CPU Degredation Over (almost) 2 Years

[MageTank]

8 hours ago, Zyndo said:

Let me explain...

 

When i said, "You do not appear to have read my post. Or you do not understand what I was saying" You responded with "What about my post makes you think I did not read yours?". I never said you didn't read it. I said you either didn't read it or you didn't understand it. Both parts of that statement are relevant to the statement. Just as in my original post, you ignored part of it in order to debate the other.

 

If you had read the entire post AND understood it, you could have come to the same conclusion as Intel Core did...

That was the exact point I tried to make. Your going on and on about throttling and boosting and a whole slough of other stuff was truly irrelevant to the topic. I thanked you for it because it was accurate information, and someone may find it useful, but it simply had nothing to do with what we were talking about.

 

Turns out I WAS wrong about the voltage part, and Lays saw to correcting me on that. You even tried it with your "proof". Nothing says proof like some random guy saying random things on the internet. Your idea "proof" for temperature vs voltage was just as valid as my wack ass statement about it. If you want to "prove" something you have to be more than just right about it. You have to actually PROVE it. Like Lays did when he commented on his 6700k temps when under dry ice vs water. That is proof. He made an assertion and backed it up with evidence. I then went to back my assertion up with evidence and was unable to. the evidence I found actually confirmed his comment, and also allowed me to find out where I ultimately was mislead.

 

 

 

But for real. This conversation has nothing to do with the OP's topic. If you did not read or did not understand the information in this post either, then I will not be commenting again to help you through it. have a good day.

I did understand it. You were just simply wrong. It's not my job to assume what you meant (which is what Intelcore did). You can't just change the context of your words to better suit the situation when you are wrong. Nothing about my post shows that I did not understand yours. You said "impact in any way". Boost throttling is impacted by thermals. I could have easily told you to just google thermal runaway and moved on, but I gave you a detailed explanation to better educate you on the impact that high thermals have on CPU's.

 

I could have also pointed out that you were wrong when you said:

11 hours ago, Zyndo said:

My apologies for my ignorance. I was mislead when I watched a LTT video that had the underlying message that "thermals do not directly affect performance". And whilst that video held true and their testing methodology was solid, it didn't address this particular issue of temperature affecting voltages and stability.

 

Thank you for correcting my mistake =)

That video is flawed. They tested completely backwards. You can't just lower the temperatures, leave voltage/clocks the same, and expect a measurable difference in performance to be seen. While it is true that colder silicon has a difference in resistance, the performance gains (outside of increasing volts and clock speeds) are almost completely immeasurable under consumer workloads. The title is simply misleading. "Does temperature impact performance?" Yes, it does. When your temperatures get too high, you run the risk of throttling, and your overclocking headroom is reduced. To use that video as evidence that temperatures do not impact performance AT ALL, is just ignorance. 

 

As for the relevance of the information I brought up to this thread, its 100% relevant. OP is likely dealing with the side effects of thermal runaway (even if its only very mild). As temps get higher, power consumption increases, which causes even more higher temps, and even more power consumption. It's a constant fight that can only be solved by better cooling solutions. It's pretty much the biggest factor to silicon degradation. Granted, it is not extreme at all. In fact, its perfectly in line with what we expect from overclocked components (most CPU's run at higher stock voltages than what is required, so you will likely never notice instability as thermal runaway occurs over the normal lifespan of the product). 

 

I am not out to make you look stupid. The way you worded your post just made it seem as if temperatures do not matter at all. To expect me to come to the exact same conclusion as someone else (that only assumed what you meant) just isn't right. If you said "within X range of temps and under X workloads" then we wouldn't be having this conversation. 

[harrynowl]

6 hours ago, Zyndo said:

I'm not going to say its impossible for me to be wrong. I'm relatively quite new to computers and overclocking, but from my understanding temperatures do not in any way affect the performance or operation of a CPU. Having lower temperatures will certainly give you more headroom to push your CPU harder if you want to (which is kind of like getting more performance out of better temperatures), but it shouldn't have a direct effect on its operation in any way.

