M17x Benchmarks thread !!!

When I had my quad @2.53 it had 2-3 FPS lower scores than 2.93. Basically, this is not that CPU dependent and Mandrake's scores are about 5-10FPS better than mine.

 

That's impressive. That's the highest I've seen on a notebook and should be playable too. And now you're only like 2% behind a W90 in DMC4 maxxed. To be fair to the W90 though, it had a huge CPU bottleneck. But factor in GPU overclocking for both systems and things should get really competitive.

 

Vicious's best score:

Scene1: 136.52 fps

Scene2: 95.75 fps

Scene3: 158.01 fps

Scene4: 100.31 fps


Mandrake's first score:

Scene1: 136.71 fps

Scene2: 97.71 fps

Scene3: 152.98 fps

Scene4: 104.81 fps


Who is behind whom?

 

They have ran Duals @ 3.0 ghz to 3.75 ghz.. they haven't had a CPU bottleneck in a while. Not after Quad got the PLL

 

New incentive to get the x9100, lol.

 

thanks for that info. i will look more into it.

so the w90 was stocked then. i like what i am seeing then. but i think the 280m can overclock higher then the 4870.(i know they are two different cards but you get what i mean?)

 

those w90 scores are the highest it gets and not sure if it was a stable system.

 

becaues ati focused on aa for the 4 series. it wins at 8aa and 2aa. 2aa actually inproves performance.

 

NO it can't AnotherGeek and the other 280 owners haven't been able to OC hardly @ all the 280 is runnign almost @ max on the clock Nvidia has been useing the same chip for so long. It looks like it finally reached it's max

It can't OC much, but it still can OC. You have to up the voltage to get a OC worth anything

 

No AA....

 

yeah i know it's at it's max but i'm not sure the stock clocks but i think the overclock is 680/1700/1000. that's maybe just straight benchmark and is not stable. but i know the 4870 is 550/850 and overdrive is 600/900. and then with vic's guide i think it's 700/950 is i am not mistaken. so one can correct me if i am wrong.

 

Dude, why do you keep saying ATI wins at AA. It is not winning anything.

 

My No AA scores:

Scene1: 180.82 fps

Scene2: 133.48 fps

Scene3: 215.60 fps

Scene4: 131.85 fps

 

you just killed a 280 stock you get repped for that.
http://img237.imageshack.us/img237/526/devilmaycry4trialdx1020qw0.jpg

 

he did not beat him in Scene 3

 

Beating one of the best dekstop GPUs out there. WOW!!!

 

See page 12 of this thread, and look at the 3rd set of numbers in the chart Vicious posted. Oddly enough, that was with the CPU stock. There was another maxxed out AA run that had much higher scores too, but not sure if it was running under Vista and Windows 7.

As for the AA argument, ATi has often dominated nVidia in this department (minus a few games). Here's a link/test which illustrates this:

http://www.firingsquad.com/hardware/sapphire_toxic_4850_dual_slot_review/default.asp

 

heh, if it can get to 700mhz core, that's well past desktop territory (and no other notebook GPU would touch it cept GDDR5 versions). Nonetheless, highest so far was 640 core and around 960 memory, which give or take is equal to a desktop 4850. And of course, there have been some overclocks on the 9800M GTX/G92 based mobile GPUs touching desktop 9800 GTX clocks. I think gettiing to desktop 9800 GTX+ would be much harder (if anyone could get their core to 738mhz), but never say never...

 

the highest i pushed my 9800m gtx is 670/1670/950. i haven't really tried to much higher since i don't have a cooler and sli but i haven't broke 80c yet.

 

bibsybob
no, the 4870's still have the record for 8x/4x.



but this was one of his regular runs with the same settings...

dmc4 demo benchmark at 1920x1080, everything set to super high, AA 8x, dx10, and got these results:
<o> > </o> >
Scene1: 136.52 fps
Scene2: 95.75 fps
Scene3: 158.01 fps
Scene4: 100.31 fps

 

everything looks even OC'd 9800m gtx, stock 4870, 280m stock for dmc4 at 8aa. can everyone agree unless we can get overclocked results for he 280m.

 

Mandrake can we get a screen shot of Crysis benchmark to prove to some people

 

Like I said a while ago, 9800GTX/4850/4870/260/280 are all in the same category, and it's relatively meaningless trying to make a competition of it. The mere fact that they are all so close is the most interesting part of this comparison... but if you have to know, hang up the towel and give it to the sli 280M, and wait for a real improvement in 2010.

And yes, I do find it hard to believe someone got 33 FPS @ DX10 Very High, WUXGA, stock GPU clocks. Call me pessimistic, but that's kind of impossible.

 

I am thinking about gatting one of these, but performance dosent seem to be a great leap from 9800, anyway when are the ATI chips coming out? And are the ATI chips based off of the real 4000 series? I would imagine if they were that they would walk all over the 280.

 

the clock speeds are to low to perform higher then them. it is dead even.

 

i certainly don't hope your comparing ddr5 4870's and 280's

 

yeah Ive been hearing a mixture on the 4870's for the M17x. Is it going to be GDDR5? 1GB VRAM? Higher clocks than the Asus W90?

also we need more people to post benchmarks! I want to see the difference in Crysis between SLI 260 and 280.

 

i'm not comparing the gddr5 because it's not released. i'm comparing 9800m gtx overclocked, 280m, 260m, and 4850(msi), 4870(asus 4850).

it's going to be 1-5fps second. overclocked i got 15fps, another geek got believe 18fps. and just add sli numbers to it.

