Does oc the memory/core clock too much cause the lifespan of gpu to diminish? Or is it only temps./overvolting?

comanzo

Honorable
Jun 18, 2016
150
0
10,690
Basically what title states. When I was overclocking my memory on my 1070, I decided to start with +400 MHz, thinking it was going to be 8400mhz, I forgot however, that it doubles, so instead it was 8800mhz. I was stable until the end of heaven benchmark. It was supposed to load the table with all the info. and instead, it crashed. I heard on some forums that going too far with memory/core clock can cause lifespan to diminish. Is this true? Or is my mistake of accidentally starting with +800mhz on memory not really going to do much to lifespan? Thanks in advance.
 
Solution


ok, lets make a test :) Run some stress test without overclock and observe the max voltage, then overclock and run the same test again and see the max voltage again. If it has no impact on voltage i think...

GRafkiD

Commendable
Aug 22, 2016
35
0
1,560
With newer cards it can barely diminish the lifespan of itself. Also your not overclocking the Memory, you're overclocking the Memory CLOCK speeds. As well as also work your way up when oc because every card is different
 

Mikel_4

Respectable
Oct 15, 2016
712
0
2,660
Reference GP104 such as founders edition comes with good power components (infineon caps, ferrite chokes, in house MOSFET drivers) when in use those caps can sustain 100 degree C, I don't think thermal and electrical conductivity will be main factor of degrading, more like its chemical and physical characteristic. Most likely will be come from matter like sulfur, chlorine, lead or any corrosive substance
 

comanzo

Honorable
Jun 18, 2016
150
0
10,690


I see. I thought the reason why they say to go up slowly was due to it diminishing life span if gone too far. Now I realize it was because every card is different and top to bottom is more difficult than working from bottom to top. On another side note, if I was to theoretically overclock it to 10ghz, which is the stock memory clock for the 1080, then wouldn't it technically be gddr5x memory at that point? From what I understood, gddr5x is simply gddr5 clocked at a higher speed. I know I probably will never get that type of overclock, however, I am just curious.
 

comanzo

Honorable
Jun 18, 2016
150
0
10,690


I believe there was a misunderstanding. I didn't up the voltage, but rather the clocks. Voltage remained the same. I do plan to keep this card for at least 2 generations. So maybe whatever comes after Volta will be my next upgrade, however, that is if I deem it worthy of an upgrade. So yes, I do plan to keep it for at least 4 years, if not, longer.
 

comanzo

Honorable
Jun 18, 2016
150
0
10,690
Ok so just for clarification, both of these situations lead to little or no difference to the lifespan of gpu.
1. GPU is clocked way too high, and is highly unstable. Only lasts for a short period of time.
2. A long period of time passes before you realize that your overclock wasn't so stable after all. In other words, a slight unstable overclock for a long period of time.
Just want to clarify if both situations lead to little or no difference to lifespan. Thanks.
 

GRafkiD

Commendable
Aug 22, 2016
35
0
1,560


You would not be running a gpu with an unstable clock so the question is redundant.

 

USAFRet

Titan
Moderator


The reason for increasing the OC slowly (step by step) is to verify the system actually works at the increased OC.

If we assume a base of 1.0, and a theoretical target of 2.0...
If you go straight to 2.0, and it crashes, you don't where the actual max was.

So you increase step by step.
1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8 ooohhhh...that's where it crashed. Back off to 1.7 and accept that.
 

comanzo

Honorable
Jun 18, 2016
150
0
10,690

Not true. There are people that do heaven benchmark and other gpu stress tests, thinking they found a stable gpu clock, only to find out later on when playing a game, that it's in fact unstable. Therefore, the clock was unstable.
 

GRafkiD

Commendable
Aug 22, 2016
35
0
1,560

I agree partially being that games do act differently than benchmarks, but if you run enough benchmarks you should crash if it is unstable.

 

JordanMihailov

Commendable
Dec 12, 2016
63
0
1,660


ok, lets make a test :) Run some stress test without overclock and observe the max voltage, then overclock and run the same test again and see the max voltage again. If it has no impact on voltage i think you can be fine.
 
Solution