Thanks, this is an amazing chip. Bought it from a fellow EVGA member Dejanh in the EVGA Market place. He advertised it as a golden CPU batch number and claimed it did 4.2ghz @ 1.25 Vcore (which it did!)

Correction! I know your father!  Asante sana squash banana!
 
 

 How do you go about getting higher Gflops? Could ram speed be an issue?

Hey Kram ol buddy ol pal. When are you going to put my new results up. I want to be 1st for Bloomfield
If anyone tries to beat me, i got 4.6ghz and 4.7ghz on the back burner just waiting to be unleashed.

Damn...I need ~1.6v for 4.6ghz  ..its too warm out now

1.35 on the IOH PLL??? The default is 1.8V! I don't think it changes the temps at all so I would bring them back up. Same applies to the CPU PLL I did not observe that it help anything to lower it. As I said it just made the CPU slower at the same clock. Raise it back up to see if your Gflops go up.

Strange... I have never found the lower CPU PLL voltage to help anything and only deteriorated my CPU benchmarks. I am not sure about the IOH PLL though. I never played with it but 1.35V does seem low.

I am about to try some overclocking with this mobo, the Highest overclock I have ever had was 4.8ghz with a GA X58A UD7 but I had to change out my watercooling for an H70 and so far the highest I have got mine is 4.2ghz at 80-82gflps   

http://valid.canardpc.com/show_oc.php?id=1419980  
 

rafale

Strange... I have never found the lower CPU PLL voltage to help anything and only deteriorated my CPU benchmarks. I am not sure about the IOH PLL though. I never played with it but 1.35V does seem low.

 
rafale, a slightly lower CPU PLL VCore is a proper adjustment for multiple video card rigs also...Shamino and Shansmi taught me that in the year 2009, and I've always run multiple
video card No.1 rigs, thus my CPU PLL VCore is at 1.65V. If ov3rclock4 runs his down as
low at 1.5V for stability, period, and it works for him, not affecting performance per se, then
you have to accept the fact that it has some validity, somewhere, somehow. 
 
In addition personally I run very low VTT voltages (125+ to 150+ MAX) and 12GB (6 x 2GB DIMMS) overclocked G. Skill RAM (1688Mhz RAM speed in ELEET/CPUZ, 1696Mhz in BIOS), which takes extra VDIMM (again, Shamino, Shansmi data as I recall it) and/or extra VDIMM+ slightly relaxed timings. I think that the relaxed timings costs me GFlops, but I'd rather have a stable 12GB RAM situation than high GFlops. High GFlops is not a true test of performance: high Mhz and stability is though!  
 
rexrzer77 

5 or 6 Gflops is a big difference. I can test with higher IOH PLL and CPU PLL if you like and show you the difference.

Need to get my 930 @ 4.44GHz in here. Could probably do 4.5GHz but I think my 1600MHz RAM is limiting me. Probably run LinX when I go to bed when it is cooler.

We're arguing apples and oranges:
 
1) Without being a stable computer my 970-equipped Classified rig would fail its FAHome assignments, have BSOD epidodes, cause me endless grief which is unnecessary, and the combination of adjustments I arrived at for its stability is similar to my 920 experience, with one caveat--it needs more VCore than I am used to giving a CPU because of my low-voltage 920's. Apparently the 970 can need a lot, or very little VCore, it's possible to get either type and neither one is necessarily better than the other.
 
2) My RAM timing now is 9-9-9-25-96-2T; G. Skill says 8-8-8-21-59-1T is recommended. My 970 passed LinX without issues with my adjustments to RAM timing and increased voltage, from 1.67V to 1.68V, no other changes whatsoever. I was getting failures on the same LinX test (20X @ 4273Mhz) the other way, (at or about iteration 18 to 19 out of the 20 necessary) so my adjustments were a good thing for LinX stability and I've kept them for my 24/7 overclock of 4215Mhz.
 
3) I don't know if I've lost or gained RAM computing power in relation to the work I do, maybe so in the absolute sense by GFlops testing, but OTOH I am stable, and completing my FAHome WU's in record times (for the CPU in question the 970...it is literally 1-2 minutes faster per iteration with the new adjustments) these days. My PPD has increased by 20K PPD, so a good result overall for the RAM timing adjustments, VDIMM adjustment, everything else the same.
 
If you base everything you do on GFlops per CPU at a given Mhz, then my GFlops with the 970 are normal at or about 80.75GFlops average at a clock speed of 4273Mhz in LinX testing. I didn't lose GFlops that I never had, you know? If the LinX run didn't complete then any GFlops that were higher than 80.75GFlops average that I am getting now are still meaningless because the computer wasn't stable enough to complete the LinX run and earn higher GFloops, if there were higher GFlops to be earned, that is. 
 
If Joe Blow over there runs 4273Mhz with RAM timings that are higher/tighter than mine, ie 7-7-7-18-59-1T, I could care less really, if he gets higher GFlops than I got, because I am successful in the same LinX run and completing FAHome Big Adv. Work Units faster than I was with stability issues, no doubt!  
 
So you tell me who was first, the chicken or the egg? 
 
rexrzer77 

rafale

Don't you have to save power in Japan?

My prefecture is fine and we haven't had a planned blackout for awhile that I know of. Not to sure on that as I have been at Sendai Airport in Operation Tomodachi with Task Force Fuji helping them get the Airport up and running by the 11th for the 30 day anniversary. By the way things were looking when they told us we needed to leave they would have it done. Also was at a few schools cleaning the grounds of debris.
When I left though on the 26th with 7 other Marines to replace another 8 we had damn near all the lights out in every building and planned power outages. Just got back yesterday and everything is back to normal.

And what is the point of having 4.0Ghz vs 4.5Ghz when it will make difference of 120 > 130FPS? Not saying you should not overclock but for i7 1366 socket CPUs i believe 4.0Ghz is Golden spot.

and who wants a 1000 dollar chip, and run it stock?