official 3D MARK 11 Score Thread

[shadow001]

Originally Posted by Roadhog_

Yeah, thats the only way you can test the stability of DDR5 is to overclock and then see if performance drops. Though the gt580 has ddr5 rated at 5000mhz, who knows if its supplied enough voltage to run at that speed.

That's not the point anyhow....The point is that my score increased 600 points just by a driver change when everything else at the hardware level was exactly the same and running at the same clocks, that's the point..


You and Mospit were going on and on about a possible problem at the hardware level or configuration of the system a few posts ago and that my score was too low, and here i am knocking at 17k's door just from a video card driver change...

[moshpit]

Originally Posted by shadow001

You and Mospit were going on and on about a possible problem at the hardware level or configuration of the system a few posts ago and that my score was too low, and here i am knocking at 17k's door just from a video card driver change...

Proving we were right, you WERE low for score. We just assumed you were already smart enough to have tested drivers before posting scores. Our mistake

Our bad was assuming you wouldn't make such an inexperienced mistake as posting scores on drivers that were sub-standard

[Roadhog]

Originally Posted by shadow001

That's not the point anyhow....The point is that my score increased 600 points just by a driver change when everything else at the hardware level was exactly the same and running at the same clocks, that's the point..


You and Mospit were going on and on about a possible problem at the hardware level or configuration of the system a few posts ago and that my score was too low, and here i am knocking at 17k's door just from a video card driver change...

Not really, you posted results for a low score. We said it was a hardware issue since that is the most logical guess with the info given. Then you posted a new score with your memory at a different clock speed. So I assumed you found that was the problem and didn't know how to overclock DDR5. Then you posted your final result with apples to apples with just a driver change, and your score is still low for your system...

[shadow001]

Originally Posted by moshpit

Proving we were right, you WERE low for score. We just assumed you were already smart enough to have tested drivers before posting scores. Our mistake

These are the latest betas released just a couple of days ago, so i used the most recent approved drivers as far as futuremark is concerned, wich were released about a month ago....


I'm sure Demowhc will also experience an increase in performance, maybe break into the 17k club once he uses these same beta drivers, does that also mean he didn't test drivers too, he used the last official ones too afterall just like i did...


The other thing that was proven is that video card memory speed doesn't matter at the performance setting in 3Dmark 11....

[moshpit]

Originally Posted by Roadhog_

Not really, you posted results for a low score. We said it was a hardware issue since that is the most logical guess with the info given.

Totally agreed. We wrongly assumed you'd have already TESTED several drivers before beginning posting results for the best one, or at least stated you were on an outdated or otherwise untested driver. Our bad for thinking you knew what you were doing, Mr. Shadow

Sorry, had to call ya Mr. Shadow. Just watched 5th Element

[shadow001]

Originally Posted by Roadhog_

Not really, you posted results for a low score. We said it was a hardware issue since that is the most logical guess with the info given. Then you posted a new score with your memory at a different clock speed. So I assumed you found that was the problem and didn't know how to overclock DDR5. Then you posted your final result with apples to apples with just a driver change, and your score is still low for your system...

Roadhog... You're full of ****.


You automatically assume a hardware problem right off the bat, and what info did i give to suggest that....That's right, nothing.


I showed just now that GPU memory speed makes no difference at the performance setting, and you went off on another tangent completely, stating about maximum attainable speed with DDR5...


And according to 3Dmarks database, wich has a few thousand entries already, i'm right in line with where i should be given the clocks i'm running and they have several SR-2 entries in the database, so i know this will come as a shock to you, but i believe them more than i do you...

[Roadhog]

Originally Posted by shadow001

Roadhog... You're full of ****.


You automatically assume a hardware problem right off the bat, and what info did i give to suggest that....That's right, nothing.

I assumed a hardware problem because I assumed that you had the software/driver side down. But guess not.

Originally Posted by shadow001

I showed just now that GPU memory speed makes no difference at the performance setting, and you went off on another tangent completely, stating about maximum attainable speed with DDR5...

That you did. But I was just giving info on overclocking DDR5 as a lot of people don't know about the ECC it has. No need to get butthurt. Was just trying to help

Originally Posted by shadow001

And according to 3Dmarks database, wich has a few thousand entries already, i'm right in line with where i should be given the clocks i'm running and they have several SR-2 entries in the database, so i know this will come as a shock to you, but i believe them more than i do you...

Sure, I believe them too, but you are also getting beat by single processor systems with lower clocked gpus. If it were me I would be going wtf.

[shadow001]

Originally Posted by moshpit

Totally agreed. We wrongly assumed you'd have already TESTED several drivers before beginning posting results for the best one, or at least stated you were on an outdated or otherwise untested driver. Our bad for thinking you knew what you were doing, Mr. Shadow

Sorry, had to call ya Mr. Shadow. Just watched 5th Element

They were the latest Nvidia WHQL drivers and approved by futuremark....You gotta do better than that moshpit.


And here's another kicker, once i use the same clocks as demowhc on the GPU's, i'll also be in the 17K club...

[shadow001]

Originally Posted by Roadhog_

I assumed a hardware problem because I assumed that you had the software/driver side down. But guess not.

Demowhc used the same drivers and i didn't see you ragging on his ass...

Sure, I believe them too, but you are also getting beat by single processor systems with lower clocked gpus. If it were me I would be going wtf.

And you also forget that they're all running at higher clocks than me, not lower....Show me one, please?

[Roadhog]

Originally Posted by shadow001

Demowhc used the same drivers and i didn't see you ragging on his ass...






