Running your CPU on auto or fixed VID

[mikeg]

Ok my setup to start with is an Asus P5QL-E with a Q9400. I am only overclocking the CPU to 3.2 gigs and have set the CPU Voltage to auto in the bios. I have noticed that with the voltage set to auto it scales it up as the I have enabled speedstep. It lovers the voltage to CPU when it is stepped down to 2.4G running 6x400 and then when I do encoding or run prime it goes up to 1.3v on the cpu. Now if I run it with a fixed voltage will it drop the voltage when speedstep lowers the CPU speed? Just wonder what others have been doing.
Mike

 

[Idontcare]

First just some terminology so we are speaking about the same thing - you don't set VID (Intel does that), you set Vcc in the BIOS, if on auto then BIOS sets Vcc to the CPU's internally listed VID.

Folks around here refer to Vcc, Vcore, and CPU voltage synonymously.

Typically when you manually set Vcore to something other than auto in the BIOS the BIOS then defeats the voltage reduction power-savings portion of C1E. The same is true if you change the multiplier from auto to a manual value, the BIOS will then defeat the multiplier change requests given by C1E on the CPU and fail to reduce the multiplier.

I am not aware of any motherboards that do not do this. If you tell the BIOS to operate your CPU at 1.3V then it is not going to reduce that voltage to 1.1V just because C1E on the CPU says to do it.

Likewise if your stock multi is 10x and you set it 9x in the BIOS then the BIOS is going to run it at 9x no matter what, and when C1E requests the BIOS reduce the multi to 6x for power savings during idle the BIOS will ignore the C1E request.

 

[mikeg]

an example is when the system is in idle mode the CPU is running at 6x400 or at 2.4gigs and the CPU voltage using the monitoring software or CPUZ shows that the voltage is at about 1.015v but if I run prime of do video encoding the CPU voltage goes up all the way to 1.305v. Now if I set in the bios to manual it always stays at the voltage I set it to. I do use the system for a lot of background tasks that don't need a lot of CPU power like downloading torrents or playing mp3s so I would like to let the CPU use speedstep to lower the power needed to run it
Mike

 

[Idontcare]

Is there a question somewhere here? I'm not following...

 

[JaBro999]

Idontcare,

Thank you for the VID/Vcc explanation.

I'm curious, if you don't change the Vcore or the multiplier and only overclock by increasing the FSB (say 333 to 400 - assuming CPU cooperates w/o voltage bump), would the CPU's C1E voltage reduction request take effect when the system is idle?

I apologize if this is getting OT.

-jb

 

[BTRY B 529th FA BN]

IDC for president :beer: ... well, moderator atleast?

 

[mikeg]

That is what I am trying to get too? I know that if I set the voltage manually it does not allow the voltage to drop when it is in idle. I don't know if it only works with this mobo also. Has anyone else done any tests to see if your systems drop the voltage when it goes to idle mode? The voltage drop does really help on the temps too. When idle system run at about 36c
Mike

 

[Idontcare]

Originally posted by: JaBro999
Idontcare,

Thank you for the VID/Vcc explanation.

I'm curious, if you don't change the Vcore or the multiplier and only overclock by increasing the FSB (say 333 to 400 - assuming CPU cooperates w/o voltage bump), would the CPU's C1E voltage reduction request take effect when the system is idle?

I apologize if this is getting OT.

-jb

Well I definitely hope its OK to ask these kinds of question mid thread, I do it all the time too

Yes, C1E will still be active and able to reduce to voltage in the scenario you propose.

The best OC you can do (from a power savings viewpoint) is one in which you only increase the FSB and you leave Vcore and multiplier alone. This keeps all the power savings features intact and functional.

Originally posted by: BTRY B 529th FA BN
IDC for president :beer: ... well, moderator atleast?

Ha ha, I don't know about president but I'll take that :beer: and say thanks!

We don't need me as a mod, mark does an awesome job already and besides we are a pretty tame bunch here in the CPU forum from what I can tell. I definitely believe we treat each other with more respect than the typical forum on the web.

I wouldn't come here if I didn't feel that way, and the day we become like that is probably the day I'll stop visiting this forum. But I don't see that happening.

Originally posted by: mikeg
That is what I am trying to get too? I know that if I set the voltage manually it does not allow the voltage to drop when it is in idle. I don't know if it only works with this mobo also. Has anyone else done any tests to see if your systems drop the voltage when it goes to idle mode? The voltage drop does really help on the temps too. When idle system run at about 36c
Mike

I know of no motherboard/bios that will decrease the voltage when the chip goes to idle if Vcore is not on auto/default in the bios.

How critical is this to you? Is it a "need to have" or a "would be nice to have" situation?

 

[nerp]

If I manually set my voltage in BIOS on my IP35-E, it still lowers voltage and raises it under load. The figures are higher on each end but C1E seems to continue working. . .

 

[Diogenes2]

Originally posted by: nerp
If I manually set my voltage in BIOS on my IP35-E, it still lowers voltage and raises it under load. The figures are higher on each end but C1E seems to continue working. . .

Same here with my DFI LP DK P35-T2RS .. I have my volts set to +.1 - with C1E enabled -
Voltage is lower when idling ..

 

[Idontcare]

That is good to know, and helpful. Wish I had known that a year ago.

 

[imported_Scoop]

Ok, maybe I should read other posts before posting next time so I won't repeat what has been said already. But I will give another vote for Abit IP35-E, as it does lower the voltage when idle like has been said.

 

[mikeg]

Thanks a lot people you guys have answered my questions.
Mike

 

[OverVolt]

So no wonder it stopped bumping my voltage down when I set a manual Vcore.

Kinda lame. I get errors in P95 with 333x9.5 on my E7200 @ 1.18vid 1.12v idle. 2hours in, the same core fails, like clock work.

I've been running with it that way for awhile, and decided to play with the voltages a bit to get them stable, and all of a sudden it stopped knocking the voltage down while idle

So I have no errors so far @ 1.45v









Ohh just kidding it's at 1.2vdroop 1.22vcc via CPUZ When I set the Vcc to Auto while overclocked, the board grossly overcompensated and set my Vcore to 1.26-1.28, which is past my maximum voltage I'll allow of 1.25Vcc.

Fascinating I know. It was stable playing games even though P95 failed it like clockwork on the same core after 2hours.