Help with oc i7-9700k on Aorus Z390 Pro

[alonbl]

My rig has i7-9700k, Aorus Z390 Pro, Corsair 16GB 3200mhz, Noctua D15, NZXT H700.

I followed the Gigabyte guide to 5GHz and here is what I got.
i7-9700k stable at @ 5GHz 1.31v with avx offset -3

On Prime95 with AVX I get 95c with no crashes

On prime without AVX I get max 87c with no crashes
In games the cpu is around 75c

Is it safe temperatures for 24/7 or I need to add to the offset?

 

[Gaidax]

Those temps are higher than average. 75c when gaming and that's in winter? You're looking at 80+ in summer.

Technically these temps are usable, but I think you should drop a notch by 100MHz and some voltage, running warm is not worth it for something like extra 100MHz.

 

[Roland Of Gilead]

Too high for me. You want to keep temps below 80c for full load OC temps.

Gaming at 75c is again too high. With the Noctua you should really be hitting much lower. I would say about the 55-65c mark for gaming load. 70-80c for stress load.

Agree with Gaidax. Drop it back a bit. Keep your prime95 temps below 80c, and work back from there. As in, maybe even drop back 200mhz for the all core OC to keep temps below 80c.

 

[alonbl]

- I will check again in the games maybe I was wrong. If you ignore the 75c in-games, does the temps of the stress tests are ok or not?
- Do you mean 80c on prime95 with avx or without?
- Do I need to lower the voltage too?

I really wanted to hit the 5 mark 🙁

 

[Supahos]

If the voltage will work lower then lower it. That's always better as it lowers temps and slows electron migration

 

[Roland Of Gilead]

- I will check again in the games maybe I was wrong. If you ignore the 75c in-games, does the temps of the stress tests are ok or not?
- Do you mean 80c on prime95 with avx or without?
- Do I need to lower the voltage too?

I really wanted to hit the 5 mark 🙁

For me 87c without AVX is too high.

Yes, that's what I mean. As Supahos said, lowering voltage will reduce temps. Ideally, it's about balance between temps, voltage, speed, to achieve stable results.

Yes, some chips will get 5ghz, some won't. It's the silicon lottery. Each chip is different, sadly.

 

[alonbl]

I will check tonight and report back, thanks guys.
BTW is it possible that the thermal paste applied wrong or the screws are not sit correctly?

 

[Roland Of Gilead]

I will check tonight and report back, thanks guys.
BTW is it possible that the thermal paste applied wrong or the screws are not sit correctly?

Yes, both of those things are possible, so it's worth checking both.

 

[alonbl]

so here are the results-
with prime 95 no avx at stock speed the min temp is 35c (my room is about 25c i think) the max temp is 70c
with prime 95 no avx at 4.9ghz @ 1.31v the max temp is 82c on 3 cores and the rest around 79c-80c

BTW winter here is around 19c-29c so its not that cold

 

[Roland Of Gilead]

so here are the results-
with prime 95 no avx at stock speed the min temp is 35c (my room is about 25c i think) the max temp is 70c
with prime 95 no avx at 4.9ghz @ 1.3v the max temp is 82c on 3 cores and the rest around 79c-80c

BTW winter here is around 19c-29c so its not that cold

Okay, so ambient temps are playing a part in your PC/CPU higher temps. BUT, you still have to factor that in.

At 4.9@1.3 and hitting 82c, you're right on the edge of it for me. Others might disagree. I would not push the OC any further, and just go with that. Testing with prime at those settings for a few hours will tell you if it's absolutely stable. Don't just run it for 20 mins or so. For my OC's , i run prime overnight for 8 hours. 20 mins for basic testing.

 

[alonbl]

so here are the results-
with prime 95 no avx at stock speed the min temp is 35c (my room is about 25c i think) the max temp is 70c
with prime 95 no avx at 4.9ghz @ 1.3v the max temp is 82c on 3 cores and the rest around 79c-80c

BTW winter here is around 19c-29c so its not that cold

Okay, so ambient temps are playing a part in your PC/CPU higher temps. BUT, you still have to factor that in.

