First time overclocking with AMD

[MarcusNumb]

Hi guys, I just want to share my story after my 1st time overclocking my little 3600x CPU with 2 different coolers.
My specs are:
CPU: 3600x
Coolers: CoolerMaster MA410M and Noctua D15s Chromax.black
Ram: Gskill Trident Z RGB 3200 cl16 2x8Gb
Case: Lianli II Mesh RGB, 3 front intake, 1 rear ex, 2 bot intake.

My first cooler was CoolerMaster MA410M. With this one, I could only OC my 3600x to 4.2Ghz all core, with 1.175v, after that no matter what I tried, I could not even reach 4.250. The temperature during 30min Cinebench r23 peaked at 78C. I thought that I did not win any lottery

Then after 10 months using that cooler, I could not resist the Noctua D15s Chromax.black so I bought it, with the idea of upgrading 3600x to 5900x. The funny thing is with the D15s Chromax.black, now my 3600x is reaching 4.4ghz all core with 1.32v, around 83C on Cinebench r23 test. 4.2Ghz with 1.175v, the difference in temperatures are 14degrees (78 then and 64 now). This makes me kinda confuse about my silicon lottery, whether i won one or not at all, but I'm just too happy with my Noctua.

I also need to mention that before this, the last time I did something with a PC was 2005, when I was in high school, so you guys can imagine that all of these stuffs nowadays is like a whole new world to me. The fulfilling feeling after doing research and trying over and over again to reach some goals with my own PC is still the same, like the first time I finished my Lego.

Now I'm kinda curious about OC my Ram. I tried 3 times with the 1usmus software, but it did not work. My goal is to reach 3600.

Please feel free to share your opinions or advice.

 

[Mussels]

ryzen has smaller dies than the intel chips, so they need better contact with a cooler, rather than just a large area
it's no shock that the premium coolers (yes, even air coolers) have better bases on them, and therefore work better

i'd take a Noctua over most AIO's in a heartbeat for Ryzen

 

[Tomgang]

I am as well overclocking amd cpu for the first time. Coming directly from a intel X58 system. I have not touched my 5600X much as this cpu is limited the low profile cpu cooler used. However i have played around with my 5950X and i got some really great results on air cooling compared to AIO and custom water loops in R23 and R20.

i have come to the conclusion, that for getting good result on amd 3000 and 5000 series cpu´s with aircooling. It´s important to have a cooler that have bofh the capacity to cool it, but also making good contact with the cpu all over the heat spreader. Thats because of Zen 2 and 3 chiplet desing. That makes good contact all over the heat spreader important and not only to the middle of the cpu like intels cpu with the die in the middle of the CPU and older AMD cpu a like.

with the setup i have. That is 5950X cooled by a Noctua NH-D15 chromax black with 2 of noctua Industrial IPPC 3000 rpm fans one is 120 MM do to memory high and one is 140 mm fan. I have not used the stock fans coming with the NH-D15 cooler, just wanted to point that out. I also used thermal grizzly kryonaut extreme paste as well. I dit these things to maximize my chances for great results on aircooling and let´s just say. This cooling setup is still exspensive compared to a good AIO, but the cooling capacity dit not dissapoint. Infact i got better results than i ever had hoped for.

So this cooling setup combined with my asus dark hero party trick called Dynamic OC switch. I got some really good results in cinebench R20 and R23. Just a quick comment: Dynamic oc switch is a feature only dark hero has as far i know. This allows me to combine PBO oc and manuel core overclock and the motherboard will swith between that on the fly depending on load. So on low load, i run in PBO mode and at high load it switch to my all core manuel oc automatic with out the need of reboot first. So that means i have boft the high single core clock and high manuelt core clock combined. Normally with out this, you have to chose between manuel oc and sacrifise high single core clock or stay on PBO for high single core clock but more limited all core clocks. With the dynamic oc swith i can do bofh at the same time.

Now with fine tuned PBO boost and made a manuel oc. I end up with a single core boost of up to 5.15 GHz (stock it never came over 5.05 GHz on even the best core) and a manuelt all core oc to 4.65 GHz at 1.375 volts. So yeah 16 cores all running 4.65 GHz on air cooling. Temp wise stock single core boost max out at 72C and 55C with all core loaded full. OC single core temp dit not change really, however all core temp dit. with the all core oc of 4.65 GHz at full and no power limits to hold the CPU back, temp peaked at 84C. So close to throttle temp, but it manage to stay below.

