Overclock.net banner

Ivy Bridge safe voltages.

14K views 17 replies 4 participants last post by  TwoCables 
#1 ·
Just dropped a cheap i7-3770K into my old i5-2500K. I had Silicon Lottery delid it, so it is running quite cool. I've managed 4.7 GHz stable at 1.45V under full AVX load. The question is - is this voltage safe or am I going to see some significant degradation? Max temperatures are 80°C.
 
#6 ·
Quote:
Originally Posted by pantsaregood View Post

1.45V is the VID being reported with additional turbo voltage. Default VID for 4.7 GHz is around 1.34V.
The VID isn't a voltage that is going to the CPU. It's informational. I can explain later (it's akin to the SPD values that show what the memory can run at). You have to go by the Core Voltage. What is it?
 
#7 ·
I'm aware that VID is the voltage the CPU is requesting. The VCore hitting the CPU is with 5mV of the VID at all times per HWMonitor. If you're trying to discern if Vdroop is landing my Vcore considerably lower than the 1.45V mentioned, no it isn't.

Anyway, this isn't getting any closer to answering the question. I'm aware of how to check voltages. The question is simply "at what point has everyone seen degradation on Ivy Bridge?"
 
#9 ·
I know VID is not the same as Vcore. What I'm trying to say is that my CPU's Vcore is closely matching its VID because of how LLC is configured.

If you want me to be more specific, VID is 1.446V and Vcore is 1.443V.

This tells you what voltage my CPU is receiving - around 1.45V. I already knew my CPU was receiving 1.45V. I'm trying to find some reference for what voltages Ivy Bridge begins to degrade at, not how to discern between VID and Vcore.
 
#10 ·
http://www.overclock.net/t/1291703/ivy-bridge-overclocking-guide-asus-motherboards/0_50
Quote:
Settings
Max voltage and temperature (Click to hide)
Voltage
Ivy Bridge is less susceptible to degrade versus Sandy Bridge and will be fine to run voltages over 1.35 as long as temperatures are good or permits.

Courtesy of Sin0822

Temperature
Ivy Bridge does get HOT, but do not let that scare you from overclocking your CPU or straying away from Ivy completely.
Thermal protection keeps your CPU from roasting too much and will turn off by itself once it reaches a certain temperature.
TJ Max for Ivy Bridge is 105C, but staying cool during stress-testing or overclocking is important! I like to stay below 85C, but during stress-testing, it's okay to go 95C.
 
#11 ·
Quote:
Originally Posted by pantsaregood View Post

I know VID is not the same as Vcore. What I'm trying to say is that my CPU's Vcore is closely matching its VID because of how LLC is configured.

If you want me to be more specific, VID is 1.446V and Vcore is 1.443V.

This tells you what voltage my CPU is receiving - around 1.45V. I already knew my CPU was receiving 1.45V. I'm trying to find some reference for what voltages Ivy Bridge begins to degrade at, not how to discern between VID and Vcore.
No, your CPU isn't receiving 1.45V. It's receiving 1.443V. The VID is simply an informational number that indicates what the CPU "thinks" it should need for that multiplier and that clock speed. It is not a voltage that is being applied to the CPU. The only way to know what voltage is being applied to the CPU is by looking at the Core Voltage.

http://www.overclock.net/t/665362/vid-voltage-identification-explained

Yes, the post is old, but the information still applies just as much now as it did then.
 
#12 ·
Quote:
Originally Posted by TwoCables View Post

No, your CPU isn't receiving 1.45V. It's receiving 1.443V. The VID is simply an informational number that indicates what the CPU "thinks" it should need for that multiplier and that clock speed. It is not a voltage that is being applied to the CPU. The only way to know what voltage is being applied to the CPU is by looking at the Core Voltage.

http://www.overclock.net/t/665362/vid-voltage-identification-explained

Yes, the post is old, but the information still applies just as much now as it did then.
Again, I'm aware that VID and Vcore are not the same. This doesn't change the fact that my CPU is closely matching its rated VID (1.446V VID yields 1.443V Vcore for AVX, 1.422V VID yields 1.42V Vcore), nor does this discussion about VID vs. Vcore have any relevance in what constitutes safe core voltage.

