Originally Posted by Wogga;14671657
i'll try 9-9-9-24 1600, 4400 HT-on, stock/725/790 with stock mem freq. and i didnt touched any settings in NV Panel so settings are stock too
add: there is a little hint (i think it is but who knows) - first 1-2 runs starts with very low fps, my record is 12 fps at benchmark startup. what i do is canceling bench after 20-30 sec and start over again - looks like some kind of warming the card =) (from idle 25C it could be some sort of coldbug)
I eagerly await your results. I'll go back to my old method (run heaven 1 pass just watching then bench) for Heaven. I also rolled-back my mobo BIOS which brought my pscore up ~200 points on average and my graphics score went up by about the same amount. It makes me wonder how much these older BIOS might be throttling performance if the newer ones throttle it like crazy. Of course, in the new BIOS half the features don;t work (incl. vcore, LLC, PWM) and this older one has the same bugs minus vcore and llc and runs the processor ~10C hotter. It also has the crazy boot cycling issue. Gigabyte needs to get their act in gear. Pronto.
Originally Posted by max883;14672129
Im runing my core 690 mem 3600. GPU1: 0.925v - GPU2: 0.913v
using nvidia driver 280.19
Do you have a verification URL?
Okay physics scores should be very similar at the same settings. Here's some food for thought concerning Gigabyte's BIOSes.
With the same BIOS, similar runs produce physics and GPU scores within ~50-75 points of each other most of the time.
I noticed drastic changes when I rolled-back to the F2 BIOS so I did a little "experiment."
System specifications in signature.
CPU-> 4.4GHz 1.296V Load 1.284V Idle
GPU-> 740/869/.975V Load .875V Idle
Programs running during testing-> Microsoft Security Essentials, 3DMark11, MSI Afterburner 2.2.0 Beta 5, nVIDIA settings (toolbar link), Intel Rapid Storage Technology, Realtek HD Audio Manager, Spybot-SD Resident, CoreTemp.
3DMark11 was run five times under each BIOS revision to eliminate 'one-off' scores. Each run was done after a clean boot and initiated only when both dies of the GPU had reached 31C stable and the processor was at 24C as read in EasyTune6 and TouchBIOS. CoreTemp read the idle at 36C. Therefore, CoreTemp is not an accurate measurement of real temperature, but does accurately represent the load-idle delta. 3DMark11 scores are represented as the average of five runs. Temperatures were taken after one run of 3DMark11, with the max of each core being logged by CoreTemp. The temperatures were then averaged across the five runs using the following formula->
RxAmax = (c0+c1+c2+c3)/4
Dx = RxAmax - 36
AmaxT = sum(R0Amax,+R4Amax)/5
DxT = sum(D0,D4)/5
-CPU Max Temps:
F2 BIOS-> 54.5C / Delta 18.5C
F3 BIOS-> 46.2C / Delta 10.2C
F2 BIOS-> P10670 / Gfx11187 / Phys9851 / Combo8734
F3 BIOS-> P10547 / Gfx10890 / Phys10186 / Combo8922
Stay far away from Gigabyte Z68 boards until they get their BIOS problems worked out. Also, BIOS does
have an effect on your scores! Don't let anyone tell you otherwise.
ETAx2: My OCD knows no bounds
Is there a Gigabyte rep on the boards?Edited by MKHunt - 8/22/11 at 11:12am