At 4.9@1.3 and hitting 82c, you're right on the edge of it for me. Others might disagree. I would not push the OC any further, and just go with that. Testing with prime at those settings for a few hours will tell you if it's absolutely stable. Don't just run it for 20 mins or so. For my OC's , i run prime overnight for 8 hours. 20 mins for basic testing.

sorry I just corrected myself - its 1.31v not 1.3v and now i see it reached 84c on 3 cores...
When I was on 5ghz @ 1.31 it was stable 12hr

so from 87c -> 5ghz its 84c -> 4.9ghz should I lower the hz to get lower temp?
Does 70c at stock speeds (prime95) is normal? Do you think re-applying the thermal paste would make a difference?

 

[Roland Of Gilead]

so here are the results-
with prime 95 no avx at stock speed the min temp is 35c (my room is about 25c i think) the max temp is 70c
with prime 95 no avx at 4.9ghz @ 1.3v the max temp is 82c on 3 cores and the rest around 79c-80c

BTW winter here is around 19c-29c so its not that cold

Okay, so ambient temps are playing a part in your PC/CPU higher temps. BUT, you still have to factor that in.

At 4.9@1.3 and hitting 82c, you're right on the edge of it for me. Others might disagree. I would not push the OC any further, and just go with that. Testing with prime at those settings for a few hours will tell you if it's absolutely stable. Don't just run it for 20 mins or so. For my OC's , i run prime overnight for 8 hours. 20 mins for basic testing.

sorry I just corrected myself - its 1.31v not 1.3v and now i see it reached 84c on 3 cores...
When I was on 5ghz @ 1.31 it was stable 12hr

so from 87c -> 5ghz its 84c -> 4.9ghz should I lower the hz to get lower temp?
Does 70c at stock speeds (prime95) is normal? Do you think re-applying the thermal paste would make a difference?

This is my opinion. I would lower it to 4.8 and voltage back to about 1.26-1.27. This seems to be what most people sit at (voltage wise). This will keep your temps below 80c, but also allow headroom to potentially boost further if thermals allow it.

70c at stock on Prime is normal for me. You can of course try a re-application of thermal paste it won't do any harm.

Reading your first post back, I am a bit baffled. Really just that prime didn't fail on any cores at 95c! That just doesn't add up.

Anyway, I think you have the idea. What you decide or end up on clockspeed wise, versus voltage is really up to you.

I always want to keep my max OC temps below 80c. That way you know it's a good balance of voltage/speed/stability.

 

[alonbl]

so here are the results-
with prime 95 no avx at stock speed the min temp is 35c (my room is about 25c i think) the max temp is 70c
with prime 95 no avx at 4.9ghz @ 1.3v the max temp is 82c on 3 cores and the rest around 79c-80c

BTW winter here is around 19c-29c so its not that cold

Okay, so ambient temps are playing a part in your PC/CPU higher temps. BUT, you still have to factor that in.

At 4.9@1.3 and hitting 82c, you're right on the edge of it for me. Others might disagree. I would not push the OC any further, and just go with that. Testing with prime at those settings for a few hours will tell you if it's absolutely stable. Don't just run it for 20 mins or so. For my OC's , i run prime overnight for 8 hours. 20 mins for basic testing.

sorry I just corrected myself - its 1.31v not 1.3v and now i see it reached 84c on 3 cores...
When I was on 5ghz @ 1.31 it was stable 12hr

so from 87c -> 5ghz its 84c -> 4.9ghz should I lower the hz to get lower temp?
Does 70c at stock speeds (prime95) is normal? Do you think re-applying the thermal paste would make a difference?

This is my opinion. I would lower it to 4.8 and voltage back to about 1.26-1.27. This seems to be what most people sit at (voltage wise). This will keep your temps below 80c, but also allow headroom to potentially boost further if thermals allow it.

70c at stock on Prime is normal for me. You can of course try a re-application of thermal paste it won't do any harm.

Reading your first post back, I am a bit baffled. Really just that prime didn't fail on any cores at 95c! That just doesn't add up.

Anyway, I think you have the idea. What you decide or end up on clockspeed wise, versus voltage is really up to you.

I always want to keep my max OC temps below 80c. That way you know it's a good balance of voltage/speed/stability.

1.26v-1.29v freeze with prime95
at 1.3v its stable but I get 84c on one core and 83c at 2 cores the rest are around 80c.
Doesn't seems like a big different compered to 5gz 1.31v 87c.

