Quote:
Originally Posted by ReelMonza
nice !!!
it needs a bit more to work now....but now you can specify VIDs for every single multi...great !
Btw. i always freeze when i have "restore CPU defaults at exit" checked...so i just leave that out all the time. Otherwise works nice :)
Is this really the case? Can you only use max 1.648 vcore when using cnq or can ANY combination normally used without cnq be used with cool and quiet and still be productive and functional?Quote:
Originally Posted by arctic-k20
Here is my settings if someone could please help me.
I am running 247 x 10 with a 180 divider on my ram cause my controller sucks. I am running 1.350v + 123% bringing up my total vcore to 1.6605v which is needed to keep my cpu stable at the full 2470mhz. My LDT is at 4x bringing total HTT to 988mhz. Now I am wanting to try cnq when I get home so am wondering how I would do this with the information I listed above.
would I set startup vid to 1.350 as that is what I need as my base vcore/vid from what I mentioned above, set my set cpu vid --> auto instead of 1.350 which is what I have it at now then would I set Set CPU VID Special Control to 123% to get me the needed 1.6605v ontop of the base vid of 1.350 in the startup vid option?
I hope I am understanding this correctly if not please correct me. It is hard to do this when im not in front of it but it is making sense to me if my idea is correct. If anything is wrong or it won't work with my needed vcore then please let me know cause unfortunately I need the full 1.350 + 123% to have my processor stable. To me it seems the only change I need to do is enable CNQ which is currently disabled as well as change SET CPU VID to auto which is normally at 1.350 and set STARTUP vid to 1.350 to replace the SET CPU VID setting which is currently at 'startup' or 1.40 being the base voltage for a winchester.
I know this is confusing but I hope someone can help. Thank you for anyone who takes the time to read this and sort me out.
ok...
CnQ provides at 5xHTT 1.0V and at max HTT 1.4V
So.. you say you need 1.66V use "special Control" = 123% which gives roughly 1.72V. (because CnQ mx =1.400V * 123% == 1.722V)
(IF my calculation is right and my assumption that max. default vcore provided by CnQ = 1.4V)
To keep it safe for a start you can also start with 112% and check voltages in windows. If not enough use 123%. I dont think there's anything between to make it exactly 1.66V
"CPU Startup" .... use 1.400 too so you have sufficient voltages while booting until CnQ is loaded.
"CPU VID" = AUTO (since CnQ will provide that later)
"CPU Spc.. Control" either 112% or 123%
you also need to download the latest "A64 processor driver" from amd websiteQuote:
Originally Posted by cantankerous
http://www.amd.com/us-en/assets/cont...ies/amdcpu.exe
and then in your power-saving-settings you set it to "minimal" which basically activates CnQ and the downthrottling of the CPU.
LMK if you have any problems.
so it appears it isn't going to work for me afterall which is what I was afraid of. I hoped you could use any combination of vid + % but it seems that is not the case. 112% only gives me 1.57v which I know isn't enough and 1.40 + 123% gives me 1.72v which is way too much and not needed to keep my cpu stable at 2.47ghz. I wish there was a way to use 1.350 x 123% to get 1.66 but if that is not the case i guess im SOL and CNQ is not an option for me. I was hoping to cut down the heat emitted into my room during the summer months which can be quite extreme as well as save on hydro while my computer is just sitting there idle. Thanks for your response though. I appreciate your help even though it didn't benefit me much.
you can always use rightmark rmclock (somone posted a link above) which has all those features !Quote:
Originally Posted by cantankerous
You can provide any wanted VID and then your "SPC Ctrl" (from bios) is added and you can try around and then find the combination which comes close to the voltage you want.
Eg. CnQ will give you 1.4 (no matter what) but in rightmark (right now) my max VID is 1.275 (+spc ctrl)
sig updated if you want to know how to set up rightmark. Its a bit of a PITA but it does work with the right settings.
thanks for all your help. I was hoping to do this whole cool n quiet thing with minimal fuss and software installation. The bios method seems easy and painless providing you can get the right voltages going but I will give an honest look at full software methods via windows and see what I come up with and decide on doing. It really isn't a big deal, just thought I would see what else my computer can do through tinkering hehehe. It is running fine just as is so really isn't going to kill me if this turns out to be more trouble than it is worth. Thought the savings of heat and energy would be nice though.
A little off topic but wondering, when LDT multi is set to AUTO what is it being set at exaclty? I have 247fsb X 4 LDT Multi bringing me 988mhz. What would AUTO set it too? Higher or lower? Just wondering if setting it statically is better than automatically. Just trying to understand the little things so I can get the best performance out of my system possible without sacrificing stability.
Thanks flexy !
