*OFFICIAL* M17x Benchmark Thread - Part 3

total, no one cares about just the cpu.

 

i think your thinking of last year...this year... everyone cares about cpu since they are i7's now.

 

only some of them can pull that, i am currently benching mine and so far i got this
http://img203.imageshack.us/img203/1034/bench4.png
using the bios options, i increased the ddr3 voltage to 1.7v and the overclocking frequency to 5%.
I am currently using the evga eleet utility, and gonna check what i get. do you want me to test anything?

 

Nah, 3dmark vantage cpu is stupid. I pull over 55k with my desktop gtx 280's running physixs. 3dmark06 is the cpu one, where a massive cpu OC can really affect your score. But a massive OC on vantage doesn't matter as much as your GPU.

 

think you had better go take a look at the world class benchmarks before making a statement like that. where some only over clock just the cpu and pretty much killed off quite a few people with over clocked cpus and gpus..(vantage) vantage is geared for both.

edit:
defining AS in over clocking cpu.... for it to make a substantial difference...you need to be in the 30k plus range on cpu or 5ghz over clock...anything under that...and it's normal scaling of cpu to gpu.

 

Go check out xtremesystems forums, people care about both no doubt, because both affect your score, but I the majority of people work on increasing their GPU score as much as possible, because it affects the score the most.

In our laptops people might care about CPU score, because there just isn't the powerful GPU's desktops have. But because Vantage CPU can be manipulated by physixs cards, I don't think its all that fancy. Hell, my desktop vantage with physics is like 26k, and without phsyics its something like 23 or 24k. So whoop But my OC'ing my video cards was awesome in Vantage.

 

How do the 260m gtx compare to 280m gtx? like a crisis benchmark, or 3dmark 06... Thank you...

 

The CPU score with physx enabled is MEANINGLESS

If there is a big spread between the cpu or gpu scores in vantage, which ever is lower will more heavily influence the overall score.

 

i was ranked in the top 20 for 5970 and gtx 295 quad and single..which means nothing.(and thats out of 1000+ people benching)
i had the highest over clocked 5970 and gtx295...again..means nothing. when someone comes by with his cpu over clocked to 5.7ghz and gpus running stock and still beating you.

dont get me wrong..i work more on over clocking gpus than cpu, because that's just how i started. so i understand where your coming from, but it's best to keep an open mind as to what can and can't be accomplished in the world of benching.

and for the record..take yur dt system and go bench any one of the earlier benchmarks...bet you can't even come close to these guys benching with single and dual core chips..
i know, because i tried....why?...because the earlier benchmarks we're based mainly on how fast your cpu is running...not how many cores you have.

speaking of benching:
anyone start benching
2001se
2003
2005
yet?

 

Show me where your overclocked 5970 but stock cpu gets beat by a stock 5970 and overclocked CPU. I guess anything is possible though but I would like to see it.

And yes, earlier benchmarks were CPU oriented which is why 3dmark06 is stupid, because it relied so heavily on the CPU. There is no dbout that my 920 @ 3.8ghz will get worked by a 980 at 5.7ghz in 06, but in vantage i will have a much closer overall score if my GPU and GPU OC is exactly the same, and the only difference is 3.8 quad to 5.7 six core.

 

Anybody ?

 

If 2 systems have the same gpu score and radically different cpu scores....the overall score will also be radically different

 

Correct, but much more so with 06 and earlier, than with Vantage.

 

I think you are a little out of your element here.....you are arguing with one of the top laptop benchers in the world and one of the top desktop benchers in the world

 

@barrok and John -

We're done here. Take it to PM if you want to continue it.

Thanks.

 

This thread is specific to M17x Benchmark results, questions, settings, etc.

This is not a thread to discuss CPUs or GPUs in other systems or benchmark results from other systems.

Stay on topic. You want to discuss other systems, head over to the hardware forum or relevant system thread.

Thanks.

 

Is anyone on here OC the 4870 with software and 10.3a beta drivers? If yes, how?

Thanks in advance. Sorry if it's been discussed before.

 

i dont think the driver matters, you can use rivatuner to overclock the cards.

 

I tried but I can't seem to get it to work... Very strange... How did you get it to work???

 

i just followed jokers instructions, and when you run rivatuner, there is a small arrow, once clicked it will show you a couple of icons, and you select the first one. what doesnt work? did you add your graphics card to the cfg file?

 

I know what your referring to, but for some reason I can't get the first icon which allows for clocking???? Very strange. I did add the 945a to the cfg file too.

 

Don't bother with the rivatuner method. Upgrade your drivers to 10.3a and use AMD GPU Tool (5870 version) to software OC. Alternatively, you can use RBE + ATi Flash Tool to OC via bios.

 

i would like to see some benchmarks for the R1 between the GTX280s and the ATIs on the Metro 2033.

i can play it on very high and 1900x1200 (AA disabled though) with more than 30 frames most of the time but fps can really go down to 12 frames.

it is widely recognized as a heavy game for a few days now since its released. something like crysis.

 

I have been playing Bad Company 2 - that is a game that needs a built in benchmark. Gorgeous!

1920*1200 all High, 1x AA, 8x AF 30-35 fps (SP)

 

battlefield with 1AA is ok. try metro 2033 with AA off and feedback..

 

i have metro 2033 with 2x aa with all very high.. using dx 10 and enabled the in game stats counter using the config file cos fraps dint seem to work with it.

got constant 40 fpses during walking around etc but drops to as low as 14-20 when monsters appear.

im on atis cat 10.3

 

Where are the 10.3a drivers? Should i uninstall the old ones?

