• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

CrystalCPUid - Voltage / Multi control - Anybody using it ?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

iamjcl

Member
Joined
Aug 4, 2003
In theory, speedstep seems like a great Idea: Lower the multi AND the Vcore when processor useage is below X %. Wall the multi (and the Vcore, but not in that order) when the task calls for it.

Would make worry about Vdroop etc... a thing of the past.

I know there is a "lag" when it kicks in / out etc..., but I don't really mind that, because I can't detect it.

Anyway, on this MB, if you enable speedstep, you have no Vcore control. But using CrystalCPUid, it works pretty well, but on my system it can only basically kick the Vcore to "low" or "high" (what you set in bios). "Low" seems to be about 1.1 V according to speedfan.

So it works great so far, but in Crystal you have all these discrete voltage choices - anybody get exact voltages to work, or is your system only doing the "hi / lo" thing ? While this works pretty good, I just wondered if anyone has gotten it to work like it looks like it should work.
 
Last edited:
CrystalCPUID worked fine for me with a number of C2D motherboards which all used different chipsets. It allowed me to change the VCore in the range 1.1x to 1.325V. The newer Intel chipsets gave the largest range whereas the older ones and the VIA chipset boards limited me to ~1.3V. This was with the same E6400.

I find RMClock gives more control than CrystalCPUID when using Speedstep, although it didn’t offer a larger VCore range on those boards I tested it on.

To sum up; you can’t over-volt with these tools and if you over-volt in the BIOS, you typically lose the ability to change the VCore when using them.
 
smilingcrow said:
you can’t over-volt with these tools and if you over-volt in the BIOS, you typically lose the ability to change the VCore when using them.

Are you saying that we can still set a high vcore in BIOS (say 1.45v), drop the multi from say 9 to 8 and then overclock the snot out of the fsb? ('cause that would be great!)
 
fritzman said:
Are you saying that we can still set a high vcore in BIOS (say 1.45v), drop the multi from say 9 to 8 and then overclock the snot out of the fsb? ('cause that would be great!)
You will need to be more precise on how you intend to change the multi and FSB before I can begin to answer this.
RMClock will allow you to choose a multi between 6 and default regardless of the motherboard used, if that’s what you are inquiring about.
So if you then want to increase the FSB using Clockgen, this shouldn’t be a problem, provided that Clockgen works with your board.

RMClock also allows you to turn off C1E so that you run with a fixed multi. Or you could for instance with an E6600 run it with a multi of 7 and 8 using Speedstep and ignore multis of 6 and 9. It’s very flexible.
 
I think you've caught what I was thinking of... I would want to set the fsb in BIOS I imagine, so it's the same every boot (hopefully something in the 425 - 450 arena), but lock the multi down to probably 8 and vcore to around 1.45.
 
fritzman said:
I think you've caught what I was thinking of... I would want to set the fsb in BIOS I imagine, so it's the same every boot (hopefully something in the 425 - 450 arena), but lock the multi down to probably 8 and vcore to around 1.45.
The problem is that you aren’t likely to be able to boot into Windows at 9 * 425+ long enough for RMClock to be able to lower the multi to 8. It’s worth a try though. The less you have loading at start-up the better, as it means RMClock can do its thing quicker.
 
here are some tests i did
http://img.photobucket.com/albums/v733/Evilsizer/ocforums/superpi 1m runs/16536.jpg
http://img.photobucket.com/albums/v733/Evilsizer/ocforums/superpi 1m runs/16670.jpg
http://img.photobucket.com/albums/v733/Evilsizer/ocforums/superpi 1m runs/14536.jpg
http://img.photobucket.com/albums/v733/Evilsizer/ocforums/superpi 1m runs/14670.jpg

BTW despite have that wide range of 1.008v up to what ever in crystal cpuid. the voltage would only jump from 1.15v to 1.325v no voltage in between could be used. This may be the lack of voltage options in the ab9's bios as the lowest is 1.325 in the bios.
 
Evilsizer said:
BTW despite have that wide range of 1.008v up to what ever in crystal cpuid. the voltage would only jump from 1.15v to 1.325v no voltage in between could be used. This may be the lack of voltage options in the ab9's bios as the lowest is 1.325 in the bios.
Both RMClock and CrystalCPUID show a range of VCore values that aren’t exactly the ones that are accessible for every CPU.
I’ve had this working on 5 different Core 2 Duo motherboards and 3 different Core Duo motherboards, so it’s probably a BIOS setting or something else.
Speedstep needs to be enabled in the BIOS as does C1E possibly and the VCore needs to be on auto.
With CrystalCPUID you need to enable the Voltage control which I’m sure you’ve done and also multiplier management. Use version 4.8.2.310 or later, as I think earlier versions weren’t very happy with Speedstep.
Otherwise, try RMClock; it seems better for working with Speedstep anyway.
 
this sounds like a great idea, so can i set the high voltage to 1.4 and low to 1.1 ? and get the mult to go from 9 to 6 for example ?

which tool of the two seems to work the best for C2D Extreme, also what settings do i need in the bios at the mo speed step is off, c1e is on and cool/quiet is set to off and voltage set to auto. I want to increase vcore but c1e has to be disabled to get and control of the vcore.
 
thedoc said:
this sounds like a great idea, so can i set the high voltage to 1.4 and low to 1.1 ? and get the mult to go from 9 to 6 for example ?

which tool of the two seems to work the best for C2D Extreme, also what settings do i need in the bios at the mo speed step is off, c1e is on and cool/quiet is set to off and voltage set to auto. I want to increase vcore but c1e has to be disabled to get and control of the vcore.
Unfortunately, with the P5W DH and every other board that I tested, once you set VCore manually in the BIOS you can no longer change the VCore using Windows utilities. So you will typically be limited to a range of 1.15 – 1.325V.

