First time cpu oc? i9-9900k 5Ghz

[zayah]

Havent tried oc'ïng cpu before other than auto stuff on bios was curious to see with all core 5Ghz and manual voltages which I set to 1.275 with 360 aio push pull are these temps ok, good, average, bad? 

Wasnt sure how to test stability other than on cinebench for 10mins and got a score of 13113. mobo rog strix z390-f


 

[Zando_]

20 minutes ago, zayah said:

Wasnt sure how to test stability other than on cinebench for 10mins

Prime95 torture test on the smallFFT setting. You will hit far higher temps than normal though, as it puts the CPU under the highest possible synthetic load. But it will usually find any instabilities in under 30 minutes of running (either your PC will crash if it's really unstable, or it'll note a thread stopped due to instability in the Prime 95 window). 

Cinebench is decent for seeing if you're close, but it won't expose instability like P95 will, so you can think you're stable then get random BSODs when playing games and such. ASUS' Realbench is also a good stress test, though I don't think it's been updated in a while. 

[LIGISTX]

34 minutes ago, zayah said:

Havent tried oc'ïng cpu before other than auto stuff on bios was curious to see with all core 5Ghz and manual voltages which I set to 1.275 with 360 aio push pull are these temps ok, good, average, bad? 

Wasnt sure how to test stability other than on cinebench for 10mins and got a score of 13113. mobo rog strix z390-f


 

I use AIDA64 for 10 hours, both the default settings and then with all CPU options disabled except for FPU. 

 

P95 can be a bit..... over the top. Its so unrealistic it can end up demanding more power then the CPU ever would see in any real world use case. 

 

Also, I had a 9900k, I can tell you if your chip came out after the 9900ks's started to existed, your OC potential is going to be pretty subpar. Intel started binning pretty hard to get good chips for the ks SKU, so normal 9900k's were hard to OC. If you have an early one though, your chances are decent at 5 GHz.

 

I was able to do 5 GHz all core, but it took IIRC 1.37v, and my temps were insane even on custom water. Under AIDA64 I would sit at ~95 which is not exactly something that made me happy, but thankfully in games it would usually be low 60's with some spikes up to low 70's depending on the game and what was going on.

 

I can also say.... 300 MHz likely isn't worth the extra power draw or heat output. BUT, whats the fun in that, right? lol. 

[zayah]

17 minutes ago, LIGISTX said:

I use AIDA64 for 10 hours, both the default settings and then with all CPU options disabled except for FPU. 

 

P95 can be a bit..... over the top. Its so unrealistic it can end up demanding more power then the CPU ever would see in any real world use case. 

 

Also, I had a 9900k, I can tell you if your chip came out after the 9900ks's started to existed, your OC potential is going to be pretty subpar. Intel started binning pretty hard to get good chips for the ks SKU, so normal 9900k's were hard to OC. If you have an early one though, your chances are decent at 5 GHz.

 

I was able to do 5 GHz all core, but it took IIRC 1.37v, and my temps were insane even on custom water. Under AIDA64 I would sit at ~95 which is not exactly something that made me happy, but thankfully in games it would usually be low 60's with some spikes up to low 70's depending on the game and what was going on.

 

I can also say.... 300 MHz likely isn't worth the extra power draw or heat output. BUT, whats the fun in that, right? lol. 

Ill test aida or prime95 tomorrow! its really late here in finalnd. Thanks for ur input i appreciate it alot!! I got my 9900k before the ks existed. Also im not sure if the 300MHz is worth either when on auto core and auto voltages i sit at around 4.8mhz in cinebench and over 10c cooler. Wanted to just give it a try and hear about other peoples opinions whether it is worthdoing or not. should i try with manual voltage and auto core or would that even make any difference? 

[zayah]

32 minutes ago, Zando_ said:

Prime95 torture test on the smallFFT setting. You will hit far higher temps than normal though, as it puts the CPU under the highest possible synthetic load. But it will usually find any instabilities in under 30 minutes of running (either your PC will crash if it's really unstable, or it'll note a thread stopped due to instability in the Prime 95 window). 

Cinebench is decent for seeing if you're close, but it won't expose instability like P95 will, so you can think you're stable then get random BSODs when playing games and such. ASUS' Realbench is also a good stress test, though I don't think it's been updated in a while. 

ill try this tomorrow!! i really wasnt sure what to do or what to try with just had cinebench installed so i thought thatd be better than nothing. also would like to hear ur opinions whether this is worth even doing when on auto voltages and core i get ~10c cooler and like 4,8ghz on cinebench or if i should set a manual voltage but core on auto or idk? and thank u so much for the help!!