When I take off the front panel of the case I get:
prime95 no axv - 5ghz 1.31v 80c.
prime95 axv - 5ghz 1.31v 90c

 

[Roland Of Gilead]

hmmm. Just make sure to test with your final settings. 80-83c is maringally to high for me. But it won't kill your processor.

Just keep an eye on the temps, and if you find it's unstable and/or crashing take the OC off. Remember you CPU boosts to 4.9 anyway. Having an all core OC only benefits some scenarios.

Yes, the difference between 4.8 - 4.9, even 5.0 is negligible wen it comes to actual FPS performance increase in games.

 

[alonbl]

hmmm. Just make sure to test with your final settings. 80-83c is maringally to high for me. But it won't kill your processor.

Just keep an eye on the temps, and if you find it's unstable and/or crashing take the OC off. Remember you CPU boosts to 4.9 anyway. Having an all core OC only benefits some scenarios.

Yes, the difference between 4.8 - 4.9, even 5.0 is negligible wen it comes to actual FPS performance increase in games.

I was supprized that I needed 1.3v for 4.8ghz with 84c on 2 cores. is it normal?
Is it really matter if I run 5ghz 1.31v 87c VS 4.8ghz 1.3v 84c? seems like 5ghz for 3c more is a better choice?

 

[Roland Of Gilead]

hmmm. Just make sure to test with your final settings. 80-83c is maringally to high for me. But it won't kill your processor.

Just keep an eye on the temps, and if you find it's unstable and/or crashing take the OC off. Remember you CPU boosts to 4.9 anyway. Having an all core OC only benefits some scenarios.

Yes, the difference between 4.8 - 4.9, even 5.0 is negligible wen it comes to actual FPS performance increase in games.

I was supprized that I needed 1.3v for 4.8ghz with 84c on 2 cores. is it normal?
Is it really matter if I run 5ghz 1.31v 87c VS 4.8ghz 1.3v 84c? seems like 5ghz for 3c more is a better choice?

i think your missing the point. If you want to achieve highest clocks, then go right ahead and dial in whatever setting you feel is best.

For 100-200mhz increase, and 4-5c increase in temps + the bump in voltage, it's totally not worth it IMO

The point being, test it at both speeds, and i'm sure you will find practically no difference in performance. You might get a few more points in some synthetic benchmark, but that's about it. So is that worth a potentially unstable system, where it's very possible any gains you have will be negated by thermal throttling?

 

[alonbl]

hmmm. Just make sure to test with your final settings. 80-83c is maringally to high for me. But it won't kill your processor.

Just keep an eye on the temps, and if you find it's unstable and/or crashing take the OC off. Remember you CPU boosts to 4.9 anyway. Having an all core OC only benefits some scenarios.

Yes, the difference between 4.8 - 4.9, even 5.0 is negligible wen it comes to actual FPS performance increase in games.

I was supprized that I needed 1.3v for 4.8ghz with 84c on 2 cores. is it normal?
Is it really matter if I run 5ghz 1.31v 87c VS 4.8ghz 1.3v 84c? seems like 5ghz for 3c more is a better choice?

i think your missing the point. If you want to achieve highest clocks, then go right ahead and dial in whatever setting you feel is best.

For 100-200mhz increase, and 4-5c increase in temps + the bump in voltage, it's totally not worth it IMO

The point being, test it at both speeds, and i'm sure you will find practically no difference in performance. You might get a few more points in some synthetic benchmark, but that's about it. So is that worth a potentially unstable system, where it's very possible any gains you have will be negated by thermal throttling?

I cleaned the dust filter and also It turned out that I set the fan curve to 100% on all fans only when the cpu pass 66c.
Now I set it to 100% when on 60c.
in AC Odyssey and BF5 its mostly 62c-66c with peaks of 71c-72c.
Here is a screenshot of HWiNFO64:
https://imgur.com/a/IAYQuy3

 

[Roland Of Gilead]

hmmm. Just make sure to test with your final settings. 80-83c is maringally to high for me. But it won't kill your processor.

Just keep an eye on the temps, and if you find it's unstable and/or crashing take the OC off. Remember you CPU boosts to 4.9 anyway. Having an all core OC only benefits some scenarios.

Yes, the difference between 4.8 - 4.9, even 5.0 is negligible wen it comes to actual FPS performance increase in games.

