*OFFICIAL* M17x Benchmarks RESULTS *Part 2*

why are you running 1920*1080 and not 1920*1200.

i just fresh installed my machine but i dont think i'm ready to do anything. i just posted something up in the stutter thread.

 

John told me to lol

Just let me know what you all want me to run. I am doing stock runs

 

version 1.05 is the correct version to have.
is crysis patched to 1.1 or 1.2 or original?

 

1.21

That is the most recent official patch I believe

 

i hope you know i'm coming for you. in what ever you run. i'm just trying to get some stuff together so it will be when everything is said and done.

 

might be...not sure if it's the best one for you to bench with..

well look at don!..go get'em tiger...grrrrrrrrrrrrrrrrr...lol

 

Hmmm, how does one uninstall a patch?

Nevermind, just going to reinstall the game

 

allot of 1080p's floating around...that's why i asked him to do both

 

That game is SLOW to load up....I am at version 1.00 now

in the interest of saving time, what version should I try out John? I am not trying to prove my benching abilities here, so just tell me what to do lol

Just want to see what these cards can do

 

i say...1.1
at least..it was for me...

 

Well so far, 1.2 is definitely better than 1.1

and 1.0 was very unstable

all the runs so far are still around 27.5 FPS though for stock clocks

1920x1200, No AA, Very High, DX10, 64-bit

 

but that's all stock right?

 

yes

but someone else was allegedly getting 31 fps stock at same settings and that has me concerned

 

And WOW, one hell of a jump from Very High to just High lol

27 fps upto 45 fps - did not realize there would be THAT much of a gap. Going to Medium was only 50 fps......

 

I was talking about stock, Scook. And the guy who did the benchmarks did the tests with GPU at stock clocks, and two runs with his CPU at stock and then with 3.2 GHz OC. His average still stayed around 31. This is why I am asking whether there may have been some kind of error in his benchmark, or vice-versa. By the way, upgrading to 1.1--at least--will help performance.

BTW:
There's a reason why high to very high yields such a gap. A large range of post-processing effects that come into play and severely tax the system. One such effect that only works best in DX10 is your 'per-object motionblur'. This is why I am sometimes a bit put off by these newer cards, because sure, when you run something like Crysis at high, they'll get amazing results in the 40+ margin. Older 8800s could barely cut 30. But, when you kick the stuff into very high, that difference narrows down with these cards getting anywhere between 27-30, and 8800s getting between 19-22.

 

veryhigh is for bragging.
game looks great at all high as far as laptops go.

i remember that now...do you have the link to his post?


kade storm:
that's why i asked him what tool he was using along with versions
so if he over clocks the gpu's..he'll gain just about 5 fps. making it about 36 fps...if it will complete...(speculation of course)
and about 60 fps at high

 

I was planning on playing the game at high, but it was an eye opener for me lol

big delta there - and I know the game is already beautiful at all high

 

I actually like very high, it looks much better if people appreciate good lighting. Not to mention, it truly determines the grunt capabilities of a card. In the M1730 forums, we have Magnus72, who specifically created a DX9 config that captures DX10 effects and allows for smooth gameplay in the 40s provided that you are running a well managed dual 8800m GTX setup. Edit: If you're going to play Crysis at high, then lose the DX10, because you're not taking advantage of any DX10 features, and so the API will only slow your frame rates by a certain margin. High - DX9 does just as good a job, and with better frame rates.

And Johnksss, I guess you got a point there, but then, why not just use the benchmark bat file that is in the binary folder?

 

ah yes...magnus is a good friend...he and i did many different types of runs together.

And Johnksss, I guess you got a point there, but then, why not just use the benchmark bat file that is in the binary folder? <---we always did. it was the new gen bencher that wanted to use the gui interfaced one. i used them all to rule out cheating amounst some of our more...over zealous members who use to frequent the benchmarking scene...lol. with these out rages claims of fps with mobile hardware beating high end desktop gpus.

 

well I found that all settings at High actually was 2 fps better with dx10 instead of dx9

 

. . .I read a few of those 'claims' back in the days of the 9800m GT SLi.

 

Wait till you hit the ice levels and where more texture is required to be streamed into your system. You won't always notice a drop in frame rates, but there might be moments where a fresh scene will play and the character's textures will still be loading and popping about.

I know this game's engine quite a bit better than the average gamer. And I know that your minimum FPS--which is lower in DX10--in this game will be something that will rear its ugly head at the most decent spots.

