New Crysis WARHEAD Benchmark Tool BETA

[maudio3]

Have they got the steam version working yet?

[Mr.John]

Originally Posted by maudio3

Have they got the steam version working yet?

Since build 0.20 steam version is working fine as reported by Guru3D's forums users. The latest build 0.24 can be download here or here

Latest updates
=======================
0.24 - 25/09/2008
Bug fix
- Program won't start if using Windows versions different than english/brazilian portuguese

0.23 - 25/09/2008
New AA 8xQ, AA 16x and AA 16xQ added for NVIDIA GFX cards.

0.22 - 25/09/2008
Bug fixes
- CUSTOM options as reported by HalJordanGL2814
- NVIDIA driver version string causing startup error (sorry!)

0.21 - 25/09/2008
New layout with new options
Added v-sync and windows mode test options
Added update check button (globe icon on bottom of screen)

0.20 - 24/09/2008
STEAM fix.

0.19 - 24/09/2008
Added missing sys_spec_GameEffects and sys_spec_MotionBlur on CUSTOM configuration dialog (Thanks again HalJordanGL2814)

0.18 - 24/09/2008
Added option to save and load benchmark queues
Added option to load custom cvars text file

0.17 - 23/09/2008
STEAM fix.

0.16 - 23/09/2008
Added support for STEAM. Thanks to HalJordanGL2814 from adrenaline forum for testing and feedback.
Added experimental system information and results submission to ranked server (not working yet on this build)
Added option to save text results after benchmarking

0.15 - 22/09/2008
Added support to WH internal flythrough timedemos

[maudio3]

Thanks a bunch going to try it out tonight.

[Mr.John]

Originally Posted by maudio3

Thanks a bunch going to try it out tonight.

Ok, please try it and tell us if it's working fine on your rig.

[chevell]

I have the steam version of Crysis warhead, it was asking me to point it to the directory where my Crysis.exe file is but there isn't one with the Steam version,so I pointed it to the Crysis warhead file in Steam which opens the bench. I have the benchmark program up but it's asking me for my registration code which you don't get with the Steam version, unless it's hiding some where and I don't know about it. Is anyone else with Steam getting this?

[Mr.John]

Originally Posted by chevell

I have the steam version of Crysis warhead, it was asking me to point it to the directory where my Crysis.exe file is but there isn't one with the Steam version,so I pointed it to the Crysis warhead file in Steam which opens the bench. I have the benchmark program up but it's asking me for my registration code which you don't get with the Steam version, unless it's hiding some where and I don't know about it. Is anyone else with Steam getting this?

The path for Steam WH is something like
C:\Program Files\Steam\steamapps\common\Crysis WARHEAD

If you're running Vista 64 the path is
C:\Program Files (x86)\Steam\steamapps\common\Crysis WARHEAD

The benchmark program will not ask for your registration code, if Crysis Warhead is running fine from Steam program it should be running without asking for any information.

[chevell]

Thanks for the help, I used those paths you described and they worked to get the benchmark open, I already had that part figured out but, it keeps asking for the registration number with a seperate lil window. I'm not sure how to find this number with Steam or where Steam might have put it. I'm using Vista x64.

Update: I can now report what the problem was, it was my fault for not running the Crysis warhead game at least once because I was working with a fresh install of everthing. Apparently the Steam account registers itself when you run the game for the first time which I hadn't done yet so it's working now, Thanks for the help Mr. John. The bench looks good now.

I go from 31 frames avg in Enthusiast settings Dx10 1680x1050 in Ambush, to 125+fps plus in performance. It looks much better with Enthusiast settings and I have no problems at all running the game in multi or single player with a single GTX280.

And Dx10 is a bit slower but doesn't get lagged like DX9 does and it looks better. What I found interesting was that adding AA actually improved times a bit, completely the opposite of what the old Crysis AA settings did.

[johnnyfiive]

Here are my results:

DirectX 9.0c GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 0xx
==> Framerate [ Min: 12.81 Max: 33.68 Avg: 20.95 ]

DirectX 9.0c ENTHUSIAST 3X @ Map: ambush @ 0 1680 x 1050 AA 0xx
==> Framerate [ Min: 5.21 Max: 25.52 Avg: 14.71 ]

DirectX 9.0c GAMER 3X @ Map: ambush @ 0 1440 x 900 AA 0xx
==> Framerate [ Min: 14.60 Max: 38.52 Avg: 23.94 ]

DirectX 9.0c ENTHUSIAST 3X @ Map: ambush @ 0 1440 x 900 AA 0xx
==> Framerate [ Min: 10.34 Max: 28.53 Avg: 17.53 ]

[largon]