I was supprized that I needed 1.3v for 4.8ghz with 84c on 2 cores. is it normal?
Is it really matter if I run 5ghz 1.31v 87c VS 4.8ghz 1.3v 84c? seems like 5ghz for 3c more is a better choice?

i think your missing the point. If you want to achieve highest clocks, then go right ahead and dial in whatever setting you feel is best.

For 100-200mhz increase, and 4-5c increase in temps + the bump in voltage, it's totally not worth it IMO

The point being, test it at both speeds, and i'm sure you will find practically no difference in performance. You might get a few more points in some synthetic benchmark, but that's about it. So is that worth a potentially unstable system, where it's very possible any gains you have will be negated by thermal throttling?

I cleaned the dust filter and also It turned out that I set the fan curve to 100% on all fans only when the cpu pass 66c.
Now I set it to 100% when on 60c.
in AC Odyssey and BF5 its mostly 62c-66c with peaks of 71c-72c.
Here is a screenshot of HWiNFO64:
https://imgur.com/a/IAYQuy3

That's not so bad. Those temps aren't too far from where I'd expect them, but still just on the edge of too much..

Typically, this is where temps should be (there are so many variables though, this is a loose guide, and also depends on ambient temps)

Idle: 25-low 30's
Gaming: 45-65 game dependent. There are one or two exceptions there. One or two games will push temps high enough, and may touch 70c.
Stress testing/max load: 70-80c.

Looking around some threads. it's seems like mid 80's for max load/stress is acceptable.

Unless you find a noticeable decrease in performance, I'd just go with the 4.8/4.9 with lower voltage, and temps at mid 80's with Prime and AVX offset. Don't go beyond 85c. See what you get. Often it's the minor tweaks that get the last few mghz out of a proc.

 

[alonbl]

hmmm. Just make sure to test with your final settings. 80-83c is maringally to high for me. But it won't kill your processor.

Just keep an eye on the temps, and if you find it's unstable and/or crashing take the OC off. Remember you CPU boosts to 4.9 anyway. Having an all core OC only benefits some scenarios.

Yes, the difference between 4.8 - 4.9, even 5.0 is negligible wen it comes to actual FPS performance increase in games.

I was supprized that I needed 1.3v for 4.8ghz with 84c on 2 cores. is it normal?
Is it really matter if I run 5ghz 1.31v 87c VS 4.8ghz 1.3v 84c? seems like 5ghz for 3c more is a better choice?

i think your missing the point. If you want to achieve highest clocks, then go right ahead and dial in whatever setting you feel is best.

For 100-200mhz increase, and 4-5c increase in temps + the bump in voltage, it's totally not worth it IMO

The point being, test it at both speeds, and i'm sure you will find practically no difference in performance. You might get a few more points in some synthetic benchmark, but that's about it. So is that worth a potentially unstable system, where it's very possible any gains you have will be negated by thermal throttling?

I cleaned the dust filter and also It turned out that I set the fan curve to 100% on all fans only when the cpu pass 66c.
Now I set it to 100% when on 60c.
in AC Odyssey and BF5 its mostly 62c-66c with peaks of 71c-72c.
Here is a screenshot of HWiNFO64:
https://imgur.com/a/IAYQuy3

That's not so bad. Those temps aren't too far from where I'd expect them, but still just on the edge of too much..

Typically, this is where temps should be (there are so many variables though, this is a loose guide, and also depends on ambient temps)

Idle: 25-low 30's
Gaming: 45-65 game dependent. There are one or two exceptions there. One or two games will push temps high enough, and may touch 70c.
Stress testing/max load: 70-80c.

Looking around some threads. it's seems like mid 80's for max load/stress is acceptable.

Unless you find a noticeable decrease in performance, I'd just go with the 4.8/4.9 with lower voltage, and temps at mid 80's with Prime and AVX offset. Don't go beyond 85c. See what you get. Often it's the minor tweaks that get the last few mghz out of a proc.

Thanks a lot for your time and effort I appreciate it

 

[Roland Of Gilead]

hmmm. Just make sure to test with your final settings. 80-83c is maringally to high for me. But it won't kill your processor.

Just keep an eye on the temps, and if you find it's unstable and/or crashing take the OC off. Remember you CPU boosts to 4.9 anyway. Having an all core OC only benefits some scenarios.

Yes, the difference between 4.8 - 4.9, even 5.0 is negligible wen it comes to actual FPS performance increase in games.