In benchmarks, you won't see much of a difference when it comes to your average. Even in the first levels, you won't see much of a difference.

But, there will be many parts, scenes and intensive gameplay points where that 'texture streaming' will be burdened for no reason under DX10, since no DX10-exclusive feature is being used anyway.

Note: I have also done runs with DX10 that show the API perform faster than DX9, but those have been under explicit conditions and usually when texture streaming is deactivated.

And the fundamental question: Why bother with a DX10 run, when you're not using a single DX10 feature?

 

Just did a set of runs under that .bat file

it seemed to be about .5 fps faster than the GUI for what that is worth

 

Excellent. I think we should just focus on the bat file tests. Keeps things a bit more, simple.

 

Scook,
I have attached a file to this message. It's a Crysis custom configuration file.

Maybe you could give it a shot and share your average benchmark rates. Above all that, you can give some feedback on your gaming experience, because its real purpose is to accomplish decent performance in-game with most of the DX10 bells and whistles. If you're bored, give it a DX10 and DX9 run.

Extract the contents of the zip file into your '. . ./Crysis/Game' folder.

 

Not sure how to run that file. Also, should it be in the game\config folder with the other .cfg files?

 

i can run crysis right now if you want i just want to know what patch you are using and what settings.

 

all the stuff I posted here was with benchmark tool 1.05 and version 1.21 of the game. All stock clocks

Do DX10, Very High, No AA, 1920x1200, 64 bit

This is a game that is known to do better on nvidia hardware so this will be interesting

 

i have to patch it up. i'm running 186.64 right now.

 

feel free to give different drivers a go

Never hunted for the best crysis drivers when I was using nvidia

 

Is it possible to run Benchmarks with Crysis Warhead (Steam)? Are they comparable to the orginal Crysis

 

it keeps on loading the game.

 

Yes, put it in the config folder with the other files.

 

and then how do I select it to run?

 

Sorry. Should've mentioned this earlier. It's an autoexec.cfg file, dude. It'll automatically load when you start the game or the benchmark. Give it a shot.

 

Find the relevant exe and set it to load as administrator. Should sort you out, otherwise you just keep loading the first level. Had the same problem til I worked that out recently.

 

Whats the latest/greatest Driver for the 4870-X2?

 

Cat 9.12 with hotfix

 

why thankyou rengsey for the speedy reply..
I have the cat 9.12..whats the hotfix?

 

http://support.amd.com/us/kbarticles/Pages/ATICatalyst912Hotfix.aspx

There is a discreption of the hot fix and also a link to download it.

 

Thanks.. so is everyone selling their R1 for R2's?
Any Idea on the expected 3dmark 06 performance jump with the i7-920xm over what u had in the R1? Just a ballpark figure would do from who ever has an idea..(speculation welcome ofcourse)..

 

Basically I will have the same spec on my R2 like my previous R1 with the exception of the CPU. I read somewhere that it's 25-26% increase in performace. Base on my OC'ed QX9300 , OC'ed 4870's the Vantage and 3DMARK06 score from my R1 , I think R2 with 920Xm would probably be around P15K - 17K Vantage , and 24K - 25K 3DMARK06. heheehe

 

lol lets hope those predictions hold true.. If they do I'd be forced to join the revolution (I thought I did just that 2 months ago)..this is the shortest lived "new" laptop ownership ever :|

 

You are too optimistic. Your CPU will stop at 5k in 3D06, compared to 4.2-5.2k of the QX9300 (depending on the OC' level, 4.2k~3-3.2GHz, 5.2k~3.8-3.9GHz), and the GPUs will stay the same. So you will be lucky to get 14-15k Vantage and 18-20k in 3Dmark06.
But dreaming is a good thing!
I dream about i9-980X + 5870CF (All OC'ed to HELL!) 17-19k vantage and 25k 3Dmark06

 

i think that sounds pretty close

 

if you bench 3dmark06 under windows xp...your scores will go up.
i use to think that it was old and out dated, but that is not the case.
you bench the correct os for the job..and xp seems to be the better choice for most when it comes to benching 2001/2003/2005/2006 and vantage being vista or windows 7

 

Indeed. It was so easy breaking some of those 3mark06 14k and 15k barriers using XP. Maybe I should've tried for 16K when I had the 9800m GTX.

 

i wana see the big improvements for real-time gaming.(if there is any)

 

i would post something but my system is horible right now.

 

Maybe you shouldn't have updated to A03. I've been telling people that most of the improvements--at least for Nvidia users--have been superficial in nature.