M17x Benchmarks thread !!!

I hate xp now, only reason i have it on one of my laptops is because of work, they will not allow anything other than xp right now, needless to say i tend to use one of my personal laptops instead

I really need to try and catch mandrake in vantage though

 

then what are you waiting for?????

 

true that...vantage is where it is AT

 

Lol were all gunning for that . Highest for me thus far is 11,543 . He wants someone to beat his score just as badly so he can try to beat his personal best .

 

well i keep pounding the F8 key when comp boots, but can't get it to go into the bios. keeps coming up with something called "Safe Mode".

or it's just waiting for my laziness to subside

 

My highest i've gotten proof of is 11,609 but i did hit 11745 once but failed to take a screenshot. "Slaps self"

 

It's called motivation....give me some!

Did you try out the 190.40 drivers yet?

 

i dont take pics of my own stuff till i actually have something worst posting .

 

That's why we submit online? haha

If you don't already own it, is only $20 for Vantage.....as much as you all like benching, being to cheap for $20 is pretty insane

 

I installed them last night before i passed out (long day) but ill be messing with them today.

 

Don't forget this also. This is not heavily OC'd. LINK

 

i own vantage . Thats another argument that would come up all the time since people spend $5000.00 on there notebook but wont pony up $20.00 because they dont feel the need to pay for benching software but its one of those arguments thats not winnable.

 

I'm legit!

 

well at the time i was the one holding the record in vantage while mandrake had it in 06, so proof was needed. Eventually mandrake surpassed me in vantage while i did it to him in 06. Mandrake and I for awhile was the only ones doing these tests.

 

As do I. EVGA was just giving them out left and right and I got 2 cards from them....laptop and desktop! haha

 

I already own

3dmark06
vantage
& pcmark vantage

 

Then why not submit whenever you have a good score? Instead of having to remember to take a screenie? You can hold and organize all your results online, is what I do.

edit: yay.....500 posts!

 

i don't beilieve i've even signed up for an orb account, i do hit the submit button but not sure if it's actually putting it up on orb, never figured out their damn site.

 

It will put it into orb but be hard to find again, just get an orb account. Just takes an email and is free and fast. Cant believe you don't have one yet!

You will be glad once you do. Makes it so much easier to check and show scores

 

aaah yes...the good old days....back when people talked badly about benching (seems that crowd is getting smaller and smaller as more options to over clock come out)...rotflmao. now we bench what ever we can. i love it! don't care what no one says. im an over clocker and benchmarker first and a sunday driver gamer next. (nothing personal to the hard core gamers)

and as quad has stated...these arguments already came up. people not wanting to spend 20 bucks on vantagemark. it got so retarted that i gave out my key to potential record breaking benchmarkers and my friends because i wanted them to have a fair advantage on the playing field.

and for the 3dmark06 xp thing...
if you look at his cpu...it's 5k. where as his other scores are far lower.
if either one of you runs it under xp, it will automatically take him, because your gpu scores will already be higher than his(speculation of course) or you can trick the system and run 06 in windows xp compatibility mode and see if that will do the trick.(more speculation though..i don't think it worked for me when i tried it)
and so far...havent seen anyone try what i said about the 2/4 core thing with a qx9300...but if someone did...what we're the results....

 

I know you mentioned something about the whole quad issue on nvidias getting the proverbial shaft but i cant remember exactly what it was and what you wanted someone to try.

 

quick test.
1: run test 1/2 & 5/6 on cores 2/3 only. (use the affinity method to run on those cores only)
2; run it with all four cores enabled

then post what the score difference is...if any with a qx9300. i know it was big with the qx9650

 

so this is really just a way to manipulate the OS scheduler to get higher scores?

if so, this gives even another reason to disregard 3DMark06 scores...haha

 

ummmmmm, let me think about that.....ummmmmmm NO.

that is a whole nother lesson there...hahaha...i remember seeing that as well.

speaking of dropping 3dmark06..the whitebook community was loving that benchmark. it was the king of kings benchmark for them. that was the talk of the town....so why you downing your benchmarking software? that's what made people take a second look at the white book?...lol

 

I just know that it (3DMark06) is not an accurate measure of a systems capabilities in this day and age

And has been proven to be non representative of the actual gaming capabilities of hardware, favors quad cores more than they influence FPS and hugely favors ATI over nvidia (thus the whitebook love) when this is not the case in actual game play.

And since I clearly misunderstand how or what would be happening with your proposed approach to the benchmark, can you please explain it some, even if just in the simplest of terms? It has me quite curious.

Thanks

 

3dmark06 usually favors dual cores not quads.

 

If that was true, than why did a QX9300 @ 3014 MHz get CPU score of 4500 vs an X9100 @ 3560 MHz getting a score of 3252.......