I was supprized that I needed 1.3v for 4.8ghz with 84c on 2 cores. is it normal?
Is it really matter if I run 5ghz 1.31v 87c VS 4.8ghz 1.3v 84c? seems like 5ghz for 3c more is a better choice?

i think your missing the point. If you want to achieve highest clocks, then go right ahead and dial in whatever setting you feel is best.

For 100-200mhz increase, and 4-5c increase in temps + the bump in voltage, it's totally not worth it IMO

The point being, test it at both speeds, and i'm sure you will find practically no difference in performance. You might get a few more points in some synthetic benchmark, but that's about it. So is that worth a potentially unstable system, where it's very possible any gains you have will be negated by thermal throttling?

I cleaned the dust filter and also It turned out that I set the fan curve to 100% on all fans only when the cpu pass 66c.
Now I set it to 100% when on 60c.
in AC Odyssey and BF5 its mostly 62c-66c with peaks of 71c-72c.
Here is a screenshot of HWiNFO64:
https://imgur.com/a/IAYQuy3

That's not so bad. Those temps aren't too far from where I'd expect them, but still just on the edge of too much..

Typically, this is where temps should be (there are so many variables though, this is a loose guide, and also depends on ambient temps)

Idle: 25-low 30's
Gaming: 45-65 game dependent. There are one or two exceptions there. One or two games will push temps high enough, and may touch 70c.
Stress testing/max load: 70-80c.

Looking around some threads. it's seems like mid 80's for max load/stress is acceptable.

Unless you find a noticeable decrease in performance, I'd just go with the 4.8/4.9 with lower voltage, and temps at mid 80's with Prime and AVX offset. Don't go beyond 85c. See what you get. Often it's the minor tweaks that get the last few mghz out of a proc.

Thanks a lot for your time and effort I appreciate it

No worries. Let me know how you get on. You can PM if you wish.

Have a great evening

 

[alonbl]

hmmm. Just make sure to test with your final settings. 80-83c is maringally to high for me. But it won't kill your processor.

Just keep an eye on the temps, and if you find it's unstable and/or crashing take the OC off. Remember you CPU boosts to 4.9 anyway. Having an all core OC only benefits some scenarios.

Yes, the difference between 4.8 - 4.9, even 5.0 is negligible wen it comes to actual FPS performance increase in games.

I was supprized that I needed 1.3v for 4.8ghz with 84c on 2 cores. is it normal?
Is it really matter if I run 5ghz 1.31v 87c VS 4.8ghz 1.3v 84c? seems like 5ghz for 3c more is a better choice?

i think your missing the point. If you want to achieve highest clocks, then go right ahead and dial in whatever setting you feel is best.

For 100-200mhz increase, and 4-5c increase in temps + the bump in voltage, it's totally not worth it IMO

The point being, test it at both speeds, and i'm sure you will find practically no difference in performance. You might get a few more points in some synthetic benchmark, but that's about it. So is that worth a potentially unstable system, where it's very possible any gains you have will be negated by thermal throttling?

I cleaned the dust filter and also It turned out that I set the fan curve to 100% on all fans only when the cpu pass 66c.
Now I set it to 100% when on 60c.
in AC Odyssey and BF5 its mostly 62c-66c with peaks of 71c-72c.
Here is a screenshot of HWiNFO64:
https://imgur.com/a/IAYQuy3

That's not so bad. Those temps aren't too far from where I'd expect them, but still just on the edge of too much..

Typically, this is where temps should be (there are so many variables though, this is a loose guide, and also depends on ambient temps)

Idle: 25-low 30's
Gaming: 45-65 game dependent. There are one or two exceptions there. One or two games will push temps high enough, and may touch 70c.
Stress testing/max load: 70-80c.

Looking around some threads. it's seems like mid 80's for max load/stress is acceptable.

Unless you find a noticeable decrease in performance, I'd just go with the 4.8/4.9 with lower voltage, and temps at mid 80's with Prime and AVX offset. Don't go beyond 85c. See what you get. Often it's the minor tweaks that get the last few mghz out of a proc.

Thanks a lot for your time and effort I appreciate it

No worries. Let me know how you get on. You can PM if you wish.

Have a great evening

Great evening to you too
quick update- I played Project Cars 2 and got 65c~