is at 1680x1050, HIGH settings
Forceware version: 176.09
Machine spec in sig
Running Vista x64 Dx10

TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 91.44s, Average FPS: 21.87
Min FPS: 14.88 at frame 1352, Max FPS: 26.98 at frame 992
Average Tri/Sec: -20707640, Tri/Frame: -946767
Recorded/Played Tris ratio: -0.97
!TimeDemo Run 1 Finished.
Play Time: 88.72s, Average FPS: 22.54
Min FPS: 12.41 at frame 1989, Max FPS: 27.32 at frame 64
Average Tri/Sec: -21067632, Tri/Frame: -934521
Recorded/Played Tris ratio: -0.98
!TimeDemo Run 2 Finished.
Play Time: 89.06s, Average FPS: 22.46
Min FPS: 12.41 at frame 1989, Max FPS: 27.37 at frame 71
Average Tri/Sec: -20985772, Tri/Frame: -934527
Recorded/Played Tris ratio: -0.98
!TimeDemo Run 3 Finished.
Play Time: 88.49s, Average FPS: 22.60
Min FPS: 12.41 at frame 1989, Max FPS: 27.37 at frame 71
Average Tri/Sec: -21131338, Tri/Frame: -934944
Recorded/Played Tris ratio: -0.98
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

 

avg fps is 22.60 ? good or bad

 

Not sure. I havent done much comparing of numbers so I dont know how this stacks up. I do know the game feels very zippy on this machine.

 

doesnt look so good

 

Here is at VERY HIGH settings:

==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 155.16s, Average FPS: 12.89
Min FPS: 8.54 at frame 1330, Max FPS: 14.74 at frame 995
Average Tri/Sec: -5192288, Tri/Frame: -402828
Recorded/Played Tris ratio: -2.28
!TimeDemo Run 1 Finished.
Play Time: 153.77s, Average FPS: 13.01
Min FPS: 7.61 at frame 1060, Max FPS: 14.81 at frame 984
Average Tri/Sec: -5072499, Tri/Frame: -390008
Recorded/Played Tris ratio: -2.35
!TimeDemo Run 2 Finished.
Play Time: 153.42s, Average FPS: 13.04
Min FPS: 7.61 at frame 1060, Max FPS: 14.81 at frame 984
Average Tri/Sec: -5067232, Tri/Frame: -388701
Recorded/Played Tris ratio: -2.36
!TimeDemo Run 3 Finished.
Play Time: 156.94s, Average FPS: 12.74
Min FPS: 7.41 at frame 1463, Max FPS: 14.93 at frame 987
Average Tri/Sec: -4955880, Tri/Frame: -388890
Recorded/Played Tris ratio: -2.36
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

 

this is my single gpu score at your same settings. differences may be that i am running XP and in dx9.

1680x1050 all high dx9 = 23.37 fps avg. single gpu

 

The last run, the screen saver kicked in I think is why FPS went down so much there.

 

ARGH recently posted 23.37 for his rig (See: http://forum.notebookreview.com/showthread.php?t=291995&page=2)

He's running XP though, which might give him another 1-2 FPS.

 

have you ran the tests with DX9 mode?

 

Still running my last tests (1024 at HIGH and VERY HIGH) using dx10

I don't know how to get it to use dx9 tho. I am using the crytek benchmark batch programs.

 

See the posts above... how does one do that?

 

And now for 1024x768 HIGH settings

==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 53.81s, Average FPS: 37.17
Min FPS: 25.69 at frame 1959, Max FPS: 49.03 at frame 991
Average Tri/Sec: 37725688, Tri/Frame: 1014974
Recorded/Played Tris ratio: 0.90
!TimeDemo Run 1 Finished.
Play Time: 51.30s, Average FPS: 38.98
Min FPS: 24.42 at frame 1287, Max FPS: 51.56 at frame 86
Average Tri/Sec: 40008240, Tri/Frame: 1026258
Recorded/Played Tris ratio: 0.89
!TimeDemo Run 2 Finished.
Play Time: 51.57s, Average FPS: 38.78
Min FPS: 18.80 at frame 1960, Max FPS: 51.56 at frame 86
Average Tri/Sec: 39792968, Tri/Frame: 1026146
Recorded/Played Tris ratio: 0.89
!TimeDemo Run 3 Finished.
Play Time: 51.29s, Average FPS: 38.99
Min FPS: 18.80 at frame 1960, Max FPS: 51.56 at frame 86
Average Tri/Sec: 40009168, Tri/Frame: 1026056
Recorded/Played Tris ratio: 0.89
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

 

to get Crysis to run in DX9 mode (for the benchmark batch files) edit them to look like this:

 

Unless I'm very confused (quite likely) these scores look great! Close to what a Dell user posted for 8800M GTX SLI, and also close to results I've seen for the 8800 GT desktop card.