It takes some playing around to get it to work but it does work, have been using it on my server for last couple of months and never had any issues. see sig
Well I thought everything was working great until I realized setting CPU VID to AUTO does default to 1.40v and there is no changing it. I installed the new driver np and installed the dashboard so I could monitor. I had Smart Guardian as well as cpu-z open so I could monitor cpu speed and temps in real time as they changed. I had windows boot up at 247 x 8 which gave me 1235mhz at 1.29v. Great I thought! Then I loaded up a few things and watched the volts and speed jump up and down as the load got higher and lower, I was really happy this was working. Problem was when I went to test with OCCT it put my processor into full load and low and behold volts were at 1.72v instead of 1.66 like I hoped it would be. It took 1.40 + 123% instead of my startup value of 1.350 + 123% like I hoped so I guess CNQ really is a no go for me. 1.72v is insane for A64 volts under air and im not going to run higher than I need to just to get this to work. Perhaps it is impossible to change the value of the CPU VID when it is set to AUTO because of the chip architecture but it would be really cool if there was a way for DFI to have the bios let you select what you want the CPU VID to default to when using AUTO for CNQ that way anyone and everyone can benefit simply by setting their CPU VID default AUTO value to what they normally use + their desired %. Probably isn't going to happen but wishful thinking right?
yeah what i said. Its what CnQ feeds the CPU with.Quote:
Originally Posted by cantankerous
They (AMD) need a more configurable CnQ so you can specify the VIDs and not take the built-in ones. That's why rmclock is far more flexible.
And thats ALSO why many overclockers who only use CnQ run into problems.
They work with their voltages or THINK they set a voltage in bios (because they rally dont understand the settings)...but CnQ puts 1.4V on the CPU PLUS the Spc. Control and they end up with 1.8V or even higher...which gives instability.
Its just not flexible enough.
Unfortunately true. At least I now understand what is going on and did have it setup correctly however just not to a desireable effect due to it not being flexible enough. If rmclock can do it more flexibly than it can't be a cpu architecture thing but a setting thing. Even if DFI never do it is it at least possible for the bios to allow a different value for CPU VID other than 1.4 for auto? The only other solution is to include a 118 or 119% option between the 113 and 123% option as that is a pretty big gap.
Either way CNQ is hard for users of high voltage ram as with CPU VID set to auto more than 3.2v is impossible due to the auto setting. I have the 3.3v rail mod on my psu to feed my BH6 and more than 3.2v is just not selectable unless you have an actual value in replacement to the auto setting.
Flexy thanks a bunch for your help matey, much appreciated. I learned a lot tonight going through this.
EDIT: wondering how much resources and overhead rmclock and other programs have while doing what they are doing to monitor and lower clocks and volts. The main reason I don't want to install all this software is because of the fact that it constantly runs in the background with high overhead and resources lost that could be towards other things. I understand I am being picky but with an only average overclock of 2.47ghz and ram running at low fsb as well I need all I can. Perhaps I should get a SD to get a hopefully higher overclock plus at lower volts. Maybe with another cpu lower volts will be fine and I can no longer worry about the 1.40v issue.
hmm.. wow.. ive never had a single issues with having cnq enabled. never noticed a clock drop either. never cause instability for me. even at ~2.4ghz when this chip tops out at 2500, not stable at all, it will do with anywhere from 1.45vcore to 1.7.
maybe I just have a really freaky CBBID. o_O
Before I explain what happens after I tried the above, first I cover what has previously been working without issue...Quote:
Originally Posted by flexy
Bios settings:
CPU VID Control : Auto
CPU VID Special Control : Auto
This allows the RM app to work properly, with the multiplier dropping as well as the vcore value - as reported by CPU-z, Central Brain Identifier and MBM5 (with the appropriate user tweaked DFI NF4 .ini file that everyone else is using). I will reiterate, this works all day long with the multi and vcore scaling up and down as required, it's rock solid.
Now, when I try your suggestion...
Bios settings:
CPU VID Control : (insert any value other than Auto here)
CPU VID Special Control : (insert any value other than Auto here)
The multiplier still drops resulting in a lower core speed but the vcore doesn't.
What I'm saying here is that in order for the RM app to be able to alter the vcore dynamically, the two BIOS settings above MUST be on Auto... at least that's how it is on my rig and as such it seems that I cannot use more than 1.45v in RM :(
hey timko,
i never said that "CPU VID Ctrl" has *NOT* to be on AUTO, i swear :)
Maybe i just used confusning names :)
about "special control": i always have that on like 123%...if i leave that on Auto i would not even KNOW what voltage i will get later. (Thus risking destroying my CPU)....IS THERE an auto setting anyway ?
Oops, my mistake :DQuote:
Originally Posted by flexy
I've just gone back into the BIOS and set CPU Vid Control back to Auto and changed the Special to the appropriate percentage to get my needed max Vid when RM is used. As such, I can now get my 2.7Ghz at 100% load as required. I was only short by 2Mhz on the HTTpreviously but it wasn't stable at 1.45v and now it is with the extra fraction of vcore I can now pump into it.
Thanks flexy :)
How can i set an voltage under 1,100V with a Venice? :stick:
Hi @ All ! Flexy nice thread and good idea using CnQ with overclocking!
So that is exactly the problem I have, I couldn't get a lower voltage,Quote:
Originally Posted by Atlantis
neither from BIOS -although I can select such a voltage- nor from RM Clock!
What might be the reason? Can anyone help the situation?! :confused:
want to add that *I* have probs running rmclock clocking down my CPU to 4xHTT (260). Its not stable. Solution w/ rmclock: Minimum downclock to 5x (instead of 4x)...voltage (for me, 5x260 == 1.00V + 23%) totals to 1.18 as MBM tells me.
Your mileage may vary, maybe the instability at 4x depends on the HTT...its just ODD...but 5x works nicely.
You can test this by setting to "minimal" performance profile in rmclock (so it doesnt upclock w/ CPU load) and run some OCCT stability tests at minimum HTT/voltage.