• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

**Official 3D Mark 11 scores**

Scores updated and thanks to all who did the test and posted the results. Well done to OptimaLrng for taking #1 spot. Those Lightnings are pumping some serious grunt.


8 Pack said:
Incorrect my benchmark scores I posted previously 13710.
Click to expand...

Great score but no proof that tesselation was on (as you supplied no link) and even HWbot requires you to submit a link. I was surprised your 3Dmark11 score with which you supplied a link was so far behind that 13710 score :confused:
 
So what is the average voltage for 1200MHz on the 7950's going to be.. can anybody give me ideas? Are the 7950 volts similar to the 7970's core volts?
 
MyBrains said:
So what is the average voltage for 1200MHz on the 7950's going to be.. can anybody give me ideas? Are the 7950 volts similar to the 7970's core volts?
Click to expand...

Can be totally card dependant, some have a higher stock voltage than others and clocks/voltages reflect this while overclocking. If you're looking for a starting point try 1.2v and see what clock speed you can get from there :)

On another note my new memory is here, hopefully bringing me some gains in the physics dept :p
 
In what way was your memory holding you back before pgi? I might be in a similar situation. Both my current and old CPU topped out at a particular point and will not boot no matter what volts go through it.

I failed to ever find what it was on the 2500k :(
 
With the 3820 the highest multiplier I can use is x43, but I have the advantage of using bclk straps to increase the base clock to 125/167/250 if I wish, increasing the base strap also clocks the memory (125 strap, x13.33 memory multiplier = 1666mhz). The memory i was using (16GB Vengeance, 1600mhz CL9) wouldn't clock for toffee without super slack timings or reducing the memory speed significantly, either way it was defeating the purpose of overclocking the CPU.

4.9ghz, ~1400mhz memory scored less than 4.75ghz with 1666mhz memory, even then the 1666mhz wasn't stable.

With this new memory I should (in theory) be able to manage 2400mhz with ease using something like 102.9 (base) x 1.25 (strap) x 18.66 (mem multiplier) = 2400mhz while also allowing me around the 5ghz mark on the CPU.

TL;DR: Vengeance sucks on X79, required an OC that it didn't like, high speed memory should work a thousand times better.
 
pgi947 said:
With the 3820 the highest multiplier I can use is x43, but I have the advantage of using bclk straps to increase the base clock to 125/167/250 if I wish, increasing the base strap also clocks the memory (125 strap, x13.33 memory multiplier = 1666mhz). The memory i was using (16GB Vengeance, 1600mhz CL9) wouldn't clock for toffee without super slack timings or reducing the memory speed significantly, either way it was defeating the purpose of overclocking the CPU.

4.9ghz, ~1400mhz memory scored less than 4.75ghz with 1666mhz memory, even then the 1666mhz wasn't stable.

With this new memory I should (in theory) be able to manage 2400mhz with ease using something like 102.9 (base) x 1.25 (strap) x 18.66 (mem multiplier) = 2400mhz while also allowing me around the 5ghz mark on the CPU.

TL;DR: Vengeance sucks on X79, required an OC that it didn't like, high speed memory should work a thousand times better.
Click to expand...

When do you recon you will have the new memory running.

Im quite keen to see how much difference going from 1600mhz to 2400mhz on the memory makes to your Physics score.

Will you be posting any before and after figures and pics with the CPU at the same speed but running the different speed ram setups.
 
Check one of my above scores (for physics) thats showing 42x (still had turbo enabled) with 1600mhz @ 9-9-11-1T.

I'll run the same clock speed again with the XMP profile for 2400mhz before I start tweaking the base for a higher clock :)

I'm going to install the memory just shortly, so results tonight providing I avoid disaster :p

Edit: floating around the 10450 mark for 42x @ 1600mhz.
 
Disaster narrowly averted, ******* compression fitting decide to turn loose under its own free will :mad:

Just going to run a test now and see what gains the memory gives.

Edit: about a 200 point physics gain from the memory: http://www.3dmark.com/3dm11/4129565

Not a clue whats going on with the GPU score (again!).
 
Last edited:
Edit: floating around the 10450 mark for 42x @ 1600mhz.
Click to expand...

For a bit fun I taking a guess at this (I have not looked at any 3820 scores apart from yours).

If you get the memory running @2400mhz with the CPU at your old gettings im guessing at a physics score of 10900.

If you then get the CPU to run @4900mhz with the memory @2400mhz im guessing at a physics score of 12720.

Just a guess lets see how wrong I am.
 
I have just seen your earlier post. Thats my prediction out the window. Have you tried tuning the memory that makes a difference. 8 Pack helped me with mine. What I found was the difference between putting the memory in and running them how they came @2400mhz and tuning them @2400mhz gained me about 140 points on the p score. This came solely from an increase in physics.

Good luck on your graphics score (I don't know anything about HD 7970s).
 
The memory certainly needs tuning, my motherboard is picking the XMP timings up as 11-13-13-2T when it should be 11-13-11-2T. Reviews for the memory had managed to tune this down to 10-11-10-1T - so that is where I'm aiming.

Also just had a lovely message from a failed overclock "Main BIOS corrupt, restoring from backup"

My luck is out tonight. Once I do get this towards 5ghz I too am hoping for around a 12K physics score.

BCCODE reported back as 'a'. Starting to think its just windows that doesn't like the OC rather than anything else. No debug code error on the motherboard when it fails either.

Edit: also getting 'CMOS Checksum Error' from the motherboard now, rolling back to an older non-beta bios.

Edit 2: Now windows required startup recovery, once recovered AMD drivers are no longer detecting my card! The only thing missing from this disaster of an evening is something bursting into flames.
 
Last edited:
pgi947 said:
Can be totally card dependant, some have a higher stock voltage than others and clocks/voltages reflect this while overclocking. If you're looking for a starting point try 1.2v and see what clock speed you can get from there :)

On another note my new memory is here, hopefully bringing me some gains in the physics dept :p
Click to expand...

Ok so i tried 1.2v and it seems like my core dont wana go above 1165 without artifacts starting to show.. (yes the artifacts are appearing when i clock up the core too much.. was told memory is the artifact culprit) but only in GPU Test 2. I left it running and got a slightly higher result from last time but i guess i'm hitting a wall with this now :( I'm also using the newly released Official 12.8 drivers that came out today which are looking a lot smoother in 3DMark11,before i noticed some slight stutter on the July 19 Beta's.

Heres my latest Score and probly the best i'm gona get out of this card unless i use Afterburner and up the volts on the memory.

Score
P9824 3DMarks

Graphics Score
9579
Physics Score
11694
Combined Score
9381

http://www.3dmark.com/3dm11/4129774

Usual 5GHz OC on the 2700K and RAM @ 1866MHz (9-10-9-27 2T) Might try 1T again and see if it makes a difference :p
 
You must log in or register to reply here.
Back
Top Bottom