 

Have I learned something wrong at some point?

Temperatures have a direct correlation with resistance in a CPU (Positive temperature co-efficient). Which mainly affects a few things to my knowledge:

 

- Voltage requirement

- Degradation (via Current Crowding)

- CPU overclocking (as a result of the first one).

 

For example at ambient room temp with water my 3570K will do 5GHz @ 1.34V, at -25c it does 5.45 at the same voltage and 5.7 at -50c.

 

But a CPU at 4.5GHz at -80c and a CPU at 4.5GHz at 80c will perform the same (provided its the same CPU), one will just require less volts.

 

___________________________________________________

 

Anandtech did a bit on degradation a while ago, although it was applying to the 45nm Hi-K MG used at the time the principle is still the same.

 

Quote

The answer is quite simple really. Each processor, regardless of silicon quality, is capable of sustained error-free operation while functioning within the bounds of the specified environmental tolerances (temperature, voltage, etc.), for a period of no less than the warranted lifetime when no more performance is demanded of it than its rated frequency will allow. In other words, rather than limit the useful lifetime of each processor, and to allow for a consistent warranty policy, processors are binned based on the highest achievable speed while applying no more than the process's maximum allowable voltage. When we get right down to it, this is the key to overclocking - running CPUs in excess of their rated specifications regardless of reliability guidelines.

 

As soon as you concede that overclocking by definition reduces the useful lifetime of any CPU, it becomes easier to justify its more extreme application. It also goes a long way to understanding why Intel has a strict "no overclocking" policy when it comes to retaining the product warranty. Too many people believe overclocking is "safe" as long as they don't increase their processor core voltage - not true. Frequency increases drive higher load temperatures, which reduces useful life. Conversely, better cooling may be a sound investment for those that are looking for longer, unfailing operation as this should provide more positive margin for an extended period of time.

 

 

The graph above shows three curves. The middle line models the minimum required voltage needed for a processor to continuously run at 100% load for the period shown along the x-axis. During this time, the processor is subjected to its specified maximum core voltage and is never overclocked. Additionally, all of the worst-case considerations come together and our E8500 operates at its absolute maximum sustained Tcase temperature of 72.4ºC. Three years later, we would expect the CPU to have "degraded" to the point where slightly more core voltage is needed for stable operation - as shown above, a little less than 1.15V, up from 1.125V.

Including Vdroop and Voffset, an average 45nm dual-core processor with a VID of 1.25000 should see a final load voltage of about 1.21V. Shown as the dashed green line near the middle of the graph, this represents the actual CPU supply voltage (Vcore). Keep in mind that the trend line represents the minimum voltage required for continued stable operation, so as long as it stays below the actual supply voltage line (middle green line) the CPU will function properly. The lower green line is approximately 5% below the actual supply voltage, and represents an example of an offset that might be used to ensure a positive voltage margin is maintained.

The intersection point of the middle line (minimum required voltage) and the middle green line (actual supply voltage) predicts the point in time when the CPU should "fail," although an increase in supply voltage should allow for longer operation. Also, note how the middle line passes through the lower green line, representing the desired margin to stability at the three-year point, marking the end of warranty. The red line demonstrates the effect running the processor above the maximum thermal specification has on rated product lifetime - we can see the accelerated degradation caused by the higher operating temperatures. The blue line is an example of how lowering the average CPU temperature can lead to increased product longevity.

I know someone whos G3258 was folding for about 2 years and it went from 4.6 1.22V to like 4.4 1.27V.

[MageTank]

3 minutes ago, harrynowl said:

Temperatures have a direct correlation with resistance in a CPU (Positive temperature co-efficient). Which mainly affects a few things to my knowledge:

 

- Voltage requirement

- Degradation (via Current Crowding)

- CPU overclocking (as a result of the first one).

