I've got my new 975BE stable at 4200mhz but the vcore is at 1.55v. If i understand what I've read here 1.55v is way high especially everyday use. I'm having trouble with this whole thing. First I thought my 940BE OC'd to 3700 no problem for a gain of 700mhz. I thought that since the 975 started at 3600 then 4300 should be doable. Nope. I guess there is a issue with going past 4000 no matter where you start. I've backed off the vcore 1.47v and 4100mhz solid. I've read though that even 1.47 is high. So I guess my question is how far can vcore go a)all out max and B) for every day use? I still have plenty to play with settings wise so I might get it stable at lower voltage by changing something else like the cpuNB multiplier and ht frequency to match it.
-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
OC 975BE not what i expected
- Thread starter Rigit
- Start date
- Joined
- Dec 27, 2008
1.55 is about the limit of what a lot of us would feel comfortable with for 24/7 use. though you will find those who advocate a lower and a higher max 24/7 voltage as being the limit. My questions for you are, is the 1.55 what the vcore is set to in bios or is that what monitoring software shows at idle in Windows or under load in Windows and 2. Do you have CPU LLC in bios and 3. If the answer is "yes" to #2, is it enabled or disabled?
Personally, I try not to exceed 1.5 in bios and 1.55 in Windows reporting software under load.
Personally, I try not to exceed 1.5 in bios and 1.55 in Windows reporting software under load.
- Joined
- Jan 10, 2012
that vcore looks much to high for 4200, 1.55v?
my 955 is locked at 1.42 vcore@4100
I just plugged in the 4200 profile and it's 1.44 but it produces 57c core temps.
my 955 is locked at 1.42 vcore@4100
I just plugged in the 4200 profile and it's 1.44 but it produces 57c core temps.
- Joined
- Apr 20, 2004
- Location
- JAX, Mississauna
This and That...
You write >> 975BE stable at 4200mhz but the vcore is at 1.55v. = If you next statement is true with at least 2 hours of Prime 95 in blend mode passing without any failure.
You write >> I've backed off the vcore 1.47v and 4100mhz solid. = Then I certainly would feel in my bones that having to run 1.55Vcore for only 100Mhz is a waste. Waste of power and likely more heat than is necessary overall for only 100Mhz.
You write >> I thought that since the 975 started at 3600 then 4300 should be doable. = When a cpu is designed they are all expected to run at a certain speed. THE design is for a certain speed, in fact will be expected to run at only a max speed.
As the fab process matures they are able to ship processors rated for a faster speed within the original design parameters. So they ship processors that do run at a faster rated stock speed but built on a design that was always made to go to a max speed. Since that is about how it works in the long-run; there is not a real reason to expect that a cpu with a faster released speed is going to overclock by an additional margin that a slower cpu did.
You write >> I still have plenty to play with settings wise so I might get it stable at lower voltage by changing something else like the cpuNB multiplier and ht frequency to match it. = Some have found that 'dropping' the multiplier by 1.0 or 0.5 and upping the FSB/CPU Frequency, has allowed them to reach their original multiplier only, Max clock using a little less Vcore. THIS certainly varies from cpu to cpu and system to system and whether or not doing so will result in using less Vcore can only be determined by the individual user testing his or her own system.
You write >> 975BE stable at 4200mhz but the vcore is at 1.55v. = If you next statement is true with at least 2 hours of Prime 95 in blend mode passing without any failure.
You write >> I've backed off the vcore 1.47v and 4100mhz solid. = Then I certainly would feel in my bones that having to run 1.55Vcore for only 100Mhz is a waste. Waste of power and likely more heat than is necessary overall for only 100Mhz.
You write >> I thought that since the 975 started at 3600 then 4300 should be doable. = When a cpu is designed they are all expected to run at a certain speed. THE design is for a certain speed, in fact will be expected to run at only a max speed.
As the fab process matures they are able to ship processors rated for a faster speed within the original design parameters. So they ship processors that do run at a faster rated stock speed but built on a design that was always made to go to a max speed. Since that is about how it works in the long-run; there is not a real reason to expect that a cpu with a faster released speed is going to overclock by an additional margin that a slower cpu did.
You write >> I still have plenty to play with settings wise so I might get it stable at lower voltage by changing something else like the cpuNB multiplier and ht frequency to match it. = Some have found that 'dropping' the multiplier by 1.0 or 0.5 and upping the FSB/CPU Frequency, has allowed them to reach their original multiplier only, Max clock using a little less Vcore. THIS certainly varies from cpu to cpu and system to system and whether or not doing so will result in using less Vcore can only be determined by the individual user testing his or her own system.
- Joined
- Jan 10, 2012
+1 Rgone.
- Joined
- Nov 12, 2002
- Location
- Rootstown, OH
I've ran about 1.65V on PHII for specific benchmarks only, with a Thermalright Silver Arrow and a 300CFM Airking Blower... But I did that knowing there was a good chance it could cause damage, and I only ran it that way for a very short period. I'd stop on everyday volts at where you are currently probably (1.55V feels fine for me with a good cooler and temps).
I'd also consider any advice these guys give above, as they may help you find something that gets you more speed/stability with less volts - running cooler is always good.
I'd also consider any advice these guys give above, as they may help you find something that gets you more speed/stability with less volts - running cooler is always good.
- Thread Starter
- #8
It is was set 1.55 in bios. I've since cut it back to 1.47 and reduced mult to 4100mhz. I'm going to do more tinkering to see if i can keep it stable at a lower vcore. Not sure about CPU PLL but I don't recall seeing it. I almost always do things through bios. I've never figured out AMD Overdrive and Easy Tuner 6 works OK but a setting there that crashes won't necessarily crash if set in bios.
- Thread Starter
- #9
I will get exact figures but cpunb is conservitive volt setting. I bumped ram 1 notch but I really don't think i need it since I'm running the ram at stock speed. I'll toy with less volts higher FSB but every time I've messed with FSB it crashed but that was done in windows not bios. FSB increase would necessitate higher memory volt since it would force memory to higher clock speed. Would it be beneficial to reduce memory multiplier and run 800mhz and push the timing? I just worry about all of the BSOD I get with messing around. I don't want to fry windows 7 to the point of reformat and start over. Also I don't use Prime 95 or any other similar utility. I find it ridiculous. In the real world it is extremely unlikely that I'll push all 4 cores to the max on multiple threads. I do do a lot of video editing and I'll run multiple instances of my conversion software. This pushes all 4 cores to 95-100% with multiple threads but not likely to be as rough as Prime 95. I ran it at 98-100% yesterday pretty much all day with no crashes.
- Joined
- Dec 27, 2008
Your instability when increasing the fsb could be caused by he ram frequency getting too high or by the HT Link frequency getting too high. In both cases it us usually necessary to reduce the starting frequency of the component to create some headroom if you are making heavy use of the FSB to overclock. Increasing the ram voltage a tad can have a beneficial effect on stability, even when the ram itself is not being overclocked and bumping up the CPUNB voltage can help too, especially if you are overclocking it.
Similar threads
