9800M-GTS beats 8800-GTX reality or bollocks talk?

Better yet overclock the 8800m GTX to the 9800m GTS speeds. It´s not hard at all to surpass those stock clocks and see what you get. Best would of course be if the other user also had an extreme CPU, then we would definitely see other results in favour of the 8800m GTX. Now as Aeris said Crysis relies on a fast CPU the faster the better and higher framerates.

Operating System also counts in, Crysis in Vista 64 is faster than Crysis in Vista 32 for instance

Then of course each one has to run Crysis as it is original and not any tweaks that boost the FPS higher, that is if anyone actually used such a tweak in the first place.

 

For the fun of it I ran Crysis DX10 all High in Vista 32 on my desktop 8800GTX at 630/1530/1000 and at my max res of 1440x900 of my monitor and got this benchmark:

TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 61.21s, Average FPS: 32.67
Min FPS: 10.76 at frame 150, Max FPS: 51.20 at frame 989
Average Tri/Sec: -32010282, Tri/Frame: -979719
Recorded/Played Tris ratio: -0.94
!TimeDemo Run 1 Finished.
Play Time: 50.43s, Average FPS: 39.66
Min FPS: 10.76 at frame 150, Max FPS: 54.61 at frame 64
Average Tri/Sec: -38340700, Tri/Frame: -966675
Recorded/Played Tris ratio: -0.95
!TimeDemo Run 2 Finished.
Play Time: 49.31s, Average FPS: 40.56
Min FPS: 10.76 at frame 150, Max FPS: 54.61 at frame 64
Average Tri/Sec: -39196348, Tri/Frame: -966418
Recorded/Played Tris ratio: -0.95
!TimeDemo Run 3 Finished.
Play Time: 49.57s, Average FPS: 40.35
Min FPS: 10.76 at frame 150, Max FPS: 54.61 at frame 64
Average Tri/Sec: -38998232, Tri/Frame: -966604
Recorded/Played Tris ratio: -0.95
TimeDemo Play Ended, (4 Runs Performed)

Now it would be interesting to see what you would get at 1440x900 since you get 35 fps average at 1920x1200 which I can´t test on my desktop. Also do note this benchmark should be run without any tweaks/configs/Autoexec.cfg´s. Just original Crysis settings.

Since my 8800GTX has the double amount of Shader Processors compared to your 9800m GTS and my clock speeds on the Core and isn´t far away your overclocked 9800m GTS it would be really interesting to see.

I did run the benchmark just through the .bat file.

On the other hand I have 33 fps average at 1440x900 DX10 all Very High.

 

his clocks aren't stock. it's a vbios over clock.

 

Yes I know, stock core is 600.

 

ummm, who's stock core is 600?

kobe_24's stock clocks are 500/800/1250
that's who i was referring to.

and as for no enhancements and completely stock. this was that first run.

and leave the negative lod bias alone.

7811 600-799-1500 2.8 ghz

9/28/2008 10:28:40 PM - Vista
Beginning Run #1 on Map-island, Demo-benchmark_gpu
DX10 1280x1024, AA=No AA, Vsync=Disabled, 32 bit test, FullScreen
Demo Loops=1, Time Of Day= 9
Global Game Quality: High
==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 90.91s, Average FPS: 22.00
Min FPS: 14.38 at frame 151, Max FPS: 25.22 at frame 1753
Average Tri/Sec: 23043226, Tri/Frame: 1047419
Recorded/Played Tris ratio: 0.88
TimeDemo Play Ended, (1 Runs Performed)
==============================================================

Beginning Run #2 on Map-island, Demo-benchmark_gpu
DX10 1400x1050, AA=No AA, Vsync=Disabled, 32 bit test, FullScreen
Demo Loops=1, Time Of Day= 9
Global Game Quality: High
==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 92.19s, Average FPS: 21.69
Min FPS: 13.63 at frame 147, Max FPS: 25.70 at frame 1749
Average Tri/Sec: 22447970, Tri/Frame: 1034776
Recorded/Played Tris ratio: 0.89
TimeDemo Play Ended, (1 Runs Performed)
==============================================================

