VCORE higher than 1.55v?

[Synn21588]

Hey all. So I consider myself to be about a 7/10 in knowledge of PC's and how they work, still learning, still reading. Recently I found myself wondering about VCORE voltages. I know the FX series chips are generally safe to overvolt to 1.5 - 1.55 volts, but can they be run higher? The reason I ask, is because (by error only) I found myself pushing 1.57-1.6 volts on my chip not to long ago when setting up my overclock, and my load temperatures, albeit pushing the limits, were still relatively acceptable. They were around 57-60 on the chip and 65-68 on the socket at that voltage. I'm not sure if I just got lucky with my cooling or the hardware itself, but I didn't consider that to be an outright "unacceptable" load temperature range, though I do recognize that it is quite high and nearing the limits. So can these chips run at voltages that high and have a reasonable amount of longevity, provided the cooling is adequate? Or will the chip suffer catastrophic failure at such high voltages?

TL;DR: Can you push higher volts than 1.55 provided the cooling keeps the temps below the thermal margin?

 

[Johan45]

According to this guide AMD recommends max 1.5v for air cooling. http://www.amd.com/Documents/AMD_FX_Performance_Tuning_Guide.pdf
It really is relevant to what you are using to cool. That being said, many have run "high" voltages for extended periods and not "killed" their chip. I see this more at OCN than here. Now high voltage may decreasese the life span by some degree but how long do you intend to run this chip? These are your parts so in the end it's up to yo but myself I wouldn't have a problem with it but I may be the wrong one to be giving advice on voltage limits. Ha ha

 

[Dolk]

I do not recommend running a CPU above the voltage limit, unless extreme cooling is present. The reason comes from complications within the silicon. A CPU that runs above 1.5v 24/7 will create too much charge on the fets, causing random spikes of energy to be released into the CPU. Energy has to go somewhere, and that somewhere mostly ends up destroying the CPU. The time in which it takes a CPU to die depends on too many factors to be calculated accurately.

Stick with 1.5V or lower, your CPU will stay happy and live a long healthy life.

 

[Johan45]

Here's the chart from that PDF. There's a big gap between the high end of air cooling and low end of extreme. I just figured water would give you a bit more room.

 

[Synn21588]

According to this guide AMD recommends max 1.5v for air cooling. http://www.amd.com/Documents/AMD_FX_Performance_Tuning_Guide.pdf
It really is relevant to what you are using to cool. That being said, many have run "high" voltages for extended periods and not "killed" their chip. I see this more at OCN than here. Now high voltage may decreasese the life span by some degree but how long do you intend to run this chip? These are your parts so in the end it's up to yo but myself I wouldn't have a problem with it but I may be the wrong one to be giving advice on voltage limits. Ha ha

Im running a LEPA ExLLusion 240 AIO with the rad fans swapped out to SP120's. I'd like to get a year or two out of the chip. Technology advances so quickly these days. I feel the same way, 1.6-1.7 volts, as long as I can keep the chip within the thermal margin, seems acceptable to me.

- - - Updated - - -

I do not recommend running a CPU above the voltage limit, unless extreme cooling is present. The reason comes from complications within the silicon. A CPU that runs above 1.5v 24/7 will create too much charge on the fets, causing random spikes of energy to be released into the CPU. Energy has to go somewhere, and that somewhere mostly ends up destroying the CPU. The time in which it takes a CPU to die depends on too many factors to be calculated accurately.

Stick with 1.5V or lower, your CPU will stay happy and live a long healthy life.

Makes sense. The transistors can only handle a finite amount of electrical energy, I'm just wondering what that limit really is...

- - - Updated - - -

Here's the chart from that PDF. There's a big gap between the high end of air cooling and low end of extreme. I just figured water would give you a bit more room.

View attachment 171396

Yeah, all things considered, seems like 1.6v wouldn't be too bad on water.

 

[Johan45]

If it were me though I would set it up with offsets and leave power saving on in bios then the voltage would cycle up and down as needed instead of a static 1.6v continually

 

[Synn21588]

If it were me though I would set it up with offsets and leave power saving on in bios then the voltage would cycle up and down as needed instead of a static 1.6v continually

No offsets available with my board's BIOS, but I've always turned back on all of the power saving features after finding a stable overclock. My board only lets me increase voltages in 0.025 increments which is really annoying, but overall I am really pleased with my GA-990FXA-UD5 R5

- - - Updated - - -

Current setup is:

FX-8320e OC'd to 4.6ghz @ 1.512v
LEPA ExLLusion 240 AIO with SP120's in push on the rad
GA-990FXA-UD5 R5 mobo
ASUS STRIX r9 380
Patriot Signature Series DDR3-1600 16gb
128GB SSD dedicated to the OS and drivers
1.5TB partitioned disk space

- - - Updated - - -

Double edit:

For some reason my board really doesn't like overclocking with the FSB AT ALL!! Thats the only other gripe I have about this board. Best I've managed was 215 FSB with multi adjusted for stock frequency, any more and it fails to POST :<

 

[Dolk]

Makes sense. The transistors can only handle a finite amount of electrical energy, I'm just wondering what that limit really is...

I wouldn't be able to calculate it without AMD's help

 

[Synn21588]

I wouldn't be able to calculate it without AMD's help

Damn AMD and their corporate secret recipes!!!!