The higher clocked dual core still got a lower score

edit: that is my best 3DMark06 score with QX9300 vs. E-Wrecked best with X9100 (that is on ORB)

 

according to these people...3dmark counts... . whether we like it or not..
http://hwbot.org/rankings/worldrecords.

my 3dmark06 score
over clocked
http://service.futuremark.com/compare?3dm06=9859841
stock
http://service.futuremark.com/compare?3dm06=10340001
1:i have a 6500 for 2.0 pretty much for both.
2:my test 5 is the only one that made a drastic increase
3:and no one else can even get to 6k on test 2.0 with a 9800m gtx
well
that's because mine are running on two cores and the cpu part of the test is running on 4 cores. thus defeating the 3dmark06 quad core bug they have for nvidia.

and as kcissem stated...it favors dual cores over quads, but ati cards some how slip though this crack in the system. thus out scoring nvidia by a far margin. (like 1500 points or so)..well...that was untill i broke past that nonsense...then it wasn't so many parties jumping off in the whitebook community any more...lmao!i

http://service.futuremark.com/search/3dmark06.action?product=3dmark06.action&doSearch=true&page=&receiver=%2Fajax%2Fsearchresult3dmark06.action&cpuId=&cpuSpeedFrom=&cpuSpeedTo=&logicalProcessors=-100&physicalProcessors=-100&hyperthreading=-100&gpuId=470&gpucorespeedFrom=&gpucorespeedTo=&gpumemoryspeedFrom=&gpumemoryspeedTo=&gpuMemory=-100&shadermodel=-100&cooperativeGpus=-100&graphicsdriver=-100&operatingsystem=-100&filterDuplicates=false&email=&newestminorversion=1&benchmarkminorversion=1&scoretype=1&scoreFrom=&scoreTo=&settings=1&resolution=-100&filtering=-100&pixelshader=-100&vertexshader=-100&forcefullprecision=-100&hardwareshadows=-100&softwarefiltering=-100

out of that list...where are you at? E-wrecked is 2nd in the world. where are you at?
and he killed your gpu score in vanatgemark...so just what are you trying to get at?

 

I see, this had never been explained to me before and I was unaware of the difference between how nvidia and ati utilize a quad core.

Thanks!

 

I will add though, that both the scores i posted for CPU are on Arima W840DI chassis using 3870x2....so that doesn't FULLY explain the score differences.....

Unless you are saying that 3DMark06 only favors duals over quads WHEN nvidia cards are being used

 

no problem.
E and i tried it on the whitebook and it didn't work. that's because it wasn't effecting the whitebook or Arima W840DI chassis like it did on the 9262

 

but again, does that mean that this trick only works in conjunction with nvidia cards?

You are stating that the different chassis is why the difference when it is maybe simply the different GPUs (and drivers)?

if someone had a 9800m GTX they could throw into a whitebook that would be conclusive data for this

 

hummm, just who is going to spend 500 bucks to find that out?

and im stating that quad/9800m nivdia = low score 3dmark. nothing else.
i have also stated the best drivers for running 3dmark06 with a 9800m gtx

there is no such thing as one size fits all when benchmarking for speed vs benchmarking for games vs benchmarking for stability....

 

I'm assuming that applies only to Nvidia Sli, because my 2.0 score dropped over 500 points and 3.0 score dropped ~200 points when I set affinity to 2 cores. I'm not sure how you ran GPU tests with 2 cores and CPU tests with 4 cores all in 1 run either...

 

which two cores?

 

3 and 4...

 

has to be sli, forgot you dont have that.

 

Ok, re-ran with cores 2 and 3, SM 2.0 down 200 but SM 3.0 where it should be with the quad. CPU score still seems low at 2342 (low for a 2.93 dual core)

It's believeable. Does this only affect G92 or GT200 as well? Or perhaps it's an Sli/Quad core problem as a whole...

 

not really sure...that's why i kept asking different people to check it out and see...try to find some common ground with it.

yes, your cpu score will be low because it's running dual core mode. to get that all to work together you have to do more than just change the cores.

 

Well I know it's low for that reason, but it should be at least 2500 yeah? My P9600 was doing 2400. Either way, quad is still faster for single GPU runs.

 

makes allot of sense there..

and for you to get a real cpu read...you will need to run that test about 10 times...just to see what your highest read for cpu is. then do that with each driver to see which one is the best driver for cpu..then again for gpu. then start your testing...but then you already know why we say...benching ain't easy peasee...

here goes my single card run for 3dmark06
13,648 quad core mode
http://service.futuremark.com/resultComparison.action?compareResultId=10334595&compareResultType=14
12,349 dual core mode
http://service.futuremark.com/resultComparison.action?compareResultId=10417435&compareResultType=14

 

my extra .3mhz helped me out there

you gpu scores are higher then mines though.

 

I've got to try that driver, my best is 13681. Your SM2.0 score is a bit higher than mine, and you've got a slight CPU advantage. Those extra shaders on the 280M aren't doing much outside of the sea dragon GPU test.

here's the compare link of my second best run:

http://service.futuremark.com/resultComparison.action?compareResultId=11326421&compareResultType=14

 

can you take a screen shot and post. i don't like switching between tabs lol. i'll edit mines.

i have a higher one it's 12,689 but i covered the setting by mistake and i don't want people to think i cheated which i didn't so i didn't post. i forgot what driver i used at the time. i think 190.15 or 185.85 dox

 

best is in my sig. difference is, my best is with dox 182.46, second best is with 186.03, so the GPU scores are a little different. but overall only 8 point diff

 

here's my best

again the settings aren't shown that's why i didn't count it. it's stocked. but some people might not believe

 

I found a killer driver for 3dmark06, Dox's 182.05 (performance). This is stock GPU clocks, and already 13k. Problem is it's unstable at my normal benching speeds, so I'll have to play around with it, but I'm confident it will break records.

 

Those should actually work with the M17x. I'll try it when I get home.

 

i'm not sure if it does, i can't remember for sure since i tried a ton of drivers but i do believe the dox 182.05 was one of them and did not work correctly. Worth a shot though as i may be thinking wrong.

 

it's times like these that i wish i had a m17x...to get in there and help find the best possible answer for situations like this here....grrrrrrrrrr