Question An argument against overclocking new computers.

[maharvey555]

I mean I can totally understand the desire to post the best benchmark for your build, but often times that's all we're shooting for without any kind of concomitant, salient change in performance. I'm glad for those of you who do, so we can better understand the limits, but frankly for me the risk to benefit ratio is too high. I think that inescapably your machine is going to be running hotter and will therefore live a shorter useful life. That being said my system is basically built for overclocking , but I will wait to deploy it until the new software catches up with it.

 

[R_1]

as to the heat that depends on the cooling.
my 1500x with stock cooling at stock speeds was running 6+ degrees C hotter than the same CPU overclocked to 4.0 with a larger cooler.
the larger cooler let me overclock beyond the turbo boost speeds while remaining cooler than it was at stock with stock cooling.
stock 3.5 boost 3.7 stock cooler installed Cinebench 20 1778 @67 degrees C
Overclocked to 4.0 with gammaxx 400 LED Cinebench 20 1997 @60.25 degrees C

 

[Eximo]

Couple of things

1) Longevity, overclocking should reduce the lifespan of the silicon, but enthusiasts who replace their systems regularly won't encounter this, so it becomes immaterial. If I wanted system to last, and I have a few that I do, they are not overclocked.

2) Getting 100% performance now, and then adjusting the system to be faster so you get 100% later. But you are discounting that you could have 112% now. And back to 100% later when you do overclock. If it were a time sensitive application it would get worse over time, but in the early days would be better with overclocking. I can see both sides, but either way you are losing that higher performance opportunity, even if it might make it more consistent. Basically the concept of over-provisioning.

3) There are measurable gains to be had in performance, so not sure where idea comes from. It can be small, but it can also be drastic depending on the task. Most benchmarks will eliminate variables, but even then you can still see improvements from overclocking.

4) Higher temperatures only relatively. Overclocking is certainly going to result in higher temperatures given the same cooling. But if I know I am overclocking I can get appropriate cooling for that goal and keep temperatures similar to that of a basic system.

5) For many it is the fun of making the attempt, not necessarily the results.

Honestly, there isn't a huge amount of risk these days if you are careful. Adding a few tenths of a volt for a 200-300Mhz improvement is within everyone's capabilities, usually. Going for maximum performance in benchmarks can be left to extreme overclockers.

Certainly not treating my current CPU well, but so far so good.

 

[zx128k]

I have an overclocked q6600 still going today.

 

[FullTank]

I have a bunch of R9 390s & 390x's, they are safe to overclock to 1100 1560 from 1040~1060 1500 ,, but not safe to overclock the memory to 1700~1724 which is their limit, even if they can handle it - shenanigans will soon happen after weeks and chip will die(live experience)

The 1100 1560 is actually a very decent boost believe it or not, as well as 1540 Vs 1500 gives more boost than 1700 Vs 1560

I have an core2duo E6400 lying around which is OCed to more than 50% boost on GHz on some decent mobo, with a stock cooler even and no issues in temps..

Had a 9800GT which outperformed every oced 9800GTX+ , only needed to flash it for extra values and to have fan speed always at 100% , but before you say anything - needed to have fan speed at 100% even on stock, because it artifacted on auto fan curve & froze then

Apparently I have golden chip of TR1920X which does 4.1GHz @ 1.32~1.344 (1.3125 in bios), it's a huge boost, but much more of a waste to watts usage since it's a power hungry chip on OC only.

Had unlucky i7 6700k chip, max it could do was 4.8@1.45 and 4.7@1.41 , but temps was not for everything, so I then done an alternative, 4.4GHz core & cache @ 1.35.. was pretty good & much colder

Nvidia GPUs generally can go sky high on OC which is good

Forgot to mention that I also have flashed timings on my R9 390/X's and that's a free 3% boost everywhere, though obviously due to a bit of a raise in FPS - temps raising a bit too, but it's a free extra if you already hit your limits on clocks

Ram timings, MHz, always a good boost to your system.. generally some values like 16 18 18 36 CR2 can go as low as 15 17 17 28~30 CR1 without changing the MHz and at the same amount of stability, and that's only main timings, timings overclocking experts also can adjust bunch of secondary timings for extra performance, and generally ram does not die today because of itself, it more oftenly does because something else has failed, like PSU or mobo

Enabling large memory pages in gpedit

Now combine all these small boosts together, CPU ghz,GPU,GPU timings(if available),ram MHz,ram timings ,memory pages, and you get yourself a much stronger machine than what you had, at very least 15%'s (at very least!) even if gpu didn't went far, yes, most of the boost will still be on your GPU side (unless your CPU is weakish for today's standards)

With all that said there is definitely some cons out there for some of the hardware, but not for all of them and in most cases it's worth it IF you know what you are doing and what to expect from what you are doing.

Overclock is totally worth it, try it, if you will be careful - you will break Nothing, just don't forget to monitor the temps, that's the same thing to do for both beginners & veterans overclockers

 

[maharvey555]

I'm not trying to be a wet blanket, but a quick glance at the topics in this forum lends some validity to my point. Many people are overclocking perfectly good systems for no noticeable improvement to performance, except on benchmark scores. I stand by my assertion that more heat = decreased longevity, and that until it buys you something tangible in terms of performance it isn't worth it!

 

[Eximo]

Personally keep my GPU at a cool 55C under a full load with a waterblock, no increase in voltage/power over what the card is limited to. Overclocked up from its base boost of 2012Mhz to 2100Mhz something, been a while since I looked. Key there is consistency. That 2012Mhz boost would only last a few minutes on the stock cooler, would settle around 1900Mhz.

So roughly 10% clock speed boost, but that translates to 10-15 FPS in the titles I run, since I am aiming for high refresh, that is a good thing.

CPU, not so much. But did add 500Mhz to its stock boost. I wouldn't call that insignificant. Sadly my 7700k is much like the 6700k above. Needs a lot of voltage to reach 5Ghz, but de-lidded and water cooled, it still keeps it well within spec. So an argument there could be made against a 'stock' cooler on a 7700k would run just as warm, but at a slower speed.

I have much more concern about the voltage, but again, if it pops tomorrow, I have funds to replace it. Even considered BIOS hacking a 9700k into my board, still might do that, even though there is little to gain performance wise.

I don't see it as a wet blanket, just a different philosophy. I treat my parts 'poorly' until I am done with them. Probably won't be selling this CPU on to anybody. GPU is fine, could always put the stock cooler back on it and sell that off.

If you argue that I spent $300+ on cooling the components, you would be right, but I would have spent it regardless, and it is protecting my investment, at what was at the time, top of the line hardware. If it lasts until DDR5 is on the market I will be happy.