Results beaten out of my 8800GTS (G92, 842MHz/2214MHz/DDR-2088) at different CPU (Q6600) speeds:

DX9, all settings on GAMER, map: ambush @ 1680x1050 0xAA 4xAF

CPU	Min	Max	Avg
2.4GHz	20.78	48.68	36.14
3.6GHz	25.21	47.03	36.38
Gain	+21%	-3.4%	+0.7%

DX9, all settings on ENTHUSIAST, map: ambush @ 1680x1050 0xAA 4xAF

CPU	Min	Max	Avg
2.4GHz	16.57	29.00	23.55
3.6GHz	16.30	28.90	23.30
Gain	-1.6%	-0.3%	-1.0%

"2.4GHz" = 9x266MHz / DDR-667 4-4-4-12
"3.6GHz" = 9x401MHz / DDR-802 4-4-4-12

CPU overclock truly did not gain anything on Enthusiast LOD. I ran three sets of tests each with 3 consecutive runs to verify the result on 3.6GHz with "Enthusiast" LOD. That's a 50% CPU OC and the gain is a round 0. Btw, at 3.6GHz the system drew ~50W more power during the benchmark than with CPU at 2.4GHz (~350W vs. ~300W).

CPU is dead.

[chevell]

Sorry but one system means absolutely nothing. Most will have better frame rates with higher clocks up too about 4.0 Ghz. Countless hours or testing on Crysis benchmark has already proved this. We have an entire thread going about this here. CPU is far from dead.

http://forums.extremeoverclocking.co...d.php?t=286518

[largon]

Heh, that linked thread is useless for comparing impact of CPU power.

[chevell]

It's obvious you don't know what your talking about and that you also can't read. Your statement is beyond rediculous. Your so wrong it's not even debatable. This forum is obviously a complete waste of my time and after reading statements like yours I have to wonder about the future of mankind. Do you really believe what your saying is true? You have so much to learn and such a small chance of learning it with your messed up attitude. Sorry you don't have the capacity to figure out what the truth is here.

[largon]

Originally Posted by chevell

It's obvious you don't know what your talking about and that you also can't read. Your statement is beyond rediculous. Your so wrong it's not even debatable. This forum is obviously a complete waste of my time and after reading statements like yours I have to wonder about the future of mankind. Do you really believe what your saying is true? You have so much to learn and such a small chance of learning it with your messed up attitude. Sorry you don't have the capacity to figure out what the truth is here.

And you are very rude.

edit:
My "messed up attitude"?!?
Heh, that's a masterpiece.

edit2:
And just to show your peculiarly passionate stand "for the CPU" is nothing but a fantasy, I ran similar benches in the original Crysis:

DX9, all settings on High, default benchmark_gpu @ 1680x1050 0xAA 4xAF


DX9, all settings on High with HP's custom config 1.3.1, default benchmark_gpu @ 1680x1050 0xAA 4xAF

CPU is dead.

edit3:
A-hahaha!

Originally Posted by chvell

I agree the benchmark [Crysis benchmark] is flawed, it's too easy to use a CRT to cheat on the bench.
http://forums.extremeoverclocking.com/showpost.php?p=3224056&postcount=11

Goodbye. You will not be missed.

[Mr.John]

There's a update available, build 0.28. Please download it and leave your feedback.

Originally Posted by largon

Results beaten out of my 8800GTS (G92, 842MHz/2214MHz/DDR-2088) at different CPU (Q6600) speeds:

DX9, all settings on GAMER, map: ambush @ 1680x1050 0xAA 4xAF


DX9, all settings on ENTHUSIAST, map: ambush @ 1680x1050 0xAA 4xAF

"2.4GHz" = 9x266MHz / DDR-667 4-4-4-12
"3.6GHz" = 9x401MHz / DDR-802 4-4-4-12

CPU overclock truly did not gain anything on Enthusiast LOD. I ran three sets of tests each with 3 consecutive runs to verify the result on 3.6GHz with "Enthusiast" LOD. That's a 50% CPU OC and the gain is a round 0. Btw, at 3.6GHz the system drew ~50W more power during the benchmark than with CPU at 2.4GHz (~350W vs. ~300W).

CPU is dead.

I did a similar test here with CPU @ stock and CPU @ 3.6 ghz and get no performance improvement in Crysis/DX 10/Very High. Did not tested WH yet but look like it didn't changed.

Please, can you run the test again with AF controlled by app and CPU @ 3.6 ghz (ambush, 16x10, AA 0x, DX9, enthusiast)

Thanks!!

Originally Posted by chevell

Update: I can now report what the problem was, it was my fault for not running the Crysis warhead game at least once because I was working with a fresh install of everthing. Apparently the Steam account registers itself when you run the game for the first time which I hadn't done yet so it's working now, Thanks for the help Mr. John. The bench looks good now.