It’s possible that other motherboards and/or possibly other chipsets will not have this limitation; I’d like to find one that doesn’t. I wrote about my experience of this with a P5W DH HERE.
 
thedoc said:
this sounds like a great idea, so can i set the high voltage to 1.4 and low to 1.1 ? and get the mult to go from 9 to 6 for example ?


Thats exactly what you can do, and exactly what I'm doing. I'm not sure exactly what smilingcrow is alluding to, but with my P5W-DH I can set Vcore to whatever it needs to be in BIOS (say, 1.5v), and speedstep etc.. is all DISABLED in bios.

Then, when XP boots, it loads Crystal, which re-enables speedstep, and gives you some control over how the CPU switches from "low" to "hi".

In "high", whatever Vcore you picked in BIOS is what is used, along with the 9X (in my case - 6600) multi.

In "lo", Vcore drops to 1.05v, and the multi drops to 6X.

Which multi's are used and the load required to switch are all adjustable in Crystal.

What I found not to work properly were the discrete voltage choices in Crystal - but I found basically if you tell it in "lo" to go to 1.xx Volts, that will end up putting 1.05 Actual into the CPU.

At the "high" / 9X multi setting, just pick (in my setup, anyway) a voltage over 1.3. All the diff. values I tried produced the exact same result: Exactly what you set in BIOS.

So you have only 2 voltages: What you set in bios, and 1.05V. Crystal switches between these 2 voltages, and 9X (or your max. multi) goes with the BIOS-set voltage, and 6X (or whichever lower one you choose) goes with 1.05v actual.

Works GREAT.

Idles at 6X / 1.05v / low temps
Load it, and it jumps to 9X / BIOS Vcore / "normal" temps.

See THIS thread.
 
Last edited:
iamjcl said:
Thats exactly what you can do, and exactly what I'm doing. I'm not sure exactly what smilingcrow is alluding to, but with my P5W-DH I can set Vcore to whatever it needs to be in BIOS (say, 1.5v), and speedstep etc.. is all DISABLED in bios.
Well that’s great news, I couldn’t get it to work for me with BIOS 0801 or 0901 Beta. I see that 1305 is the current BIOS, which version are you using?

Also, which version of CrystalCPUID are you using? Edit. I checked the link that you gave which answered the second question.
 
Last edited:
I'm currently using 1305, but I know it was OK with 1101, and almost positive also OK with 801 (which I flashed back to day or 2 ago to check something - crystal running all along) didn't have problems, but cant be 100% on 801.
 
iamjcl said:
I'm currently using 1305, but I know it was OK with 1101, and almost positive also OK with 801 (which I flashed back to day or 2 ago to check something - crystal running all along) didn't have problems, but cant be 100% on 801.
It can be easy to miss if you aren’t looking for it, especially if you only had 801 loaded for a short time.

I noticed a big change with the Gigabyte DS3 between BIOS F3 and F4. With F4, if you left VCore on auto and increased the FSB, the VCore jumped to a fixed setting of ~1.4V regardless of whether EIST was enabled. A weird one!
These are early days for tuning a BIOS for C2D, so we should expect oddities stuff like this.
 
Stilletto said:
So is this only efective for the C2D?
no, just cpus that use EIST or speedstep. as its a power saving feature to slow the cpu down when not in heavy use.
 
smilingcrow said:
I tried RMClock with a P5B Deluxe using the latest BIOS and it works exactly the same way, with a fixed lower VCore of 1.05V.
I’ve tested this more thoroughly since and it seemed to be more flexible than I first thought; I could set any VCore value using RMClock and CPU-Z (1.36) registered the change of voltage.
But when I tested the power consumption with both cores fully loaded with Prime95, there was very little difference in the power consumption figures for the system when I changed the VCore setting with RMClock. Ironically, with the VCore at 1.05V the power consumption was consistently 1 or 2W higher than when it was at 1.1376V.
So CPU-Z (1.36) is inaccurate with this motherboard with BIOS 0507.
It doesn’t mean that the same holds true for the P5W DH, but it’s worth investigating.

I haven’t tested with the newly released BIOS 0614 yet; hopefully this rectifies this anomaly.
 
Trying to work out how to get Crystal to kick in during bootup, and just don't get it.

I read the piece about putting this "
/CQ = Enable Multiplier Management (Code Name:Crystal'n'Quiet)
/HIDE = Hide Main Dialog" in the command line, but in what file?

Once that is sorted, then I presume I simply paste a shortcut to the .exe into the normal startup section under all progs.

Thanks (in advance)
 
Just place a shortcut in your "startup" folder to the .EXE file for Crystal.

Here is mine as an example:

"D:\Program Files\Overclocking\Crystal cpuid\CrystalCPUID.exe" /CQ /HIDE

That is the "target" in my shortcut.
 
Back