 

Last dx10 Results
1024x768 at VERY HIGH

==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 87.59s, Average FPS: 22.83
Min FPS: 15.00 at frame 1354, Max FPS: 28.50 at frame 998
Average Tri/Sec: -15357593, Tri/Frame: -672554
Recorded/Played Tris ratio: -1.36
!TimeDemo Run 1 Finished.
Play Time: 85.75s, Average FPS: 23.32
Min FPS: 14.56 at frame 1246, Max FPS: 29.61 at frame 64
Average Tri/Sec: -15400648, Tri/Frame: -660300
Recorded/Played Tris ratio: -1.39
!TimeDemo Run 2 Finished.
Play Time: 86.37s, Average FPS: 23.16
Min FPS: 11.45 at frame 1941, Max FPS: 29.61 at frame 64
Average Tri/Sec: -15251356, Tri/Frame: -658592
Recorded/Played Tris ratio: -1.39
!TimeDemo Run 3 Finished.
Play Time: 85.90s, Average FPS: 23.28
Min FPS: 11.45 at frame 1941, Max FPS: 29.61 at frame 64
Average Tri/Sec: -15327664, Tri/Frame: -658300
Recorded/Played Tris ratio: -1.39
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

 

And my last test (unless anyone wants anything else)
This one is back to 1680x1050 HIGH running DX9

==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 87.01s, Average FPS: 22.99
Min FPS: 18.10 at frame 1951, Max FPS: 27.87 at frame 980
Average Tri/Sec: -21743134, Tri/Frame: -945901
Recorded/Played Tris ratio: -0.97
!TimeDemo Run 1 Finished.
Play Time: 82.97s, Average FPS: 24.11
Min FPS: 18.10 at frame 1951, Max FPS: 28.35 at frame 982
Average Tri/Sec: -22514862, Tri/Frame: -933999
Recorded/Played Tris ratio: -0.98
!TimeDemo Run 2 Finished.
Play Time: 82.94s, Average FPS: 24.11
Min FPS: 18.10 at frame 1951, Max FPS: 28.63 at frame 1002
Average Tri/Sec: -22527104, Tri/Frame: -934251
Recorded/Played Tris ratio: -0.98
!TimeDemo Run 3 Finished.
Play Time: 82.96s, Average FPS: 24.11
Min FPS: 18.10 at frame 1951, Max FPS: 28.74 at frame 987
Average Tri/Sec: -22542054, Tri/Frame: -935031
Recorded/Played Tris ratio: -0.98
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

 

Looks like Dx9 give a 1 or 2 FPS boost over the same settings using Dx10.

 

Good stuff sligsystems!

If you feel the need for some dumb reason to go with one hard drive, I’ll gladly take one off your hands. You have an awesome setup, and it’s nice of you to take the time to fulfill our greedy wants.

 

though I think those benchmarks are pretty good for a single mobile GPU... you can get slightly better scores if you optimize the system

disable the resource hogs (in Vista):
- Theme: switch to Classic Mode (to disable Aero)
- Drive Indexing service
- Windows sidebar
- Auto Defrag
- System Restore
- Volume Shadow Copy (Volume Snapshot Service or VSS)
- Microsoft Software Shadow Copy Provider service
- AntiVirus / Spyware suites

XP does not have that much crap... thats why its more optimal for gaming and benchmarks with little tweaks.

 

Opps! Looks as if I was confused after all... I was thinking this run was at 1920x1200. So please ignore what I said.

 

I dont think those scores are any different from the 9800GT. Can anyone with a single GT do the crysis benchmarks?

 

Yes, that appears to be normal. Here's a quote from ExtremeTech:

DirectX 9 Performance
As we mentioned earlier, there really is little reason to run Crysis in DX10 mode. Right now, the game runs slower in DX10 than DX9 on every card we tested. With the configuration file hack mentioned earlier, the Very High options are available to you even in DX9, and looks nearly the same as DX10.

http://www.extremetech.com/article2/0,2845,2222416,00.asp

 

Why don't you turn off SLI and give it a try?

 

^^ good idea.

 

Thanks for the tips. I am hoping someday, that if the 9800 GTX SLI ever comes out, I will be able to upgrade my machine at that point. But I am not sure this box would support it. Hopefully it would just be a matter of getting another card and a bridge...but who knows.

 

I think it will happen, and you'll have a heck of a machine when you upgrade.

 

from these benchmarks it appears the 9800gtx is really any more powerfull than the 9800gt or the 8800gtx for our laptops.

 

I've run both the built in Crysis GPU test and the one available for download.


I get the following FPR with one 9800 GT

 

This is an excellent run for comparison purposes. Even running Vista you ran a little faster in dx9 than ARGH and I do with XP.