And you also forget that they're all running at higher clocks than me, not lower....Show me one, please?

http://3dmark.com/3dm11/1222238

[moshpit]

Originally Posted by Roadhog_

http://3dmark.com/3dm11/1222238

Significantly lower clocks, in fact. 850mhz.

[shadow001]

Originally Posted by Roadhog_

http://3dmark.com/3dm11/1222238

I'm 600 points higher, and he's still using his CPU clocked 200Mhz faster?...Wich matters a lot at the performance setting

I don't know why he's clocking his memory to 2200Mhz, but his cards are clocked lower on the GPU side by 50Mhz?


What were you proving again?...

[Roadhog]

Originally Posted by shadow001

I'm 600 points higher, and he's still using his CPU clocked 200Mhz faster?

I don't know why he's clocking his memory to 2200Mhz, but his cards are clocked lower on the GPU side by 50Mhz?


What were you proving again?...

My point is that your score is coming from the PHYSICS benchmarks. But in the GRAPHICS benchmarks, yours is really slow imo.

[moshpit]

Originally Posted by shadow001

I'm 600 points higher, and he's still using his CPU clocked 200Mhz faster?...Wich matters a lot at the performance setting

I don't know why he's clocking his memory to 2200Mhz, but his cards are clocked lower on the GPU side by 50Mhz?


What were you proving again?...

Just curious, what was that second socket for again?

[shadow001]

Originally Posted by Roadhog_

My point is that your score is coming from the PHYSICS benchmarks. But in the GRAPHICS benchmarks, yours is really slow imo.

Horse****....I gained several hundred points just by moving my CPU from 3.8 to 4.0 Ghz in the first 4 graphic test runs....They don't use all the cores, but still very sensitive to clock speeds just the same.

[Roadhog]

Originally Posted by shadow001

Horse****....I gained several hundred points just by moving my CPU from 3.8 to 4.0 Ghz in the first 4 graphic test runs....

lets see it?

[shadow001]

Originally Posted by moshpit

Just curious, what was that second socket for again?

If the benchmark used all the cores in all the tests, including the graphics tests, it would be game over for any single socket setup and you know it...

[moshpit]

I think "really slow" is an exaggeration. He's about 1000 points off the average for graphics. That's not really slow. That's a little slow in my book, but possibly chipset driver related for as minor of a problem as it seems.

[moshpit]

Originally Posted by shadow001

If the benchmark used all the cores in all the tests, including the graphics tests, it would be game over for any single socket setup and you know it...

That's my whole point. "If" doesn't even count in horseshoes or hand grenades.

Edit: Keep in mind, horseshoes and hand grenades are very tolerant of not hitting the mark.

[Roadhog]

Originally Posted by moshpit

I think "really slow" is an exaggeration. He's about 1000 points off the average for graphics. That's not really slow. That's a little slow in my book, but possibly chipset driver related for as minor of a problem as it seems.

for graphics score though. Gaining a few fps on those is a huge difference.

[shadow001]

Originally Posted by Roadhog_

lets see it?

This one is at 3.7 Ghz but you get an idea just the same:

[moshpit]

Originally Posted by Roadhog_

for graphics score though. Gaining a few fps on those is a huge difference.

True. But I've seen a bit of play from one chipset to the next, and this isn't technically an X58 system, it's a server chipset. Intel may not have optimized the PCIe controller in the same way, more for stability then speed (another HUGE reason to avoid dual socket rigs for a home machine).

[shadow001]

Originally Posted by moshpit

I think "really slow" is an exaggeration. He's about 1000 points off the average for graphics. That's not really slow. That's a little slow in my book, but possibly chipset driver related for as minor of a problem as it seems.

I'm matched with demowhc pretty much in that last run, and he's running 30Mhz faster on the GPU's and using a CPU clocked 600Mhz faster.....

[Roadhog]

Originally Posted by shadow001

This one is at 3.7 Ghz but you get an idea just the same:


http://s765.photobucket.com/albums/x...37ghzpart1.png

isnt that on the broken drivers? I do lose fps from 3.7 to 4ghz, but same fps from 4ghz to 4.5ghz.

[moshpit]

Originally Posted by shadow001

I'm matched with demowhc pretty much in that last run

Are ya? Then seems like "a little low" was more accurate, as I thought. You're not having any problems that cannot be easily blown off as chipset driver level optimizations. Minor stuff, and obviously easy enough to offset.

[shadow001]

Originally Posted by Roadhog_

isnt that on the broken drivers? I do lose fps from 3.7 to 4ghz, but same fps from 4ghz to 4.5ghz.

They're the official WHQL release from Nvidia....The 270.61 release.

Make another excuse please....

[Roadhog]

Originally Posted by shadow001

They're the official WHQL release from Nvidia....The 270.61 release.

Make another excuse please....

facts are not excuses.

[shadow001]

Originally Posted by moshpit

Are ya? Then seems like "a little low" was more accurate, as I thought. You're not having any problems that cannot be easily blown off as chipset driver level optimizations. Minor stuff, and obviously easy enough to offset.

More like going into the 17k club simply by matching his GPU clocks ...

[shadow001]

Originally Posted by Roadhog_

facts are not excuses.

Then go bust Nvidia's balls as to why a beta driver is much faster than their latest WHQL driver set, not mine....

[Roadhog]

Originally Posted by shadow001

Then go bust Nvidia's balls as to why a beta driver is much faster than their latest WHQL driver set, not mine....

i will.