High vtt needed is weak board or cpu?

[xpower]

hi everyone,
actually my sig rig required more vtt for 4000 uncore. thinking is cpu IMC is weak. and not the board. what you guys think? any other way to determine the weaker one

[acebmxer]

I have used my cpu i7-920 in 3 different boards, EX58-UD5, X58A-UD5, X58A-OC. Qpi/vtt is close enough on each board to say its the imc in the cpu. There was little variation but that could be how the board handles qpi/vtt as the X58a-oc took less qpi/vtt then the other boards but then again it over volts qpi/vtt by up to 0.05v.

I have a 930 also but I have not used it all the boards but it requires more qpi/vtt then my 920 does at any given speed.

Thanks from my experience.

[FedericoUY]

Agree, every board has not only Vcore drop, but any other voltage may have a increase or decrease in it's original / setted value, but in a perfect controlled voltage board, the chip is the one that needs more or less vtt.

[xpower]

just testing 4Ghz uncore, 920, on this board but it seems like its need 1.6v to handle that uncore. qpi board that supply the volts to the cpu IMC. meaning the cpu is the culprit. that cant
handle?

[acebmxer]

Heres my 920 at 220 blck and 3960 uncore with memory at 1760 @ 1.395 qpi/vtt set in bios. Actual voltage more like 1.45 there about.

I think some cpus cant do 4ghz uncore but could be wrong. yes its the cpu.

[xpower]

thanks aceb
testing 200x15 with uncore 4000 and ram at 1000 with 2:10 ratio vcore at 1.0v. linx gives BSOD with 124 code. vtt is at 1.65v.

[acebmxer]

You sure your cpu can do 3.0ghz on 1.0v vcore? I have seen if a voltage is so far out that 101 and 124/050 be reversed.

[xpower]

You sure your cpu can do 3.0ghz on 1.0v vcore? I have seen if a voltage is so far out that 101 and 124/050 be reversed.

yes it can do 3.0ghz on 1.0v. getting 124 is because of less vcore?

[Leeghoofd]

124 is usually VTT/QPI voltage...

Lower uncore to eg 3600 and keep VTT at sub 1.4ish level... To me 1.65VTT is for bragging rights...and will kill something soon... Btw not many CPU's do 4000 uncore stable on positive temps... No idea what you want to achieve at 3000mhz CPU clocks and 4000uncore... seems weird... I would do the inverse

[xpower]

124 is usually VTT/QPI voltage...

Lower uncore to eg 3600 and keep VTT at sub 1.4ish level... To me 1.65VTT is for bragging rights...and will kill something soon... Btw not many CPU's do 4000 uncore stable on positive temps... No idea what you want to achieve at 3000mhz CPU clocks and 4000uncore... seems weird... I would do the inverse

nothing so special to achive . just curious to test 4ghz uncore and qpi 4ghz speed but seems like it is taught for 920.

[xpower]

problem is when i set the ram to 2000, uncore goes to 4000.
just wana run ram to its rated speed.

[Leeghoofd]

Why not try 1866Mhz C8-9-8-24 with 3600Mhz Uncore... have you tested the performance difference between 1800C8 and 2000C9 ?

Your 920 doesn't seem to have a stellar IMC, so you have to look for some extra ooomph elsewhere... running 1.65 QPI is not the way to go ...

[xpower]

vtt 1.7
for uncore 4000Mhz
on air 24/7