[LIGISTX]

3 hours ago, zayah said:

Ill test aida or prime95 tomorrow! its really late here in finalnd. Thanks for ur input i appreciate it alot!! I got my 9900k before the ks existed. Also im not sure if the 300MHz is worth either when on auto core and auto voltages i sit at around 4.8mhz in cinebench and over 10c cooler. Wanted to just give it a try and hear about other peoples opinions whether it is worthdoing or not. should i try with manual voltage and auto core or would that even make any difference? 

So, if auto is setting it to 4.8 GHz all core.... that may be slower then just full defaults.

 

9900k at default settings will run 2 threads at 5 GHz, and all threads at 4.7. You may actually gain more by having a few threads boost to 5 ghz when needed vs having them all locked at 4.8. If I wasn't able to get a 5 GHz all core, I would have left mine at stock.

[zayah]

5 hours ago, LIGISTX said:

So, if auto is setting it to 4.8 GHz all core.... that may be slower then just full defaults.

 

9900k at default settings will run 2 threads at 5 GHz, and all threads at 4.7. You may actually gain more by having a few threads boost to 5 ghz when needed vs having them all locked at 4.8. If I wasn't able to get a 5 GHz all core, I would have left mine at stock.

depending on a task it goes from 5ghz -4.8ghz 

[zayah]

11 hours ago, Zando_ said:

Prime95 torture test on the smallFFT setting. You will hit far higher temps than normal though, as it puts the CPU under the highest possible synthetic load. But it will usually find any instabilities in under 30 minutes of running (either your PC will crash if it's really unstable, or it'll note a thread stopped due to instability in the Prime 95 window). 

Cinebench is decent for seeing if you're close, but it won't expose instability like P95 will, so you can think you're stable then get random BSODs when playing games and such. ASUS' Realbench is also a good stress test, though I don't think it's been updated in a while. 

Im not sure i understand what went wrong?

 

[Zando_]

2 hours ago, zayah said:

Im not sure i understand what went wrong?
-snip-

First, I always check these options so it puts everything in one window:

 

"Merge Main & Comm & Workers" + "Merge All Workers"

As for why it stopped, Workers 2, 6, and 8 had hardware instability so they spat out and incorrect number. Probably up vCore a teeny bit and that should even out, as any actual major instability would have crashed the whole PC.

14 hours ago, zayah said:

Also im not sure if the 300MHz is worth either when on auto core and auto voltages i sit at around 4.8mhz in cinebench and over 10c cooler. Wanted to just give it a try and hear about other peoples opinions whether it is worthdoing or not.

300Mhz or above is usually worth it, 200Mhz (4.8 to 5.0) isn't unless you really need every last frame. I'd personally still go for 5.0 just because I like the nice even number, but your gaming experience shouldn't be noticeably affected.

[zayah]

3 hours ago, Zando_ said:

First, I always check these options so it puts everything in one window:

 

"Merge Main & Comm & Workers" + "Merge All Workers"

As for why it stopped, Workers 2, 6, and 8 had hardware instability so they spat out and incorrect number. Probably up vCore a teeny bit and that should even out, as any actual major instability would have crashed the whole PC.

300Mhz or above is usually worth it, 200Mhz (4.8 to 5.0) isn't unless you really need every last frame. I'd personally still go for 5.0 just because I like the nice even number, but your gaming experience shouldn't be noticeably affected.

Thank you so much for the settings this is so much easier to read LOL Ill give it another go tomorrow ill try maybe 1.28-1.3 and see how it goes, and yeah its like 200Mhz difference depends sometimes cores boost to 5Ghz but id say on average cpu heavy tasks its at 4.8 all core

[zayah]

18 hours ago, LIGISTX said:

So, if auto is setting it to 4.8 GHz all core.... that may be slower then just full defaults.

 

9900k at default settings will run 2 threads at 5 GHz, and all threads at 4.7. You may actually gain more by having a few threads boost to 5 ghz when needed vs having them all locked at 4.8. If I wasn't able to get a 5 GHz all core, I would have left mine at stock.

sorry my bad not on all core all the time but on cpu intensive tasks like during cinebench i see it more all core 4.8ghz than those boosts, but desktops and i think in games (i have to double check the in game one) its usually 2 or 3 threads on 5ghz