 

It would be great if you could rerun that with no AA to get a direct comparison with the 9800M GTX. A higher resolution would be good too.

 

Well.... maybe a little... 5-10% perhaps. Still, we haven't seen it at the highest resolutions.

 

Just to see what all 3 can do, I’ll add the 8800M GTX. Now it will be on the old platform, so scores might be slower than if it were running on the new platform. But having a good idea of what all three can do, should be a good reference point.


T9300/8800M GTX/3GB DDR 2 (4GB actual memory) running at 667


http://img254.imageshack.us/img254/434/crysisbenchex5.jpg


http://img508.imageshack.us/img508/7150/crysis2mo3.jpg

 

not sure how you achieved 38 avg. fps on your 1024x768 run with 2x aa. did you use a 3rd party benchmark tool or the one supplied by the game, Benchmark_GPU?

i achieved 36.60 avg. fps with no aa on that same run.

one thing i noticed is that when you choose the option that supposedly changes all the settings for you to a pre-defined low, med, or high, it will not change the objects quality setting and you are forced to change that manually.

 

Doesn't seem much faster, I got an average of 18.04fps in Dx10 running Vista x64. I was looking foward to replacing my 8800gtx with the 9800gtx.
: (

 

the 9800M GTX only has 16 more shaders over the 8800M GTX.

so the performance difference would not be that high.

 

Your 1680x1050 run is virtually identical to the single card run I just made.

---------------------------------------------------------------------

8/30/2008 5:11:01 PM - XP
Beginning Run #1 on Map-island, Demo-benchmark_gpu
DX9 1680x1050, AA=No AA, Vsync=Disabled, 32 bit test, FullScreen
Demo Loops=3, Time Of Day= 9
Global Game Quality: High
==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 1 Finished.
Play Time: 80.64s, Average FPS: 24.80
Min FPS: 12.80 at frame 139, Max FPS: 29.23 at frame 997
Average Tri/Sec: -24475430, Tri/Frame: -986892
Recorded/Played Tris ratio: -0.93
==============================================================

Completed All Tests

<><><><><><><><><><><><><>>--SUMMARY--<<><><><><><><><><><><><><>

8/30/2008 5:11:01 PM - XP

Run #1- DX9 1680x1050 AA=No AA, 32 bit test, Quality: High ~~ Overall Average FPS: 24.80

---------------------------------------------------------------------

Our scores are roughly the same as those Sliqsystems got with the 9800M GTX. Giving him a 5% "Vista Discount", it looks as if these tests show that the 9800M GTX is about 5% faster in Crysis than the 8800M GTX and the 9800M GT. I hate to acknowledge it, but this is what the notebookcheck game comparisons showed.

 

Come on, why are my benchmarks being questioned when I provided you with all the information? I play the game with all high settings, so there is nothing for me to go in and tinker with. It’s the same utility that was posted before mine that was being used, and did not see it questioned. I’ve even had someone question my 3DMARK score?? WTH? I’m not posting anything that I can’t do, so what you see is what you get. I didn’t post it to bloat or anything ( not on purpose anyway ), I just wanted to see comparisons of all three cards. I’m sure I can raise that score as soon as I get the full version of Crysis, since it seems to run faster than the demo.

 

You, me, and probably hundreds of others!

 

Yes, and maybe that's why they were the only ones to show it. With that being said, the 9800GTX is still faster!

 

Wow those are way better than mine.

 

Glad to take one for the team But even thought the 9800 GTX is only marginally faster than 8800, for a new system, I would definately go with 9800. To upgrade to one from 8800, its probably not worth it.

I am just waiting for the day I can sli up this 9800 GTX to its long lost brother somewhere in the future. Should perform pretty well I would think.

 

sliqsystems,


Wait till you’ve had time to tweak that machine and use the latest drivers, it will scream! Regardless of anything else, you have a kick azz system! All of your scores are out of the box, just give it time.

 

.............

 

Here it is at the higher res no AA

 

Single 9800M GT set to the same 1680x1050 HIGH game settings:

 

Can we get a GPU-Z screenshot?

 

wow that 9800M GTX is a big disappointment. $400 more for pretty much no visible performance gain at all Dell and Alienware will be probably charging $1000 for the SLi setup in their systems for a performance that is so outdated now Hope Nvidia will get their act together and release GT280M or GX2M or some new card to close up the gap between laptops and desktops.

 

I hate to tell you this, but the third party benchmark is totally screwed up when set for dx10 and high/medium/low settings. Try the Crytek benchmark.

 

I did, look ↑ but it did the test with AA.

How do I run it w/o AA?

 

Look at the Crysis built in GPU test and the 3rd party app. Results are the same.