The question was "do Ivy Bridge CPUs tend to degrade at 1.45V?"
 
#13 ·
Quote:
Originally Posted by pantsaregood View Post

Again, I'm aware that VID and Vcore are not the same. This doesn't change the fact that my CPU is closely matching its rated VID (1.446V VID yields 1.443V Vcore for AVX, 1.422V VID yields 1.42V Vcore), nor does this discussion about VID vs. Vcore have any relevance in what constitutes safe core voltage.

The question was "do Ivy Bridge CPUs tend to degrade at 1.45V?"
It's not at 1.45V. The VID is not a voltage that is being applied to the CPU.

Your CPU is at 1.443V. It doesn't matter if it's close to the VID. It doesn't mean that the VID is a voltage that is being applied to the CPU. It's not a voltage that's being applied to anything at all.

ANYWAY....

Degradation will really only occur if you have the CPU under heavy loads often, but it might only show up after a year or two - and you might only need to decrease the core clock or increase the voltage upon seeing degradation. In other words, yes, but it probably won't kill your CPU.

Again, do not go by the VID. Ever. Go by the Core Voltage.
 
#14 ·
Quote:
Originally Posted by pantsaregood View Post

Again, I'm aware that VID and Vcore are not the same. This doesn't change the fact that my CPU is closely matching its rated VID (1.446V VID yields 1.443V Vcore for AVX, 1.422V VID yields 1.42V Vcore), nor does this discussion about VID vs. Vcore have any relevance in what constitutes safe core voltage.

The question was "do Ivy Bridge CPUs tend to degrade at 1.45V?"
short answer:

yes.

a bunch of stuff to explain it:

through out the last few years, puts it ~3-5 years after ivy, a few folks who had been using 1.45-1.51 Vcore under water started to report/post about instability issues and needing to lower core speed. those individual reports would generate input from others and out of (i'm guessing) a dozen or so "samples" a few were ok but by ~5 years (just recent) i only recall one person stating no change.

on a side note:

though your temps look fine, that still seems like a lot of Vcore for 4.7 or did i read something wrong?

if so, you might want to consider dialing back to were you start to get diminishing returns for voltage esp. if you got the chip used.
 
  • Rep+
Reactions: pantsaregood
#15 ·
Quote:
Originally Posted by looniam View Post

short answer:

yes.

a bunch of stuff to explain it:

through out the last few years, puts it ~3-5 years after ivy, a few folks who had been using 1.45-1.51 Vcore under water started to report/post about instability issues and needing to lower core speed. those individual reports would generate input from others and out of (i'm guessing) a dozen or so "samples" a few were ok but by ~5 years (just recent) i only recall one person stating no change.

on a side note:

though your temps look fine, that still seems like a lot of Vcore for 4.7 or did i read something wrong?

if so, you might want to consider dialing back to were you start to get diminishing returns for voltage esp. if you got the chip used.
Thanks! Rolled back 50mV to a stable 4.6 GHz. 4.7 GHz behaves a bit different from any other CPU I've ever seen - it requires a fairly significant amount of extra voltage, but 4.8 GHz only needs an additional 0.015V over 4.7 GHz.
 
#16 ·
thumb.gif
 
#17 ·
1.45v is way too much for ivy bridge and will degrade it within 6 months from personal experience to the point that you won't be running the same overclock after this time.

Stay under 80c and 1.375v at the max 24.7. To be on the safe side I'd stay under 1.35v.
 
#18 ·
Quote:
Originally Posted by $ilent View Post

1.45v is way too much for ivy bridge and will degrade it within 6 months from personal experience to the point that you won't be running the same overclock after this time.

Stay under 80c and 1.375v at the max 24.7. To be on the safe side I'd stay under 1.35v.
But your degradation experience happened when you were Folding 24/7.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top