Beginning Run #3 on Map-island, Demo-benchmark_gpu
DX10 1680x1050, AA=No AA, Vsync=Disabled, 32 bit test, FullScreen
Demo Loops=1, Time Of Day= 9
Global Game Quality: High
==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 92.35s, Average FPS: 21.66
Min FPS: 13.99 at frame 147, Max FPS: 25.36 at frame 1746
Average Tri/Sec: 22405310, Tri/Frame: 1034561
Recorded/Played Tris ratio: 0.89
TimeDemo Play Ended, (1 Runs Performed)
==============================================================

Beginning Run #4 on Map-island, Demo-benchmark_gpu
DX10 1900x1200, AA=No AA, Vsync=Disabled, 32 bit test, FullScreen
Demo Loops=1, Time Of Day= 9
Global Game Quality: High
==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 100.65s, Average FPS: 19.87
Min FPS: 14.49 at frame 153, Max FPS: 23.07 at frame 1753
Average Tri/Sec: 19186234, Tri/Frame: 965550
Recorded/Played Tris ratio: 0.95
TimeDemo Play Ended, (1 Runs Performed)
==============================================================

Completed All Tests

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

9/28/2008 10:28:40 PM - Vista

Run #1- DX10 1280x1024 AA=No AA, 32 bit test, Quality: High ~~ Last Average FPS: 22.00

Run #2- DX10 1400x1050 AA=No AA, 32 bit test, Quality: High ~~ Last Average FPS: 21.69

Run #3- DX10 1680x1050 AA=No AA, 32 bit test, Quality: High ~~ Last Average FPS: 21.66

Run #4- DX10 1900x1200 AA=No AA, 32 bit test, Quality: High ~~ Last Average FPS: 19.87


so even if his sound like they are some pretty far numbers....what about these stock numbers?

now, if there is something wrong with the benchmark tool, then by all means let me know. but as far as who's version of what stock is suppose to be...is in the eye of the beholder.


gtx280: stock....

 

Ok..no tweaks, just rain via interface..64 bit:
9/30/2008 10:29:12 PM - Vista 64
Beginning Run #1 on Map-island, Demo-benchmark_gpu
DX10 1400x1050, 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: 59.74s, Average FPS: 33.48
Min FPS: 15.80 at frame 150, Max FPS: 40.99 at frame 1748
Average Tri/Sec: 28113388, Tri/Frame: 839732
Recorded/Played Tris ratio: 1.09
!TimeDemo Run 1 Finished.
Play Time: 53.62s, Average FPS: 37.30
Min FPS: 15.80 at frame 150, Max FPS: 41.73 at frame 993
Average Tri/Sec: 31676292, Tri/Frame: 849302
Recorded/Played Tris ratio: 1.08
!TimeDemo Run 2 Finished.
Play Time: 53.31s, Average FPS: 37.52
Min FPS: 15.80 at frame 150, Max FPS: 42.24 at frame 997
Average Tri/Sec: 31880090, Tri/Frame: 849784
Recorded/Played Tris ratio: 1.08
TimeDemo Play Ended, (3 Runs Performed)
==============================================================

Completed All Tests

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

9/30/2008 10:29:12 PM - Vista 64

Run #1- DX10 1400x1050 AA=No AA, 64 bit test, Quality: High ~~ Overall Average FPS: 37.41


