I'm having... hmm... experience with this board...

 

It is very good so far, seem like CPU's need less VTT and vDIMM to run Elpida Hypers at 4000/2000/7-7-7-20-1T.

 

But, it also killed my favorite CPU OK, it might have died anyway, but it's like borrowing a car to a friend and it breaks while he was driving. You will not hold him responsible, but it just leaves a bad taste in your mouth.

 

Booted at 5GHz at 1.55V. Set vCore at 1.6V, pressed Multi+ once and made CPU-Z s/s at 5.2GHz. Restarted.

Booted at 5GHz. Then set vCore at 1.645V (my 990X ran benchmarks many times at under 1.65V), then I pressed Multi+ button twice. Took CPU-Z s/s at 5.4GHz.

Rebooted. Noticed that vCore is 1.7V. Set vCore again at 1.645 and noticed that it does jump at almost 1.7V (checked with DMM). Rebooted.

 

CPU worked next few hours, while I was trying to make it run at 4.2/4.0/2.0@7-7-7-20-1T unsuccessfully, then it just stopped at C1. Dead.

 

 

llc1? or no loadline calibration used?

llc2 or whatever, higher one. But there was no load and it usually stayed as set in BIOS when idling, with some small increase in voltage under load, like 0.01, but no 0.05V.

did you play with the vrm frequency switches?

Of course, but not at that moment. VRM freq was at 800kHz IIRC.

Usually didn't power off while changing freq, but I never changed it in Win, I did it while idling in BIOS.

use stock frequency...its only needed when you are at the max load say 6ghz vantage.

Gigabyte boards do have a bad reputation of killing CPUs.

But you were asking for trouble yourself - running 1.65 vcore with ambient cooling

use stock frequency...its only needed when you are at the max load say 6ghz vantage.

 

I'll keep that in mind. Thanks.

 

Gigabyte boards do have a bad reputation of killing CPUs.

But you were asking for trouble yourself - running 1.65 vcore with ambient cooling

 

Yeah, I know, but never expected trouble around that corner.

I was very cautious about VTT, as it is known to be higher then expected on this board. But it is always higher.

vCore was spot on/slightly higher, and than jumped much higher.

I know and accept the risk of overclocking, but this IS a problem with board.

1.65v is too much for gt, 1.65v for 45nm bloomfields were slow killers. 1.65v is just too much imo.

 

as for the board, it works ok for me. it does overvolt-gigabutt trick to make things work fo sure...

Oh man sorry to hear that, bad luck there . It really sucks losing a good chip I know how you feel!

 

You were playing with fire at those volts and cooling. Gigabyte boards set vcore a bit higher once you start going past 1.65, if you had a play with other boards in the past this jump would be familiar. It only happens around 1.65 in bios. This is normal. You really should have had subzero cooling for anything past 1.6 mate, I wouldnt stuff around next time no matter what board you use.

 

use stock VRM unless you specifically see an improvement. I very much doubt you will see anything change on standard cooling. It might make a difference with some CPUs subzero however. Always test extensively with stock on Gigabyte first though and then gradually work your way up that's my advice.

Edited August 31, 201114 yr by dinos22

1.65V, ht enabled and benching. That's a dangerous combination on air cooling.

 

You got balls, sir.

I killed a good 980x once because vcore was still on 1.8v after benching with LN².

Forgot to clear the bios and booted it up on air.

Cpu was still alive but all CPU intensive tasks did not work.

Could not even get Wprime to work at stock.

Oh man sorry to hear that, bad luck there . It really sucks losing a good chip I know how you feel!

 

You were playing with fire at those volts and cooling. Gigabyte boards set vcore a bit higher once you start going past 1.65, if you had a play with other boards in the past this jump would be familiar. It only happens around 1.65 in bios. This is normal. You really should have had subzero cooling for anything past 1.6 mate, I wouldnt stuff around next time no matter what board you use.

 

use stock VRM unless you specifically see an improvement. I very much doubt you will see anything change on standard cooling. It might make a difference with some CPUs subzero however. Always test extensively with stock on Gigabyte first though and then gradually work your way up that's my advice.

 

Thanks, and thanks for the heads up, I'll remember it good, well it will be carved in my memory for sure.

I'm just sorry it never made the LN2 run, as I was planing one soon. This board just asks for it.

 

I was just playing with VRM freq, it's low again.

 

1.65V, ht enabled and benching. That's a dangerous combination on air cooling.

 

You got balls, sir.

 

Tell it to the girls, as I almost feel like loosing one.

 

I killed a good 980x once because vcore was still on 1.8v after benching with LN².

Forgot to clear the bios and booted it up on air.

Cpu was still alive but all CPU intensive tasks did not work.

Could not even get Wprime to work at stock.

 

Sorry to hear that. I hope mine would boot, so I can tell it's not dead completely, just crippled, veteran...

Sorry for the loss, but it just common sense when up 1.65+Vcore HT on to use subzero otherwise its harakiri

Edited August 31, 201114 yr by dumo

you never know dude the new chip could be even better. a mate of mine got a new chip and it doesnt 6.4GHz 3D eeek

Common sense is not my strongest side although I wasn't pushing it intentionally so much over 1.65V and it was just for CPU-Z s/s.

 

There will be new chip only if I buy one, cause this one is ES, no RMA.

Here it is... 990X... Preparing for the last ride...

 

 

 

Well, it's not all black. As I said, this is very good board...

 

Memory timings as on s/s (RTL +2 +2) VTT=1.355V / 1.380V actual, vDIMM=1.72 / 1.715 actual.

It is 0.04V less VTT and 0.06V less vDIMM then on R3E. Compare link here

vCore=1.4375 / 1.4380V actual, 0.045 less then on R3E. This 980X is not bad at all.

 

nice rams

you never know dude the new chip could be even better. a mate of mine got a new chip and it doesnt 6.4GHz 3D eeek

 

Hehe, Intel send me a better one, that's true

Thanks honda, it does have some kind of stability too...