 

[Johan45]

One thing I do know is that the 9xxx series have a default VID of 1.55v and up. So AMD is willing to take the chance, the caveat is they say it needs WC and sold them with a 120 AIO

 

[JrClocker]

Im running a LEPA ExLLusion 240 AIO with the rad fans swapped out to SP120's. I'd like to get a year or two out of the chip. Technology advances so quickly these days. I feel the same way, 1.6-1.7 volts, as long as I can keep the chip within the thermal margin, seems acceptable to me.

Just remember that power dissipation goes up as:

Voltage_Increase^2 * Frequency_increase

For example, if your 10% increase in frequency requires a 20% increase in voltage, then your power would increase by:

1.2*1.2*1.1 = 1.584 or 158.4%

A 10% increase in frequency with a 10% increase in voltage would be:

1.1*1.1*1.1 = 1.331 or 131.1%

 

[Synn21588]

Just remember that power dissipation goes up as:

Voltage_Increase^2 * Frequency_increase

For example, if your 10% increase in frequency requires a 20% increase in voltage, then your power would increase by:

1.2*1.2*1.1 = 1.584 or 158.4%

A 10% increase in frequency with a 10% increase in voltage would be:

1.1*1.1*1.1 = 1.331 or 131.1%

Maybe being 10 hours into a 12 hour shift is killing me, but what do you mean by this?

 

[JrClocker]

The temperature you get depends upon the power dissipation and the thermal impedance of your cooling system:

Your cooling system has to get rid of the heat, but it has built in resistance. You have to be able to get rid of the heat (power).

For a given frequency, increasing the voltage by 10% (1.1) will increase the processor power dissipation of the processor by 121% (1.1 * 1.1 = 1.21).

For a given voltage, increasing the frequency by 10% (1.1) will increase the processor power dissipation by 10% (1.1).

If you increase both at the same time, you multiple the 2 (1.21 * 1.1 = 1.311 or 131.1%).

http://www.overclockers.com/forums/showthread.php/765728-Overclock-Temp-Calculations-101

 

[Synn21588]

The temperature you get depends upon the power dissipation and the thermal impedance of your cooling system:

Your cooling system has to get rid of the heat, but it has built in resistance. You have to be able to get rid of the heat (power).

For a given frequency, increasing the voltage by 10% (1.1) will increase the processor power dissipation of the processor by 121% (1.1 * 1.1 = 1.21).

For a given voltage, increasing the frequency by 10% (1.1) will increase the processor power dissipation by 10% (1.1).

If you increase both at the same time, you multiple the 2 (1.21 * 1.1 = 1.311 or 131.1%).

http://www.overclockers.com/forums/showthread.php/765728-Overclock-Temp-Calculations-101

Interdasting

 

[ShrimpBrime]

At about 5.2ghz your going to be around 300w over 1.5500v. Give or take depending on voltage.

Lower temps yield use of lower volts.

Pc hang or freeze is sign of low volt. Pc shut down is thermal protection by board or surge protection by psu.

Edit. One more thing. 68c socket temp is not acceptable. Most boards (asus in particular) while on auto throttle at 65c.

Seek better cooling.

 

[Synn21588]

At about 5.2ghz your going to be around 300w over 1.5500v. Give or take depending on voltage.

Lower temps yield use of lower volts.

Pc hang or freeze is sign of low volt. Pc shut down is thermal protection by board or surge protection by psu.

Edit. One more thing. 68c socket temp is not acceptable. Most boards (asus in particular) while on auto throttle at 65c.

Seek better cooling.

Yeah I noticed the throttling. At 4.6ghz w/ 1.512v temps are around 52 on the cpu and 60 on the socket. So without a custom loop, it just doesn't look like i'll be able to push my CPU to 4.7 or higher. Maybe on a better board, but with my current setup, 4.6 is the max I can OC to with stability. Still, a 3.3ghz processor pushed to 4.6ghz is nothing to sneeze at.

 

[JrClocker]

Still, a 3.3ghz processor pushed to 4.6ghz is nothing to sneeze at.

39 % performance boost is great in my book.

 

[ShrimpBrime]

Yeah I noticed the throttling. At 4.6ghz w/ 1.512v temps are around 52 on the cpu and 60 on the socket. So without a custom loop, it just doesn't look like i'll be able to push my CPU to 4.7 or higher. Maybe on a better board, but with my current setup, 4.6 is the max I can OC to with stability. Still, a 3.3ghz processor pushed to 4.6ghz is nothing to sneeze at.

The fx processor responds very well to chilling. The 8300 under LN2 only two cores maxed at 7685mhz at 2.0 volts on Asus M5A78L MX plus. While running less cores doesnt strain VRMs too terribly. This is one option you could try perhaps. Run as a quad and see how it effecfs thermally.

 

[Synn21588]

The fx processor responds very well to chilling. The 8300 under LN2 only two cores maxed at 7685mhz at 2.0 volts on Asus M5A78L MX plus. While running less cores doesnt strain VRMs too terribly. This is one option you could try perhaps. Run as a quad and see how it effecfs thermally.

I thought about dropping to 6 or 4 cores, but right now my rig is literally in pieces as I just ordered a Air 540 case. Also going to be returning my AIO because the mounting points broke and it's only 13 days old!!! Thats two LEPA AIO's ive had to return now. I really wanted them to be great, just because I like an underdog, and while their performance was competitive, they lack in quality it seems.