And 32bit:
9/30/2008 10:35:07 PM - Vista 64
Beginning Run #1 on Map-island, Demo-benchmark_gpu
DX10 1400x1050, 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 0 Finished.
Play Time: 58.76s, Average FPS: 34.04
Min FPS: 22.43 at frame 159, Max FPS: 41.28 at frame 1770
Average Tri/Sec: 28603524, Tri/Frame: 840340
Recorded/Played Tris ratio: 1.09
!TimeDemo Run 1 Finished.
Play Time: 53.60s, Average FPS: 37.31
Min FPS: 22.43 at frame 159, Max FPS: 42.45 at frame 1768
Average Tri/Sec: 31705414, Tri/Frame: 849707
Recorded/Played Tris ratio: 1.08
!TimeDemo Run 2 Finished.
Play Time: 53.35s, Average FPS: 37.49
Min FPS: 22.43 at frame 159, Max FPS: 42.54 at frame 1775
Average Tri/Sec: 31872944, Tri/Frame: 850180
Recorded/Played Tris ratio: 1.08
TimeDemo Play Ended, (3 Runs Performed)
==============================================================

Completed All Tests

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

9/30/2008 10:35:07 PM - Vista 64

Run #1- DX10 1400x1050 AA=No AA, 32 bit test, Quality: High ~~ Overall Average FPS: 37.4

2 pics below:
64bit run..


32bit run:


Eh..not much difference.

 

My video card was overclocked when I did the benchmarks, and I can’t come close to 32FPS in Crysis DX10 1900x1200 all high. Even if I had an extreme processor I don’t think it would touch that, as it should only give me 1 or 2 FPS more.

3DMARK06 in Vista I get 11256 with GPU and CPU overclock, Crysis only went up to 19FPS average in Crysis at the same resolutions with both CPU and GPU overclocked. Both my SM scores in 3DMARK06 still stays above 5k.

NOTE: Using Crysis demo to test

 

i can't claim that title yet myself...only my sm 2.0 is in the 5k while my 3.0 is boarder 5k. if you click on the 11,115 (in my sig) you will see just how close.

 

I just have Vista for the hell of it, but I will be testing drivers to raise that score well pass the 11300 mark. I will also be testing drivers for that Crysis benchmark, to see what it does. There is nothing with reason that tells me you should be getting 32FPS average in Crysis 1900x1200 all high. Those are desktop like numbers, which should reflect into desktop like synthetic scores.

But hey, you posted what you got and that is that. I’ll see if there is a driver that will give me on average, a 10+ FPS raise. And even then, it would fall short of 32FPS.

 

hell!.. i can't touch E's score on any of them...lol

 

LOL!

I hear ya!

 

i thought i was doing something with my little old 26 frames...*LMAO* guess i was fooling myself...*LOL*

 

Whewt, I found one of my old 3DMark06's benchmarks with my Sager NP9262 (did it 1 day before the motherboard stopped working, cards were downclocking themselves and had to turn SLi on and off to get the cards to work, although the 2nd card's clocks were at roughly 200 MHz Core, 400 MHz Memory and 500 MHz Shader...

http://img385.imageshack.us/img385/4546/capturemt8.jpg

3DMark 06's Results (Fully-Working 1st Card, Failing 2nd Card):

12,836 3DMark's.

SM 2.0 Score: 4,840.

HDR / SM 3.0 Score: 6,066.

CPU Score: 3,857.

 

Which still tops my 19FPS. The thing that is stunning to me, is that your score doesn't drop too much when changing from lower to higher resolutions. Even when websites post their unofficial benchmark scores, the change from one resolution to another is drastic, like on mine. You're dropping only 5FPS going from 1280x1040 to 19001200, which doesn't make sense. Unless...Vista 64bit makes up for that, which I do happen to have installed. But hell, even desktops drop vital FPS when changing from low to extreme resolutions, which going form 1280x1040 to 1900x1200 is.

 

all them runs we're under vista 32 bit. i just through in my 64 bit drive to run some test now.

 

hummmmm, i think i found out why the scores are very different.

 

When one card is failing, it can mess up your total score.

 

ok, looks like your going to get a second chance to prove the 8800m gtx is better, because this benchmark has a major hole in it....*LMAO* so all involved will need to re run their test.

edit: let me know when your ready to start kobe_24

 

What do you mean? I'm using the demo version, so I can't tweak anything...besides, the very first test I posted was from the default benchmark setup found in the bin folder of Crysis.

 

Indeed. :/ I miss my laptop.