 

Unfortunately it's 1920x1080(Mandrake's score) vs 1920x1200(dektop GTX 280).

 

it's not to far off. if he bring it up the only thing he will lose is scene 3 which is shaders.

 

Sure. Hope it's not too small.

 

yeah it is small. i'm using orca browser to zoom in on the picture and it's still blurry

 

Well considering a weakened 4870 (512mb, GDDR3) is still holding its own, it says a bit about having desktop derived cores. But at the same time, the GTX 280M is based off the desktop 9800 GTX+/GTS 250. Adding an extra 512megs of memory on top of adding GDDR5 to the current 4870 could add as much as 30% more performance without changing clocks, and at least 8-10% with settings that don't require a lot of memory.

 

Better?

 

yes much better thanks.

 

could you plz run the same with very high details please

 

Does it require tweaks like the description says?

 

hm do you mean to run in very high mode? i change my settings directly in the game options on very high and run the benchmark gpu bat(or exe) in the crysis director. i didnt use the other crysis benchmark tool.

 

Interesting. Sounds like I'm doing more than I have to.

 

this is what you do. you run the game and change your settings, then you use the benchmarking tool. the setting on the benchmarking tool doesn't work except for the res, dx9,dx10 and 64bit. the shadows,textures, etc won't change unless you change it in the game.

 

can someone try this. i don't think it applies to the m17x since you can disable the 9400m in the bios

http://forum.notebookreview.com/showpost.php?p=5012414&postcount=1429

 

Do you need to keep the game running change the settings then without closing run the benchmark tool? AA on or off?

 

i'm not sure with aa on or off. but no all you have to do is open the game, change your settings and then appy it. then you can exit it and then run the benchmarking tool.

just keep the settings on the benchmarking tool the same as the game so you don't get confused of your settings.

regarding aa however, i would change it in the game just to be sure.


also motion blur isn't in the benchamarking so make you have it on the same settings also.


i notice with the patches (motion blur) you willl get a big performance hit. so can you tell us what version of crysis are you running?

 

The graphics go so fast it's not easy to monitor. Changing to high settings in the game and turning on AA or leaving it off doesn't affect the numbers I'm getting. Still averaging 33 and maxing at 55. I'm up to patch 1.2.1.

Edit: I'll try to take a video of it tonight when there isn't so much glare.

 

i stand corrected as i just set everything to very high in crysis and then i set the benchmarking tool to high and got
6/27/2009 12:45:24 PM - Vista 64
Beginning Run #1 on Map-island, Demo-benchmark_gpu
DX10 1920x1080, AA=No AA, Vsync=Disabled, 32 bit test, FullScreen
Demo Loops=2, Time Of Day= 9
Global Game Quality: VeryHigh
==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 153.88s, Average FPS: 13.00
Min FPS: 9.53 at frame 1950, Max FPS: 15.31 at frame 998
Average Tri/Sec: -3986710, Tri/Frame: -306736
Recorded/Played Tris ratio: -2.99
!TimeDemo Run 1 Finished.
Play Time: 150.46s, Average FPS: 13.29
Min FPS: 9.53 at frame 1950, Max FPS: 15.47 at frame 988
Average Tri/Sec: -3881773, Tri/Frame: -292028
Recorded/Played Tris ratio: -3.14
TimeDemo Play Ended, (2 Runs Performed)
==============================================================

Completed All Tests

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

6/27/2009 12:45:24 PM - Vista 64

Run #1- DX10 1920x1080 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Last Average FPS: 13.29

high specs stocked
6/27/2009 12:51:49 PM - Vista 64
Beginning Run #1 on Map-island, Demo-benchmark_gpu
DX10 1920x1080, AA=No AA, Vsync=Disabled, 32 bit test, FullScreen
Demo Loops=2, Time Of Day= 9
Global Game Quality: High
==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 95.03s, Average FPS: 21.05
Min FPS: 15.86 at frame 1951, Max FPS: 26.49 at frame 990
Average Tri/Sec: -18272626, Tri/Frame: -868198
Recorded/Played Tris ratio: -1.06
!TimeDemo Run 1 Finished.
Play Time: 90.49s, Average FPS: 22.10
Min FPS: 15.86 at frame 1951, Max FPS: 27.12 at frame 1000
Average Tri/Sec: -18915258, Tri/Frame: -855818
Recorded/Played Tris ratio: -1.07
TimeDemo Play Ended, (2 Runs Performed)
==============================================================

Completed All Tests

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

6/27/2009 12:51:49 PM - Vista 64

Run #1- DX10 1920x1080 AA=No AA, 32 bit test, Quality: High ~~ Last Average FPS: 22.10




but i know in my whitebook when i did this, it wouldn't work unless i changed it in the game first. i am 1000% sure of this and that's why i always did it. i may have forgotten to install a framework or something

btw everything is stocked.

 

Do you always run it with 2 loops? I've been running all my tests at 1920x1080 based on recommendations here.

I'm running Vantage right now @ default settings.

 

oh yeah my bad i should also run 1920*1080. i'll do that and replace what i have.

i run 2 test because i don't feel like running 3 and 1 is not enough. anything can change while benchmarking. (windows update starts for no reason, antivirus decides to scan because it wants to annoy you. you get me?)

 

I use exclusions in my AV software.

I'll try the Crysis benchmark with 3 loops.

 

First run...

 

I must be blind. Where do you select how many loops?