Just Benchmarked Crysis With 4850's

Just did a couple of runs with the crysis benchmark tool.

One at 1920x1200 on high DX 10 which came back as min15 max 39 and average of 27.8

If anyone wants me to do any different configurations just let me know.

 

Nice Marshall, now give Very high a shot. You dont have your CPU in your sig ?

 

You're kidding me.

Bloody hell, now I REALLY want those cards. Considering I've only got a 1440x900 screen (yes im p00r) I'm looking at playing any game out there on max graphics and 1440x900.

Presumably that benchmark was at 0xAA? Or was AA on?

 

The latest 2 sets of ATI drivers have like a 15% crysis boost in them too, so make sure you are using new drivers if possible.

I was able to get like 40fps average I think all maxed out on 1920x1080 DX9 but in DX10 my frame rate went to hell.

I think Quadzilla had some really good scores too in Crysis while he still had the W90.

 

Ya my max at 1920x1080 DX 10 very high was 29fps. Don has the screenshot around somewhere and i know its still in the W90 forums .

 

Don has ever screen shot or benchmark ever posted

 

Right just running again at 1920x1200 on very high DX 10 64bit and 32bit. Will post results in 5 mins.

 

I'd like to see what it is at DX9, as there's not much of a difference between DX9 and DX10.

 

1920x1200 On High And DX 9:

Min 19fps
Max 41.9fps
Ave 29.65

1920x1200 On Very High And DX 9

Min 18.7
Max 40.54
Ave 28.97

Interesting there is hardly and difference between high and very high. Gonna start overclocking the GPU's and see how things go.

 

Which driver are you using?

Nice results!

 

1920x1200 Very high avg 29fps, that is amazing.

 

So would a p8600 at 3.0 and a pair of 260 get above 20fps on very high dx10 1200p?

 

Am using the standard ATI 9.7 just modded with MobModder.

 

why not load windows 7 then use un modded ATI drivers, it is good performing and very stable (vs instability I always had with modded drivers)

 

Why not use dx10?

 

Hmmm, the Crysis benchmarking tool doesnt seem to like me; it just loads a level and doesn't actually benchmark!

I'll happily post some Crysis scores from Vista if I can get the damned thing to work!

 

juat for comparison here is the m980 doing crysis

did you have AA on? i had 8xAA on this one

http://forum.notebookreview.com/attachment.php?attachmentid=37007&d=1249658794

 

we need that in the m17x benchmark thread

since that is what your m980nu is really up against

 

Was AA definately working? I seem to remember a while back people complained that 8xAA wasn't working in DX10, and AA/Crysis tends to get low results even on the high end desktop cards?

Good result with or without though

Could you run again with 2/4x AA to see whether the results hold up?

 

Because DX10 has very little graphical difference from DX9 and you still have to pay in FPS.

It's the same for me, it just loads the starting level if I ever try it I've no idea how to get it working.

 

here's quad's w90 crysis run

 

Most shocking stat....2:02:28 AM

 

some people work nights and/or stay up

 

Lol. i didn't notice that.
what did you get with this setting?

 

I'll need to hunt that down.

 

Lol 2:00am is late ? That was 3 days after my Bday and the W90 was a gift to myself so i was having some fun .

Oh and back then the GPUs could only overclock 50mhz on each so that was basically stock . Cant imagine what it would get now since they made some real gains on overclocking on that system with the GPU.

 

When you're used to waking up at 3:30am it's very late. I'm half asleep right now.

 

oh ok. so you were 600/900?

 

Yep exactly.

 

Lol well for someone who has to wake up at 3:30am i guess it would late lol but then again if i had to wake up that early i would not be awake right now either .

 

i still would be

 

If i were 20 years old i would to . Dont worry though you will find out one day .

 

I'm 21 this month so i think i'm starting to know what you mean.

 

i hate subway sometimes i work from 5- 1am then the next day i work 10-5



sorry for offtopc

 

it's ok. i go off topic all the time.
here's a ramdon benchmark sinlge 3870 crysis high no aa 1920*1200
so we have the 4870x2 from the w90 and the 3870 now we need the 4850 and the 280m
Crysis
2/20/2009 10:35:19 PM - Vista 64
Beginning Run #1 on Map-island, Demo-benchmark_gpu
DX10 1900x1200, AA=No AA, Vsync=Disabled, 64 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 0 Finished.
Play Time: 124.63s, Average FPS: 16.05
Min FPS: 6.84 at frame 1943, Max FPS: 19.22 at frame 991
Average Tri/Sec: -15140137, Tri/Frame: -943451
Recorded/Played Tris ratio: -0.97
!TimeDemo Run 1 Finished.
Play Time: 115.14s, Average FPS: 17.37
Min FPS: 6.84 at frame 1943, Max FPS: 19.32 at frame 997
Average Tri/Sec: -16150893, Tri/Frame: -929844
Recorded/Played Tris ratio: -0.99
!TimeDemo Run 2 Finished.
Play Time: 114.74s, Average FPS: 17.43
Min FPS: 6.84 at frame 1943, Max FPS: 19.42 at frame 983
Average Tri/Sec: -16220077, Tri/Frame: -930550
Recorded/Played Tris ratio: -0.99
TimeDemo Play Ended, (3 Runs Performed)
==============================================================

Completed All Tests

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

2/20/2009 10:35:19 PM - Vista 64

Run #1- DX10 1900x1200 AA=No AA, 64 bit test, Quality: High ~~ Overall Average FPS: 17.4

 

they ARE NOT 4870's

it makes me angry @ the stupidity of the nameing ...

 

Lol lets not even get into that again .

 

If you call GTX280m as GTX280m why not call W90's 4870x2 as 4870x2?

lalalalala

 

i don't say 4870x2 i say 4870's in Xfire

 

Not a big deal unh?.

After all... X is the multiplication sign

 

Lol. i didn't know the x2 would cause people to get mad. x2 is when there is master/slave. even though i know it's flashed master/master to work right.

and it's shorter.
4870x2 4870's 4870 crossfire 4870 xfire 4870 xfired

 

4870 CF hehe

 

NO

4870x2 is when there is 2 GPU/dies on onePCB

research don

 

as far as drivers are concerned, just needs to be master/slave gpu configuration actually....look at the 3870's in M17/whitebook

master/master gives the NON x2 name

 

They just need to end the mobile naming scheme all together based on the desktop cards To bad average joe consumer wouldnt know what the heck they are buying otherwise. On the other hand that person who thinks they are getting the desktop card in there laptop will be very disappointed.

 

yes yes and yes i agree