I will re-run 3DMark 06 and post my results with SLi Enabled, and SLi Disabled once I get my laptop back from Sager.

 

I will not get anything done right now, as I'm on battery power. I just happen to wake up, and see some might have thought I was using default video card settings, which I wasn't.

But que us in on what you've found?

 

has nothing to do with tweaking. has all to do with what the in game resolution is set to. i set the in game resolution to 800x600. then exited out then ran the benchmark at 1900x1200 and do you know what it gave me for results? an average of 45 frames. that's a big "hell to the nah!" *LOL* so i went back in the game and set it to all high and the in game resolution to 1900x1200 and this gave me my true scores. my 26 dropped to 24 frames....*LOL* and my 31 frames for the 1280x1024 dropped to 27.

so go test it both ways

edit: getting ready to run them all again and post the results.

 

I don't blame you, I would miss mine too!

Although my D810 is working fine and does great, I would miss the heck out of this one.

 

Hmmmm.....?

Now go into your bin folder and use the default benchmark, but set everything to high and 1900x1200 before you click on GPU benchmark?

 

Forget the Crysis tool, use the default benchmark. Just remember to set everything to the resolution you want before you start the test, ok?

 

I have to get up early, so I'm going back to sleep. I will be back on when I get up to see what has changed, goodnight!


Just remember you have a GPU and CPU test, so click on the appropriate one.

 

have a good one.

 

Wow!, that explains it all, thanks for sharing the information, johnksss, shall this help for more accurate scores and comparations.

Hehe, x3, the laptop I have to use is kinda like your D810, it is a Gateway MX3631m, with the following specs:

1.6 GHz Pentium M, Intel GMA 900, 512 MB of RAM at 400 MHz, 14" Display at 1280x768..., 80 GB at 4,200 RPM Hard Drive.

It feels so bad to use it, the chassis is falling apart, literally, after 2 years of heavy-duty but careful use.

 

it will sure level the playing field some what. see... you only get 1 run per resolution, but that run is in a series of 4. when doing the loop each loop your score gains a frame or two because it no longer has to wait to reload the scene.(speculation of course) doing it the other way means your over clock can't be outrages due to the fact that every run is concurrent and at a higher resolution each time. also it has to reload the scene. this should help keep scores allot closer....

it's still not full proof because if you dont keep it honest and run your res in game as the same res in the benchmark. you can pretty much get away with any frame rate you want.

 

I agree.

Good hypothesis, but I think that the loading of the scene is more Hard-Drive related than anything, maybe except for the textures loaded to the card, but, this is an hypothesis as well.

That is true, hopefully people will play sweep, but just in case, it'd be better to run the .bat benchmark included with Crysis until that the bug gets reported and fixed, that way, the scores cannot be modified at will so easily.

 

Okay, thanks for the info!, John.

Edit: Where'd your post go?! xD It just, poofed all of sudden, quick, the Men In Black are coming! x3

 

nah, it reloads from scratch... i forgot i had it set to run 3 per resolution.

edit: because i was looking at the wrong thing at the time of posting.

 

Okay Dokey! Thanks for the update!

 

I made a thread in this same "Gaming (Software and Graphics Cards)" forum about the bug in the Crysis Benchmark Tool and gave you proper credit as the discoverer of the bug, johnksss, so that the benchmarks get more realistic, requested and hopefully the thread will be stickied.

 

i'm not sure if im the first person to find this out.

 

That´s what I said the tool has a bug in it. I mean even me with a 1440x900 max res of my desktop computer can select 1920x1200 but the game still runs at my default res. However the tool outputs the same framerate I had at 1440x900 for the 1920x1200 res if you know what I mean.

So as I said this one is flawed. The real one, just the pure .bat file gives the right benchmark right away in 4 runs. Run that one instead and see what you get.

I were a little sucpicious myself since those framerates first off is near my 8800m GTX SLI scores and very near many high end desktop scores. Personally the 9800m GTS is basically an 9600m GT desktop and that card meanwhile good is nowhere near any of the higher end ones even with some overclocking.

