Lets see how important the CPU is, anyone with a 6600GT/s939/s754/intel rig wanna bench F.E.A.R MP Demo?

[dug777]

I'm running a 2.1GHz tbred-b (167x12.5)/1 gig hynix cheapo RAM (single channel of course)/ASUS A7V8X mobo/SB Live! 24-bit/380w Truepower/Sparkle 6600GT (570/1.145) rig here, and it would be very interesting to see how much the CPU and single channel RAM penalise me...

I suggest the benchmark in the MP Demo, 1024x768, 0xAA/8xAF, all settings maximum, so computer and graphics card on maximum, then go into graphics card and turn soft shadows off, set 0xAA/8xAF, with 'high quality' set in the forceware drivers. Run it three times then take an average for each score you report, if that makes sense?

Sounds ok?

EDIT:

Here are my results, exactly following what i said above

18.33 FPS min
38 FPS avg
95 FPS max

below 25 fps - 21.33%
25-40 fps - 47%
above 40 fps - 31.66%

EDIT2: that was done with my 2nd monitor enabled, but i doubt that would have had any noticeable impact. I used the 81.84 drivers.

EDIT4: using the 81.85s now, will post my scores soon as well.

Average of five runs ( i was bored ok )

19.2 FPS min
37 FPS avg
91.4 FPS max

below 25 fps - 20.4%
25-40 fps - 49.1%
above 40 fps - 30.2%


so a small (and very consistently repeatable) drop in the avg....

EDIT4: Now to compare with Inappropriate4AT, i clocked down to 500/900. (using the 81.85s so only comparable to my second set of scores above). Average of four runs.

18.75 FPS min
33.25 FPS avg
81.5 FPS max

below 25 fps - 28.5%
25-40 fps - 47.75%
above 40 fps - 23.75%

 

[dug777]

anyone?

 

[dug777]

someone at least?...it would be really interesting to see how much of a difference a fast CPU/memory subsystem helps...

 

[Sunrise089]

hoping someone can help you with this, I'm interesting in seeing the results

 

[n7]

I could run that with my X2 & X850XT PE, but i don't think that'd be a very fair comparison

 

[dug777]

Originally posted by: n7
I could run that with my X2 & X850XT PE, but i don't think that'd be a very fair comparison

not really

but yeah, cmon folks, i KNOW there are a whole truckload of you on here who have 6600GTs...it'll only take 15minutes maximum (including a fresh restart if you're into that sorta thing)...

 

[moonboy403]

i do have my 7800 gtx results in the other thread....lol

 

[dug777]

no one?

 

[martinez]

Link me to the demo and I'll do it for you.

I have Leadtek extreme 6600GT clocked 550/1120, 3500+ Clawhammer@stock(939), 2 gig value RAM in dual channel, 550W Antec Truepower, Seagate NCQ 120 gig HDD

 

[dug777]

http://3dgxfer6p.3dgamers.com/%5E305869...3dgamers/games/fear/fear_mpdemo_en.exe

awesome, that would be great martinez

 

[Inappropriate4AT]

570/1.145 is a pretty high OC, you should lower it to stock for a better comparison. I have a 6600gt, but I'm running it on an Intel setup, and there's no way I'm gonna get it to clock at 570/1.145.

 

[dug777]

Originally posted by: Inappropriate4AT
570/1.145 is a pretty high OC, you should lower it to stock for a better comparison. I have a 6600gt, but I'm running it on an Intel setup, and there's no way I'm gonna get it to clock at 570/1.145.

yeah i can do that, what would you like?

 

[Inappropriate4AT]

500/900 is stock for mine, I haven't played with it enough to find my max stable OC. I'm also runnin the 78.01's and too lazy to d/l other drivers. Want me to bench it anyways?

 

[martinez]

I'm on 81.84 atm too, forgot to mention that, so we should get a good comparison. There will be a bit of a delay while I dload that demo, as I have a torrent coming in and I'm only getting about 2KB a second :/. Bloody Optus DSL blows. But I shall get on that ASAP.

 

[Hacp]

Shouldn't it be 1280/1024 instead because we all know its less gpu limited at the other setting?

 

[Inappropriate4AT]

Intel Mobile Celeron @ 2.81ghz 187mhz fsb
Abit IS7 865PE, 350watt Rosewill generic, AV-710 sound
2x512mb Corsair pc3200 2.5-3-3-7 Dual Channel
Rosewill 6600GT @ 500/900 78.01

Settings as described:
14 minimum
26 average
59 maximum

50% below 25
39% 25-40
11% above 40

 

[dug777]

Originally posted by: Inappropriate4AT
Intel Mobile Celeron @ 2.81ghz 187mhz fsb
Abit IS7 865PE, 350watt Rosewill generic, AV-710 sound
2x512mb Corsair pc3200 2.5-3-3-7 Dual Channel
Rosewill 6600GT @ 500/900 78.01

Settings as described:
14 minimum
26 average
59 maximum

50% below 25
39% 25-40
11% above 40

cheers for that, i'll try at those speeds now

 

[dug777]

Originally posted by: Hacp
Shouldn't it be 1280/1024 instead because we all know its less gpu limited at the other setting?

thats the point.