I go from 31 frames avg in Enthusiast settings Dx10 1680x1050 in Ambush, to 125+fps plus in performance. It looks much better with Enthusiast settings and I have no problems at all running the game in multi or single player with a single GTX280.

And Dx10 is a bit slower but doesn't get lagged like DX9 does and it looks better. What I found interesting was that adding AA actually improved times a bit, completely the opposite of what the old Crysis AA settings did.

Thank you for reporting and thanks for the results of your GTX280.

[largon]

Mr.John,

DX9, all settings on ENTHUSIAST, map: ambush @ 1680x1050 0xAA 0xAF (application controlled)

CPU	Min	Max	Avg
2.4GHz	17.60	31.50	24.81
3.6GHz	17.34	31.71	25.17
Gain	-1.5%	+0.7%	+1.5%

I'll also do this cpu test and report back

[FooArm]

i think ill try that out

[johnnyfiive]

my latest results, which surprised me...

10/12/2008 11:50:18 - Microsoft® Windows Vista™ Ultimate 64-bit Service Pack 1

DirectX 9.0c GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 0xx
==> Framerate [ Min: 14.14 Max: 47.58 Avg: 32.25 ]

DirectX 9.0c ENTHUSIAST 3X @ Map: ambush @ 0 1680 x 1050 AA 0xx
==> Framerate [ Min: 8.85 Max: 34.64 Avg: 24.06 ]

I got worse results in XP which makes no sense, but I'm not complaining. I'm back on Vista64. CPU is at 2.42GHz (Phenom 9600BE), 3870's are at 837Mhz/1134Mhz.

[largon]

Whoa!
That's a crazy improvement over the your WinXP results.

Hmmm...
I think I'll try Vista64 Ultimate aswell as I just happen to have it in dual boot...

[johnnyfiive]

Originally Posted by largon

Whoa!
That's a crazy improvement over the your WinXP results.

Hmmm...
I think I'll try Vista64 Ultimate aswell as I just happen to have it in dual boot...

Yeah, I know! Big improvement!

[FelipeV]

My tests, CPU stock and CPU overclocked with memory @ 1000Mhz.

[EastCoasthandle]

It would be nice to be able to stretch the Results box (or the program itself) so you can take a screen shot of all the information in the results section of this program.

[Binge]

Processor: Intel(R) Core(TM)2 Duo CPU E8600 @ 3.33GHz @ 4441 Mhz
CPU ID: Intel64 Family 6 Model 23 Stepping 10
Operating System: Microsoft® Windows Vista™ Ultimate 64-bit Service Pack 1
Physical memory: 4.00 GB
Display adapter: ATI Radeon HD 4870 X2 1024 MB
Driver version: Catalyst 8.10 (8.541-080923a-069997C-ATI)

================================================== ============
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 85.82s)
!TimeDemo Run 0 Finished.
Play Time: 41.09s, Average FPS: 48.68
Min FPS: 34.76 at frame 95, Max FPS: 62.72 at frame 1407
Average Tri/Sec: -49226368, Tri/Frame: -1011245
Recorded/Played Tris ratio: 0.97
!TimeDemo Run 1 Finished.
Play Time: 38.47s, Average FPS: 51.99
Min FPS: 34.76 at frame 95, Max FPS: 64.77 at frame 1987
Average Tri/Sec: -51027272, Tri/Frame: -981506
Recorded/Played Tris ratio: 1.00
!TimeDemo Run 2 Finished.
Play Time: 38.10s, Average FPS: 52.49
Min FPS: 34.76 at frame 95, Max FPS: 65.49 at frame 1995
Average Tri/Sec: -52311228, Tri/Frame: -996586
Recorded/Played Tris ratio: 0.99
TimeDemo Play Ended, (3 Runs Performed)
================================================== ============

//////////// Summary \\\\\\\\\\\\\

10/15/2008 19:51:56 - Microsoft® Windows Vista™ Ultimate 64-bit Service Pack 1

DirectX 10 ENTHUSIAST 3X @ Map: airfield flythrough @ 8 1280 x 1024 AA 0xx
==> Framerate [ Min: 34.76 Max: 65.13 Avg: 52.24 ]


--------------------------------------------------------------------------------
Ambush was much more taxing on these settings. Lows around 24, highs around 70, but the average never got above 46.