 

Well, I’m sure it’s more than just that.

But luckily, it’s starting to hit home with johnksss a fellow overclocker that realizes some things isn’t what it may appear to be. But one thing should ring a bell, and that is something has to be wrong.

Anyone who has a WUXGA display can attest, that changing resolutions on Crysis (one of the most demanding games visually thus far), will affect game performance.

There is just no way to get 30FPS in Crysis all high 1280x10424 DX10, and then get 26FPS in Crysis all high DX10 1900x1200. That is only a 5FPS drop going from low resolution to extreme resolution in one of the most demanding games out right now. If that were the case, why then even bother to drop down in resolutions? There are about 4 higher resolutions to choose from after 1280x1024, which is still below 1900x1200. So that is like dropping only 1FPS per resolution change, which doesn’t make any sense what so ever. Not only would it not make sense, it is impossible for such a demanding game.

There is a big difference between Crysis high 1280x1040 and Crysis high 1900x1200, it’s not even close. So to almost get the same results in those resolutions shouldn’t be that close. We can point to my test for that. 1280x1040 (33FPS), 1680x1050 (25FPS) and 1900x1200 (18FPS). Those are major drops in performance, which is consistent with what it should be. So I go from 33FPS low, to 18FPS extreme and that is just with GPU overclocking. Add that with the fact that more shaders handle higher resolutions better, then something is slightly out of the norm.

I only overclocked the GPU to speeds everyone can get, not something out of reach to anyone. So anybody who gets the 88GTX/98GT should be able to get the same results. Overclocking the CPU and GPU does indeed affect Crysis results in lower resolutions. Since my FPS went from 33 to 38.
Still, 1900x1200 only resulted in a 1FPS increase.


Let’s face it. If a single card laptop can average 30FPS plus in Crysis all high 1900x1200 DX10 mode, the laptop has won. You put that in SLI mode, and you’re even going to be future proof for the next 10 years! Hell, E is claiming to achieve 34 and up on average, which destroys non tweaked and some tweaked SLI setups. If the 98GTX would average 35FPS when it came out, we would all give our right or left kidney for two of those. Desktop owners would throw away their beast to get a hold of two of those cards, to get the same performance if not better with the laptop. SLI would be getting well over 60FPS, and 3DMARK06 scores would be well over the 20k barrier.

 

Definitely, in SLI I see a huge improvement compared to a single card. To me 35 fps average in 1920x1200 DX10 high with a 9800m GTS with 64 SP´s is surreal considering that would even beat my 8800m GTX SLI which in fact totally destroys a single 9800m GTS without a sweat in performance and especially Dell´s SLI solution in the XPS M1730 which in fact is dual GPU´s on one PCB, sort of a GX2.

But who knows maybe he has the 280m GTX

 

That is true.

Yeah, the bug behind the Crysis Benchmark Tool has been unveiled, hopefully the game will keep fair.

Yes, going up to higher resolutions indeed decreases the frame-rate, in some cases, dramatically, such as in Crysis.

I agree on that, it is just not possible, 1920x1200 are a lot more of pixels that need to be computed and then filled compared to 1280x1024.

Yes, something had to be out of norm, it was just not possible that a card with higher clocks but less shaders could output such "olympic" framerates at 1920x1200 on Crysis.

Hehe, Crysis should be classified as "Bad Against Everyone", 1 FPS increasement on 1920x1200, at that resolution, that is an achievement to be proud of, and I am sure you are.

Lol, I agree, the 9800m GTS would become so popular Nvidia would rename it as the new "280 GTX", and Gateway would become a lot more richer than they are right now.

Yeah, SLi scales very well.

Yeah, it is like something they would air on "Ripley", and true, a 8800m GTX SLi could even use a 9800m GTS as a yo-yo; Dell's SLi solution is nice.

 

And Sager´s too SLI solution is really nice.