 

Thanks Joker!!! Also, I noticed that I now have GPU acceleration in Photoshop CS4 since upgrading to 10.3a! Just in case anyone cares.

 

One more thing. Joker what settings do you use for your system? I have the Diamond Thermal Compound applied to both CPU and GPU's. Just curious.

 

Thanks for the heads up. Good enough reason for me to update.

 

Here are the specs:

2.0GHz Intel Core2 Quad Q9000 Processor, 4GB System Memory, 320GB+640GB Hard Drives; Dual Layer DVD Burner, 17.0" Display(1080p); Dual NVIDIA GeForce GTX 260M - SLI Enabled, Windows 7 Home Premium Operating System (64-bit)

What could i expect from this system that i just bought? after i OC it? 3dmark 06 etc...

 

Here is a score for the top 5 mandrake. its 19,548 3dmarks
http://img29.imageshack.us/img29/2088/71717703.png
i got up to 27 mutlipliers, cpu overclock from the bios 5%, 1.7 voltage, vcore 150mv. why is it such a low score compared to yours? and why did you stop benching?

 

Is there a way to make the AMD GPU Clock Tool apply clocks after reboot or do you have to manually do it every time?

 

I am not sure whether its possible, but i wouldn't recommednd doing that, its going to be bad for your graphics card. just use higher clocks when you need them the most, and then disable them.

 

Here is a shot of my clocks at boot using the same bios settings as you.
Does it look ok?

http://img502.imageshack.us/img502/7555/cjclocks.png

 

Everything seems to be okay. Try running 3dmark06, and see how much you get. Thats where i knew there was a problem with my system.

 

Why does your Eleet say multipliers at 9 and core speed only in the 1K area? Mine is hitting around 3.7-3.8K

Running 3DMark now. First run was in the mid 18's now running with clocks up to 27 like you.

Edit: OK Just reran and got 19100 which is a bit closer to your score. Why is the test only pulling from one core though? Is it supposed to do that?

 

The m17x uses the amount it needs at that time. and then drops to low clocks, when it doesn't need them.
My first run was about mid range 18K, about the same i guess. 19,100 is normal, but i would have expected you to get better scores seeing as you've got 8gb of ram.
I've got a few questions, did you use the bios overclocking features with eleet utility? What bios are you using?

 

A02 and yes the Eleet bios clocker.
How do you get the clocks to fluctuate when needed? Edit:nevermind power settings.
What do you put min processor state at? Mine is now at 55%
I'm using the Diamond Thermal Paste too! Just applied it last night.
Do you need to reapply the eleet clocks after every reboot?
Oh yeah, I'm also using a CryoLX.

 

For high performance which is what i use for benchmarking, i put it at 100%.
I currently installed windows 7 on my second hard drive. i jsut used the drivers i needed , i did not install anything else other then 3dmark06 and vantage. How hot do your temps get when benchmarking? run real temp while benchmarking. You do need to reapply the clocks everytime you reboot.
whats a CryloLX?

You might want to try increasing your graphics card clocks when benchmarking.

 

Gotcha...
I'll check my temps later tonight. Thanks for the quick responses.

CryoLX - http://www.newegg.com/Product/Produ...ries+-+Notebooks/Tablet+PCs-_-NZXT-_-34997340

 

good luck in benchmarking! Don't get too addicted

 

When benching, the amount of RAM isn't going to make a differance. It's pretty much all CPU/GPU dependent

 

Bleh, okay I'm having a crazy problem with my M17x. I set all of the settings to High Performance (Windows and the Alienware Command Center) and both of the GPUs are enabled in SLI but I'm only getting a score of 7,418 in 3dmark06 and just as bad in 3dmark05. Any idea what could be causing this? GPU-z reads the clocks to be correct as well and Powermizer is disabled.

CPU score: 2,932
SM 2.0 score: 2,914
SM 3.0 score: 3,033

Specs:
Intel Core 2 Quad Q9000 @ 2.0 Ghz
GTX 260M SLI @ Stock
4GB DDR3 1066 Mhz RAM
320 GB 7,200 rpm HDD
1440x900 resolution LCD

 

stealth mode on perhaps?

 

I had the same problem with my m17x r2, turns out the screws for the cpu were loose. I'm not saying that is you're problem, but put the CPU under pressure and check whether it usesthe frequency it is set at. It also might be that stealth mode is enabled, like the guy above me suggested.

 

Well Fn+F7 disables Stealth Mode doesn't it? It was enabled when I first received the laptop I believe, but at that point neither GPU even showed up in device manager, only the 9400M intergrated. After I hit Fn+F7 the computer restarted and both GPUs appeared in device manager. As I said I checked GPU-z and they're both reading to be running at stock speed (550/950/1350 I think). I could check CPU-z but the CPU score seems accurate, it's the SM 2.0 and 3.0 that seem lacking. Isn't 2,932 accurate for a stock Q9000? I've also checked the BIOS and the settings are all correct there also.

 

3dMark06 is cpu dependant, so if your cpu is weak or something is wrong with it your going to get a low score. Download HWinfo32 and check the temp of your cpu temps . If your temps are normal try downloading the latest drivers.

 

CPU-z reads the core speed to be 2000.0 Mhz and using HwMonitor I got these temps at this very moment, seems about right.
Core 0: 38C
Core 1: 40C
Core 2: 37C
Core 3: 37C
Also I've already tried using stock drivers, 195.62 from nVidia and Dox's 195.62

 

i dont think those temps are high. on the contrary.