sham, for your opinion this bios is faster tha 61j for OC'ing?

44 was bad..

firerx

Hey Shammy,
I'm like the new bios.

Thanks,

glad to hear tat!

Balthazor

Just wanted to point out that I also saw my two 5970s exhibit performance degradation in both the 44 and 49 BIOSes, with all other settings being the same.  And, just as with the 44, once I went back to S61J my benchmark scores are back where they should be.
 
Mlippold:  What application were using for comparison when testing your 5970s?  I used 3DMark Vantage.

Yep.   Same, 3DMark Vantage.   I'm running 10.3pre drivers on Win7 64bit.  

Moltenlava

Alex2190

yeah i think alot of the issues stem from people not knowing what settings they should run with...and when they flash they forget to change things back...

Yup...

this was not the case for me though

No the Intel SATA Controller was set to RAID, I reset it to RAID otherwise my RAID array would no come back online.  All other settings i went through thoroughly like i always do. 

to alex2190
You really should look at the person before making assumptions, possibly some new guy would make a mistake like that...

dont get mad at me because your board doesnt work bro, even the most veteran people forget the stupidest crap. sometimes you double check and say "damnit, i missed that one thing.." turns out that one thing is AHCI or something like that

Orici

EVGA_JacobF

• Sets ICH Voltage to 1.1v as Default

Why did you do this?

i believe it helps with the USB bsod and other issues.

i think they found that some southbridge chips were poorly binned, so they needed a touch more voltage to be stable. i naturally bump my voltage on the on board chips by a bit to be stable...never had any issues

nullcorp

Crap. I just received a 5970 and really wanted to do my first i7 overclock with that new bios.

What bios would you guys recommend for an E759 ?

I recommend you get baseline scores at stock speeds and frequencies using the S61J and this 49 BIOS, observe what memory divider you get with 49 (because it seems to vary, with or without Turbo Performance), and compare.  Use whichever BIOS gives you the best results.  A single 5970 should get very close to a 11000 GPU score on the Extreme benchmark (at stock frequencies); if you see Extreme GPU scores in the sub-10250 range then this is lower than expected.  These are just rules of thumb, and you really have to test for yourself to determine your own baseline for comparison.
 
I'm not sure why the 44 and 49 BIOSes degrade performance with dual 5970s, and I haven't tested a single card to troubleshoot.  All settings were the same when comparing BIOSes.

fyi. my 3dmark06 score dropped over 10% with this bios upgrade. all original bios settings verifed. my old 8800 and i need all i can sqweeze out of it for now. something in this release is bottlenecking the video. back to the older code for me.
 
benchmark this bios, evga guys, please!
 
920 &  E758

Orici

Alex2190

Orici

EVGA_JacobF

• Sets ICH Voltage to 1.1v as Default

Why did you do this?

i believe it helps with the USB bsod and other issues.

Does it help my HDD's that are getting some bad arttributes in the SMART readings ?

ID Cur Wor Thr RawValues(6) Attribute Name
03 197 196 _21 000000000C4E Spin-Up Time
C6 100 253 __0 000000000000 Uncorrectable Sector Count
C8 100 253 __0 000000000000 Write Error Rate

Have you got your HD set for ACHI? If you do then go into bios a change ACHI to RAID. This will not affect the drive configuation. Reboot then go back into bios. Go back to where you set RAID. You'll see AHCI on very top. Hit Enter. You'll see setting for Spinup. Make sure it's disabled.

Same problems persist on this release, I'll be sticking with S61J for the 760:
 
Problem #1
PCIe x16 Slot #2 and #4 still drop PCIe bandwidth to x8 for Slot #1 and Slot #3.  Just need a confirmation at this point whether this is a known bug or working as intended and the FAQ on PCIe bandwidth is incorrect and needs to be corrected or not:

http://www.evga.com/support/faq/afmmain.aspx?faqid=58513

E760 - X58 SLI Classified

PCIe slot 1: 16x (8x if slot 2 is filled*)
PCIe slot 2: 8x
PCIe slot 3: 16x (8x if slot 4 is filled*)
PCIe slot 4: 8x
 
*not affected when running 1x PCIe cards like a sound card or Killer NIC

Again, this is incorrect as an x1 card in PCIe slot 2 drops PCIe slot 1 bandwidth down to 8x as seen in the picture below.  Verified with S61D, S61J and now BIOS 44:



Again, just need a simple confirmation whether this is something that's a bug or working as intended so I can stop wasting time checking it every BIOS release and look for alternative solutions, like custom chipset cooling.  Ideally, EVGA could release a Classified North Bridge cooler without the bump preventing larger x1 cards in the top x1 slot, or sell Classified coolers so we can modify it ourselves.

Problem #2
Cas Latency Memory Timings after resume from S3 Sleep are incorrect.  Timings are set 8-8-8-24 in BIOS and work correctly until resume from S3 Sleep, the timings change to 19-8-8-24.  Seems as if there were some problems with regard to the Turbo Performance option and RAM timings so this may be related.  This doesn't happen with S61J which I've reverted back to now.

 

2 poor BIOS releases in a row, as expected, consumer confidence wanes with every such release.

After years of playing around with new technology I am not surprised that there are bugs with every new bios release. Would be nice if once these bugs become apparent a new beta hotfix to be released. Anyhow on my 759 I have also noticed and confirmed the cas latency changing after coming out of S3 sleep in Win7 on bios 49 from '8' to '11'.  The only thing that is overclocked right now is my memory controller at 1600 (memory rated at 1600 actually). Turbo and VT and CXe all turned off on a 920 D0 Stock frequency. At least Everest reports my Pci-e lanes operating at 16x. But dear Evga can we please have a working bios and put this problem to bed...

Thank you.


Edit: Got to say at least they're trying and releasing frequent updates compared to other companies.

My CAS# will go from 9 to 11 when coming out of S3.  This is down from the SZ1C Bios where it used to go to 18 so I guess it is an improvement.   I see no impact on benchmarks or stability of this.

yes.  i am not disagreeing that it is problematic and should be fixed as the problem has been around for a long time.

however, i think the S3 state in general sucks, especially for us rev1.0 owners who can't take advantage of nice overclocks.  Anything over 190blck and S3 outright doesn't work.

KrzyInuYasha

Well my HDDs and I thank EVGA for getting rid of that damn power off/on after 5 sec on every system reset.

Is that so bad for the HDDs?

lam_das

Orici

KrzyInuYasha

Well my HDDs and I thank EVGA for getting rid of that damn power off/on after 5 sec on every system reset.

Is that so bad for the HDDs?

Yes it is. Head parking and power down for HDDs should not occur during a reset. Its not required and it increases wear and tear on mechanical HDD. If you are lucky and have SSDs, you can ignore this.

OK, so this BIOS fixes that but VT isent working?

I don't think i need VT, so it's safe to flash then?

rodymcamp

well this bios has been working great for me so far. I just tested the s3 issue about the cas latency and mine goes from 7 to 11, pretty weird. I rarely use s3 sense it has never worked right on my 759 classified, but every thing else seems to work.

Are you running Auto vcore or did you manually set your vcore?