However when I ran 1440x900 max res on my desktop 8800GTX that is still a very fast GPU with 128 SP´s, double amount compared to a 9800m GTS and my clocks are set at 630/1530/1000 with a 384 Wide Bus then something is wrong considering I had an average fps of 40 at 1440x900. I mean my 8800GTX is basically faster than a standard 8800 Ultra, then this 9800m GTS is something out of the air beating even an Ultra in Crysis

1920x1200 I can´t even test since I don´t have such a monitor for my desktop. But I would assume my framerate would drop to around 30 fps average if I am lucky with a single 8800GTX.

Then how in the world can a 9800m GTS with 64 shaders output 35 fps average at 1920x1200, that is 5 fps lower average than me at 1440x900, tell me that, something if flawed here and it is not my 8800GTX at all with even a Core2Duo 2.93 GHz, 3+ GHz if I want though I don´t see any huge improvement in fps in Crysis at 3+ GHz Now I have said mine.

 

still messing with that benchmark and the whole thing is faulty. you can cheat using the .bat mode as well. **** shame.

ill post some shots in a bit.
but the 9800m gts is averaging 15 to 18 frames on high settings at 1920x1200
25 to 30 frames on high settings at 1280x1024

and as far as flubbing those scores...you would just set shaders to medium. then run your .bat file. which can push them up another 5 to ten frames from there. post your pick and non the wiser.


so far..the most reliable one is vantage mark, if you leave the geforce boost off. it's about as close to a real comparison your going to get. short of all of us in the same room testing under the same conditions. to many side variables to throw in. and everyone wants to show they got the better deal for what they spent. in some cases this could be true... in others...not even close. i even went so far as to take pictures of the screen with a digital camera to see if this would help, but that is even up for scrutiny as well.

tisk.. tisk... tisk...

and on a side note: overclocking is classed in 2 classes and not one.
1: class one is strait clocking for score and max clocks for a test run.
2: class two is all about style and how stable your system is (lower clocks with better performance is "better" not the other way around. and it's different clocks for different games.(technically, but who in their right mind is going to be switching clocks every different game and resolution they play? the person with the low end card, of course!. *LOL* j/k) hell...all this over clock stuff and i run my system with normal speeds at the end of the day. that is really why i bought it. to do what it's suppose to do at it's stock class. and everything after that..... is a plus.....

edit: i just ran 1 gtx280 & it scored 46.4 frames with no over clocking of gpu or cpu or ram (1440x1050)

 

Thanks.

Hehe, yes, it was like if the 9800m GTS had taken "Super Pills", and beat the 8800 GTX Ultra.

True, 1920x1200 is a... extreme resolution, a "poor" single 8800 GTX would let out some smoke trying to render that.

I agree, something had to be wrong for that to happen; well, depending on the CPU, it will boost the frames a little, since physics and AI (Crysis' AI is... well, exotic) can utterly slow down a game no matter what Graphics Card(s) you have, and the framerate goes up when those operations take a smaller time to do, thus, making the game look more fluid.

I wonder what framerate I'd get if I installed Crysis and ran the benchmark, and if I had my computer (almost on its way to Sager), Dual 8800m GTX's on SLi at Stock Speed, 2.66 GHz Q9450 Quad Core, if the day before my computer "died", I ran 3DMark 06, I had to turn SLi on and Off to make it work, but the second card was working at the slowest 2D Speed (301 MHz Core, 500 MHz Shader, 400 MHz Memory), and I got this 3DMark 06 result:

http://img385.imageshack.us/img385/4546/capturemt8.jpg

 

Aww, ugh, the .bat is faulty too... >.<, I guess we cannot use Crysis as a benchmark anymore now, thanks for the update, updating the Crysis Benchmark Bug thread...

Those framerates are way more realistic, thanks for the honest results once again, johnksss.

Ugh, I hope people do not begin exploit the bugs in both benchmark tools and begin to cheat everywhere.

Nice idea, 3DMark Vantage W/O GeForce Boost shall be it, then, that is true, good idea about the photographs, man, I guess I will validate my scores with photography if needed, but I doubt that would be needed because when the "Score" window of 3DMark comes out, it does not let you click anywhere else, or am I mistaken?