 

[dug777]

Originally posted by: Inappropriate4AT
Intel Mobile Celeron @ 2.81ghz 187mhz fsb
Abit IS7 865PE, 350watt Rosewill generic, AV-710 sound
2x512mb Corsair pc3200 2.5-3-3-7 Dual Channel
Rosewill 6600GT @ 500/900 78.01

Settings as described:
14 minimum
26 average
59 maximum

50% below 25
39% 25-40
11% above 40

if you have time, here are the 81.85s.

http://nzone.com/object/nzone_downloads_winxp_2k_32bit_81.85.html

i did a 500/900 to compare, but i think there's a big difference between the 78.01s and 78.01s, since i can't imagine my platform is THAT much faster than yours

my scores: Now to compare with Inappropriate4AT, i clocked down to 500/900. (using the 81.85s). Average of four runs.

18.75 FPS min
33.25 FPS avg
81.5 FPS max

below 25 fps - 28.5%
25-40 fps - 47.75%
above 40 fps - 23.75%

 

[A554SS1N]

Wow... That's quite a difference between the Tbred at 2.1Ghz and the Celeron at 2.8Ghz, so this particular game is quite CPU bound. The important figure is the minimum framerate, as this is where all the jerkiness happens, adn the average seems most likely affected by the much lower maximum framerate. Very interesting to see. Hopefully more will post their results. If I had broadband I would download the demo as well...

 

[Hacp]

Originally posted by: dug777

Originally posted by: Hacp
Shouldn't it be 1280/1024 instead because we all know its less gpu limited at the other setting?

thats the point.

No, don't you really want to test how important the CPU is at settings you would normally play? To do this, 1280.1024 would be best. At that low of a resolution, the person with the better CPU would def get way better results, since the gpu isn't taxed at all.

 

[dug777]

Originally posted by: Hacp

Originally posted by: dug777

Originally posted by: Hacp
Shouldn't it be 1280/1024 instead because we all know its less gpu limited at the other setting?

thats the point.

No, don't you really want to test how important the CPU is at settings you would normally play? To do this, 1280.1024 would be best. At that low of a resolution, the person with the better CPU would def get way better results, since the gpu isn't taxed at all.

i play FEAR at 1024 normally (prefer the eyecandy to the higher res)

 

[A554SS1N]

Originally posted by: Hacp

Originally posted by: dug777

Originally posted by: Hacp
Shouldn't it be 1280/1024 instead because we all know its less gpu limited at the other setting?

thats the point.

No, don't you really want to test how important the CPU is at settings you would normally play? To do this, 1280.1024 would be best. At that low of a resolution, the person with the better CPU would def get way better results, since the gpu isn't taxed at all.

Putting it in higher resolution takes away the dependance on the CPU for framerate, which is what is being tested. The 6600GT is merely a common standard that is used to test the CPU usage (I.e. everyone tries same card with difference CPU's). It would thus in this case be better to lower the resolution to 640x480 but with all settings on highest to see the CPU effect, but 1024x768 will suffice. The results would then form a guide for those with different graphics cards on working out what their machine may perform like in the game CPU wise.

 

[Topweasel]

Originally posted by: A554SS1N

Originally posted by: Hacp

Originally posted by: dug777

Originally posted by: Hacp
Shouldn't it be 1280/1024 instead because we all know its less gpu limited at the other setting?

thats the point.

No, don't you really want to test how important the CPU is at settings you would normally play? To do this, 1280.1024 would be best. At that low of a resolution, the person with the better CPU would def get way better results, since the gpu isn't taxed at all.

Putting it in higher resolution takes away the dependance on the CPU for framerate, which is what is being tested. The 6600GT is merely a common standard that is used to test the CPU usage (I.e. everyone tries same card with difference CPU's). It would thus in this case be better to lower the resolution to 640x480 but with all settings on highest to see the CPU effect, but 1024x768 will suffice. The results would then form a guide for those with different graphics cards on working out what their machine may perform like in the game CPU wise.

It Might not be worth it, to explain to him. He is a smart guy but I remember when they were benching games on the first couple of DC Nvidia drivers he stated the same thing and it was explaine 20-30 times and he still did not get it. Again I am not trying to make fun of Hacp but this is one subject that he seems to not understand very well.

 

[Inappropriate4AT]

I'll give the 81.85's a whirl. Just woke up and have nothin better to do.