 

For example at ambient room temp with water my 3570K will do 5GHz @ 1.34V, at -25c it does 5.45 at the same voltage and 5.7 at -50c.

 

But a CPU at 4.5GHz at -80c and a CPU at 4.5GHz at 80c will perform the same (provided its the same CPU), one will just require less volts.

 

___________________________________________________

 

Anandtech did a bit on degradation a while ago, although it was applying to the 45nm Hi-K MG used at the time the principle is still the same.

 

I know someone whos G3258 was folding for about 2 years and it went from 4.6 1.22V to like 4.4 1.27V.

This is 100% correct. Your temperature directly impacts the resistance of the silicon itself. Whether or not temperatures impact performance, depends on a few variables:

 

Are the temps high enough to throttle the CPU? Is it a CPU that has thermal/power target limits (T SKU's, U SKU's, etc), is your overclock limited by thermals? If the answer to any of this is yes, then reducing the temperature (through delidding or upgrading your thermal solution) will result in an increase in performance. In your analogy, you are 100% correct that a CPU that is clocked exactly the same, and is running hotter (but not throttling) it will perform exactly the same. 

[done12many2]

14 hours ago, Webrider99 said:

As well cooling hasn't changed much, I've been using a Swiftech H320 the whole time, only difference is I had to RMA it to have the pump replaced. Which didn't affect temps at all, improved them if anything.
I've also cleaned the loop with vinegar/water solution a couple months ago cause there was gunk build up and Swiftech was kind enough to send me coolant concentrate to replace the old fluid. So loop is clean that's not an issue (anymore).

 

I don't see where you mentioned the disassembly of your CPU water block to thoroughly clean it after this "gunk build up" incident.  Simply running vinegar/water through the loop will not effective clean any build ups in the CPU water block.  You need to disassemble it and scrub the channels on the plate with a brush and some alcohol to remove any build up.

 

If you combine this with what others have stated about your TIM and the impact of temperatures on required voltages, I think it's safe to assume that you're not experiencing "degradation".  This word is just thrown around too much.

 

 

[KnightSirius]

19 hours ago, Savageiste said:

Factor in thermal paste and your cooler. If you reapply new thermal paste and clean up your cooler whether heatsink or radiator and fan(s) clean up completely

As I mentioned in my original post, and again above. I recently rebuilt so TIM isn't an issue, nor is the cleanliness of my rad.
It could be my actual CPU block though, not sure. Again I wish I had a custom loop not in use that I could throw in to see if the temps change much hmm...

[FPS-Russia]

So what happen to people running 4ghz phenom ii with same voltage for 5 years? Father has this as hand down from me and is still run 5 years later.

[KnightSirius]

8 hours ago, done12many2 said:

 

I don't see where you mentioned the disassembly of your CPU water block to thoroughly clean it after this "gunk build up" incident.  Simply running vinegar/water through the loop will not effective clean any build ups in the CPU water block.  You need to disassemble it and scrub the channels on the plate with a brush and some alcohol to remove any build up.

 

If you combine this with what others have stated about your TIM and the impact of temperatures on required voltages, I think it's safe to assume that you're not experiencing "degradation".  This word is just thrown around too much.

 

 

Wasn't very clear on that part. I did disassemble and clean the block with ketchup (acidic) and a toothbrush. Then I ran vinegar water through the system over night and cleaned the block with what it accumulated again. I repeated this 4 times and there was nothing left coming out of the loop.

It looked like crap when I first opened it, I attached a photo. Afterwards it was well cleaned also attached a picture.
 

So I am definitely experiencing degradation to some degree.

[KnightSirius]

3 minutes ago, FPS-Russia said:

So what happen to people running 4ghz phenom ii with same voltage for 5 years? Father has this as hand down from me and is still run 5 years later.

Dunno, different chips, different architectures, different manufacturers, lots of variables. Can't really be compared to my scenario.

[FPS-Russia]

Ok i see you now.