I support your ideology about overclocking, even though I do not overclock, because I want to preserve my Sager NP9262's life for as long as I can.

 

oh nah..not like that...*LOL*
i meant i was trying to find a full proof method for crysis benchmarch, but it can still be flubbed even with a picture. and i had everything up in there as well...*LOL*

 

Ah, alright then, re-updated the thread on the issue and corrected the error.

 

i think we have a few things going on at the same time... let me see what you changed.

edit: ok. running the .bat file is flawed. no way to say what the res your really running at. and you can change the log file as well. not to mention a few other things.

this is what i was talking about. meaning taking a picture of everything....

 

testing the warhead benchmark tool. this one seems to be a bit more on par with keeping the flubbed scores down to a min. but we shall see.... i only just started on it so give me about a day...

 

The .bat file is not faulty, just the person using it will have to be honest with what they are posting. You have to go in to system settings, and setup everything before you run the test. So if you want to run 1900x1200 all high DX10, then set it to that before you run the test.

Even better, use the demo with no tweaks…you will really see the differences between resolutions since you won’t have anything tweaked, to give good performance with great visuals.

I’m a straight shooter, I don’t have time to try and fool or trick someone in a test or game, as I’m only fooling myself. I know I talk some chit sometimes (Quaker humor), but I’ll definitely play fair.

We overclock things to make them somewhat fast. To overclock something and expect it to give not yet produced performance from even the next generation card and beyond is just ridiculous…as we’re trying to just overclock to reach the next level.

The 98GTS is a good card, no doubt! We see that with the scores you are posting. Nothing wrong with you posting low 1900x1200 all high DX10 scores, that is expected with single cards. The good thing is, you can run all high DX9 mode and get better results without losing much visual.

I want to apologize to johnksss, sorry bro!
But you can see why I thought that, right? This is all done in fun and I commend you for your honesty.

Oh…I took the 06 score to 11,290

 

Thanks for the update information again, john, you are really helpful and hopefully people will see what is really going on behind both of the Crysis Benchmarks.

Okay, nice picture, and nice framerate too, man, still, taking pictures is a really good idea to certify scores.

 

but of course. im all about getting to the bottom of things myself...my scores we're looking to be a bit to good, but i wasn't going to argue...*LOL* i ran them test like 30 times doing different things. then when i couldn't catch my man E. i was like...oh hell t the nah!. *LOL* then i ran my gtx280. then scouted the forums for a high performing 9800m gt/9800m gtx card. then started dissecting files. i played the game to see my scores. then started messing with internal settings. and low and behold...my score jumped up to 50 frames this time for 1900x1200. that's another big hell to the nah!.. . im not about flubbing the scores and i like to play fair and keep the field fair as well.

i found that running the game all high and drop shaders down one knotch gives you a pretty **** good picture and a playable framerate at 1900 res. better at 1680 though.


i hear ya kobe_24. and of course no hard feelings.
and if i wasn't a strait shooter, i wouldn't have never went back to find out the problem. i too know pretty close to how much these cards can go and what they can do with a range of different cpu's. and i was a bit amazed myself when i seen the first set of results! ya dammn skippy! still i was skeptical so i kept testing, but 36 frames at 1900 told me something is wrong. so instead of pointing fingers or getting bent out of shape...i investigate. i dont ask people what they do to over clock their systems anymore.(i can pretty much figure it out)(for dook- problem solving-*LOL*) use to do that along time ago and found that people will only give you about 75 - 85 percent of the whole answer. *LOL* so when i found the error and seen it works in bat mode as well as gui mode, i knew my scores we're not correct. so on to admitting that i was wrong. i have been wrong before. you just have to be a bigger man or woman to admit it.... so all i can do is run them correctly and post the results.

on another note.. stock 1920x1200 score
http://service.futuremark.com/resultAnalyzer.action?resultId=8364766&resultType=14