A .08v drop under load?

[Methodx]

Hey all, I just got my new Fortron 530 w psu in, which a lot of overclockers say is awesome.

I'm loving it so far, but when I set my vcore to 1.7v in the bios and windows boots up, both the hardware monitoring software that came with my mobo and cpuid-z say I'm only getting 1.66v.

Should I consider my cpu to be getting 1.7v or 1.66?

Also, under extreme load (2 instances of prime95 running the worst torture test simultaneously) my vcore will drop down to 1.62 v. I know that one can expect the vcore to drop a little under load, but this is now a .08v drop under what I set the vcore to. Is this something to be concerned about?

[larrymoencurly]

Unless the PSU puts out a +5V or +12V way, way out of whack, the Vcore is affected virtually only by the voltage regulator built into the mobo.

[stan03]

could you post your other rails also?

[Methodx]

Right now i have my vcore set at 1.65v in bios.

My software reports:

VCore - 1.60v
+3.3v - 3.20 v
+5v - 4.95v
+12v - 11.52v
-12v - -12.11v
5VSB - 4.87v
VBAT - 3.20v

My source is Windbond Hardware Doctor Version 3.4.1.1 which came on CD with my mobo.

I am idle right now, all of those numbers go down even more under heavy load. With 2 instances of prime95 running vcore will go down to 1.58v and occasionally hit 1.57v for a brief second.

Should I consider my vcore to be 1.60v even though in bios it is 1.65v? Would it be safe then to raise it another .05v to get an "actual" 1.65v?

[larrymoencurly]

I don't know if a 1.9% change matters, but if the BIOS is as accurate for this voltage as it often is for the PSU voltages, you may want to take measurements with a meter. The CPU core voltage or something very close to it is usually present on one of the MOSFETs, either at the metal tab one one end or a pin on the other end. You don't want your meter probe to slip or you could ruin the mobo and CPU and maybe even weld the probe in place. Don't measure at the MOSFET pins without first putting pieces of cardboard between them to prevent slippage. The measured voltage may be slightly higher than the actual CPU core because there may be a resistor inline with the output to sense current and shut down the regulator if the amps go too high.