[EastCoasthandle]

Originally Posted by Binge

Processor: Intel(R) Core(TM)2 Duo CPU E8600 @ 3.33GHz @ 4441 Mhz
CPU ID: Intel64 Family 6 Model 23 Stepping 10
Operating System: Microsoft® Windows Vista™ Ultimate 64-bit Service Pack 1
Physical memory: 4.00 GB
Display adapter: ATI Radeon HD 4870 X2 1024 MB
Driver version: Catalyst 8.10 (8.541-080923a-069997C-ATI)

================================================== ============
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 85.82s)
!TimeDemo Run 0 Finished.
Play Time: 41.09s, Average FPS: 48.68
Min FPS: 34.76 at frame 95, Max FPS: 62.72 at frame 1407
Average Tri/Sec: -49226368, Tri/Frame: -1011245
Recorded/Played Tris ratio: 0.97
!TimeDemo Run 1 Finished.
Play Time: 38.47s, Average FPS: 51.99
Min FPS: 34.76 at frame 95, Max FPS: 64.77 at frame 1987
Average Tri/Sec: -51027272, Tri/Frame: -981506
Recorded/Played Tris ratio: 1.00
!TimeDemo Run 2 Finished.
Play Time: 38.10s, Average FPS: 52.49
Min FPS: 34.76 at frame 95, Max FPS: 65.49 at frame 1995
Average Tri/Sec: -52311228, Tri/Frame: -996586
Recorded/Played Tris ratio: 0.99
TimeDemo Play Ended, (3 Runs Performed)
================================================== ============

//////////// Summary \\\\\\\\\\\\\

10/15/2008 19:51:56 - Microsoft® Windows Vista™ Ultimate 64-bit Service Pack 1

DirectX 10 ENTHUSIAST 3X @ Map: airfield flythrough @ 8 1280 x 1024 AA 0xx
==> Framerate [ Min: 34.76 Max: 65.13 Avg: 52.24 ]


--------------------------------------------------------------------------------
Ambush was much more taxing on these settings. Lows around 24, highs around 70, but the average never got above 46.

This is a prime example why the "results" portion of Framebuffer programs should be expandable when taking screenshots of the program. When you want to authenticate your results there is more credence towards that screenshot showing the very thing that is posted above.

[Mr.John]

Originally Posted by Binge

Processor: Intel(R) Core(TM)2 Duo CPU E8600 @ 3.33GHz @ 4441 Mhz
CPU ID: Intel64 Family 6 Model 23 Stepping 10
Operating System: Microsoft® Windows Vista™ Ultimate 64-bit Service Pack 1
Physical memory: 4.00 GB
Display adapter: ATI Radeon HD 4870 X2 1024 MB
Driver version: Catalyst 8.10 (8.541-080923a-069997C-ATI)

================================================== ============
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 85.82s)
!TimeDemo Run 0 Finished.
Play Time: 41.09s, Average FPS: 48.68
Min FPS: 34.76 at frame 95, Max FPS: 62.72 at frame 1407
Average Tri/Sec: -49226368, Tri/Frame: -1011245
Recorded/Played Tris ratio: 0.97
!TimeDemo Run 1 Finished.
Play Time: 38.47s, Average FPS: 51.99
Min FPS: 34.76 at frame 95, Max FPS: 64.77 at frame 1987
Average Tri/Sec: -51027272, Tri/Frame: -981506
Recorded/Played Tris ratio: 1.00
!TimeDemo Run 2 Finished.
Play Time: 38.10s, Average FPS: 52.49
Min FPS: 34.76 at frame 95, Max FPS: 65.49 at frame 1995
Average Tri/Sec: -52311228, Tri/Frame: -996586
Recorded/Played Tris ratio: 0.99
TimeDemo Play Ended, (3 Runs Performed)
================================================== ============

//////////// Summary \\\\\\\\\\\\\

10/15/2008 19:51:56 - Microsoft® Windows Vista™ Ultimate 64-bit Service Pack 1

DirectX 10 ENTHUSIAST 3X @ Map: airfield flythrough @ 8 1280 x 1024 AA 0xx
==> Framerate [ Min: 34.76 Max: 65.13 Avg: 52.24 ]


--------------------------------------------------------------------------------
Ambush was much more taxing on these settings. Lows around 24, highs around 70, but the average never got above 46.

Great improvement. A funny/strange behaviour I'm seeing on the ranked benchmark is NVIDIA doing very well on Flythrough benchmarks and performing worse on other timedemos which are more demanding as they simulate more normal gameplay (physics, particles, etc.).

Compare your results here

Open the link on IE, the scripts are a little buggy yet and don't work as expected on Firefox/Chrome/Opera.

Originally Posted by EastCoasthandle

This is a prime example why the "results" portion of Framebuffer programs should be expandable when taking screenshots of the program. When you want to authenticate your results there is more credence towards that screenshot showing the very thing that is posted above.

That should not be hard to implement