12900ks 24/7 O.C safe?

[prayth90]

Recently clocked my 12900ks for a 24/7 application. Using XTUI found what seems to be a stable core voltage of 1.370V. After finding this I locked that voltage in on the bios because even locked on XTU I had variable and unstable core voltages. After that I went on to push my p an e cores to 55,55, 53 remaining and 42 for the e cores with TB time maxed and TVB R23 runs will peak 95C @ 338W. My two questions here come from lake of experience. 1) is this a reasonable clock or do I run a high risk here 2) if this is reasonable is there anyone within suggestions on going a bit further.

12900ks
MMSI MEG Z690 ACE
DDR5 5200MHZ
3090TI STRIX OC
2x 360x44mm rad
1x 360x58mm rad
Quantum velocity2
2x d5 serial @1gpm

 

[ratirt]

Maybe you should focus only on pcores. I would leave ecores out.

 

[ir_cow]

If your at 5.5GHz @1.37V and below 100c, you have a golden cpu. Congrats

 

[prayth90]

Thank you I will look at only p cores then and bring e cores back to 40. Apologies if I miss led on p core clock speed I don't believe it is. Two are at 5.5 the rest are at 5.3. Thanks again

 

[Count von Schwalbe]

If that is a fixed OC, and you can't get it higher than 95C with no thermal limit offset, than I can't see any reason it could have an issue.

 

[GerKNG]

as usual on intel since skylake:
stay under 1.35V (under full load!) and under 1.45v idle if you want to keep your CPU for longer than 2-3 years without degrading it.

 

[ir_cow]

as usual on intel since skylake:
stay under 1.35V (under full load!) and under 1.45v idle if you want to keep your CPU for longer than 2-3 years without degrading it.

12th Gen is similar. 1.35V~ . 5.3 GHz is still good. But you will be pulling a lot at that freq. I mean you already see 330+ watts in R23 . Only way I can do that without hitting 100c is to delid the CPU. I'm guessing you did the same?

 

[Deleted member 185158]

338W.

12900ks

2x 360x44mm rad
1x 360x58mm rad
Quantum velocity2
2x d5 serial @1gpm

Yes very nice.
Very able to dissipate 1153 BTU/hr

Pic?!

 

[prayth90]

Yes a QA

12th Gen is similar. 1.35V~ . 5.3 GHz is still good. But you will be pulling a lot at that freq. I mean you already see 330+ watts in R23 . Only way I can do that without hitting 100c is to delid the CPU. I'm guessing you did the same?

Yes delid and conductonaut went nicely. Results were favorable. I'm also looking into one of the contact frames.

 

[dgianstefani]

1.37 is borderline too much as others have said.
Personally I would change the tune to 2x5.4 and the rest at 5.3, and see if you can get that with voltage of 1.35 or lower v.

If its cool though, you may be OK, just be aware the cpu may degrade.

 

[Vario]

In my recent experience, bone stock they boost to 1.4V and 100C, either Intel set these up to fail long term or they may be tougher than expected.

 

[dgianstefani]

In my recent experience, bone stock they boost to 1.4V and 100C, either Intel set these up to fail long term or they may be tougher than expected.

That's temporary burst voltage on a couple of cores under load. Completely different to a static oc set at a constant voltage.

 

[prayth90]

Yes very nice.
Very able to dissipate 1153 BTU/hr

Pic?!

View attachment 262824

Yes give me a bit I'll get a decent shot for you id say the only downside is the case went to micro center picked up the case I liked within my budget ( no reaserch). So I have a o11d xl

Yes very nice.
Very able to dissipate 1153 BTU/hr

Pic?!

Downside o11d XL I know didn't know

1.37 is borderline too much as others have said.
Personally I would change the tune to 2x5.4 and the rest at 5.3, and see if you can get that with voltage of 1.35 or lower v.

If its cool though, you may be OK, just be aware the cpu may degrade.

Thank you and everyone here it has been great to get some more specific info I will be going back to bios base clock settings and lock the core voltage at 1.350 do a couple r23 runs and game for a while to check the stability.

In my recent experience, bone stock they boost to 1.4V and 100C, either Intel set these up to fail long term or they may be tougher than expected.

I noticed similar results this all started with my first set up no custom loop 360 aio didn't stand a chance with no adjustments

 

[P4-630]

1.37 is borderline too much as others have said.
Personally I would change the tune to 2x5.4 and the rest at 5.3, and see if you can get that with voltage of 1.35 or lower v.

If its cool though, you may be OK, just be aware the cpu may degrade.

And besides benchmarking, how much do you really win with this OC in real world usage , at stock speeds it's already a top CPU.
Is it worth all that extra heat?

 

[prayth90]

And besides benchmarking, how much do you really win with this OC in real world usage , at stock speeds it's already a top CPU.
Is it worth all that extra heat?

I have been coming to the same conclusion it all started with my attempt to lower stock temps. It has now turned into a bit of game and to broaden my understanding of everything going on

 

[Vario]

I have been coming to the same conclusion it all started with my attempt to lower stock temps. It has now turned into a bit of game and to broaden my understanding of everything going on

Your cooling is much nicer. From starting at 5.2, I run half the wattage just dropping it -200 MHz to 5.0GHz on air. Any additional 100MHz comes at a steep heat premium.

 

[Deleted member 185158]

Yes give me a bit I'll get a decent shot for you id say the only downside is the case went to micro center picked up the case I liked within my budget ( no reaserch). So I have a o11d xl

That took a lot of time and effort for sure. Super clean looking, I'd rock that!

The overclock is respective of it's cooling. Hats off sir!

 

[ir_cow]

And besides benchmarking, how much do you really win with this OC in real world usage , at stock speeds it's already a top CPU.
Is it worth all that extra heat?

Not a whole lot. One could argue for dropping P1=P2 to 190. Lose 10% over stock and call it a day. But whats the fun in that if you already have a custom loop.

 

[Solid State Brain]

Intel specifications suggest that the default TJmax is 90 °C for the 12900KS, and IccMax still 280A (like the 12900K).
I think 24/7 safe settings should at the very least respect these values.