Also with the scores i got. In the Cinibench thread. I dit not only get the higgest score for an air cooled 5950X. I got the secongs higgest multi score for a 5950X cpu and the third higest single core score for any of the CPU´s it seems and beating out several AIO´s and custom loops in the process with a air cooling setup. All throw i dit what i cut to maximize the cooling capacity. Well the results is also talking for it self i´ll say.

Now to my scores.

R20 with the combined oc

R23 stock cpu

R23 with the combined OC

If there are any one that dosent belive my claims. Then go tjeck the cinebench thread your self and messure up my scores with the leader board.

Post your Cinebench R23 Score

Download Cinebench R23 Benchmarks FFXV Benchmark Cinebench R23 "Multi" Scores are clickable, leading to the original post. NameCPUCoreClockSingleMultiCooling mirrormaxEPYC 7742 (x2)128C/256T@ 3160 MHz 100981 cbAir nepuEPYC 770264C/128T@ 2499 MHz 48844 cbAir Bret WeeksRyzen Threadripper...

www.techpowerup.com

ryzen has smaller dies than the intel chips, so they need better contact with a cooler, rather than just a large area
it's no shock that the premium coolers (yes, even air coolers) have better bases on them, and therefore work better

i'd take a Noctua over most AIO's in a heartbeat for Ryzen

Based on my own exsperience i just have with my 5950X and a Noctua NH-D15. I will say your claims is true. That´s what i have seen first hand now with my own setup.

 

[MarcusNumb]

I am as well overclocking amd cpu for the first time. Coming directly from a intel X58 system. I have not touched my 5600X much as this cpu is limited the low profile cpu cooler used. However i have played around with my 5950X and i got some really great results on air cooling compared to AIO and custom water loops in R23 and R20.

i have come to the conclusion, that for getting good result on amd 3000 and 5000 series cpu´s with aircooling. It´s important to have a cooler that have bofh the capacity to cool it, but also making good contact with the cpu all over the heat spreader. Thats because of Zen 2 and 3 chiplet desing. That makes good contact all over the heat spreader important and not only to the middle of the cpu like intels cpu with the die in the middle of the CPU and older AMD cpu a like.

with the setup i have. That is 5950X cooled by a Noctua NH-D15 chromax black with 2 of noctua Industrial IPPC 3000 rpm fans one is 120 MM do to memory high and one is 140 mm fan. I have not used the stock fans coming with the NH-D15 cooler, just wanted to point that out. I also used thermal grizzly kryonaut extreme paste as well. I dit these things to maximize my chances for great results on aircooling and let´s just say. This cooling setup is still exspensive compared to a good AIO, but the cooling capacity dit not dissapoint. Infact i got better results than i ever had hoped for.

So this cooling setup combined with my asus dark hero party trick called Dynamic OC switch. I got some really good results in cinebench R20 and R23. Just a quick comment: Dynamic oc switch is a feature only dark hero has as far i know. This allows me to combine PBO oc and manuel core overclock and the motherboard will swith between that on the fly depending on load. So on low load, i run in PBO mode and at high load it switch to my all core manuel oc automatic with out the need of reboot first. So that means i have boft the high single core clock and high manuelt core clock combined. Normally with out this, you have to chose between manuel oc and sacrifise high single core clock or stay on PBO for high single core clock but more limited all core clocks. With the dynamic oc swith i can do bofh at the same time.

Now with fine tuned PBO boost and made a manuel oc. I end up with a single core boost of up to 5.15 GHz (stock it never came over 5.05 GHz on even the best core) and a manuelt all core oc to 4.65 GHz at 1.375 volts. So yeah 16 cores all running 4.65 GHz on air cooling. Temp wise stock single core boost max out at 72C and 55C with all core loaded full. OC single core temp dit not change really, however all core temp dit. with the all core oc of 4.65 GHz at full and no power limits to hold the CPU back, temp peaked at 84C. So close to throttle temp, but it manage to stay below.

Also with the scores i got. In the Cinibench thread. I dit not only get the higgest score for an air cooled 5950X. I got the secongs higgest multi score for a 5950X cpu and the third higest single core score for any of the CPU´s it seems and beating out several AIO´s and custom loops in the process with a air cooling setup. All throw i dit what i cut to maximize the cooling capacity. Well the results is also talking for it self i´ll say.

Now to my scores.

R20 with the combined oc
View attachment 210528

R23 stock cpu
View attachment 210529

R23 with the combined OC
View attachment 210530

If there are any one that dosent belive my claims. Then go tjeck the cinebench thread your self and messure up my scores with the leader board.

Post your Cinebench R23 Score

Download Cinebench R23 Benchmarks FFXV Benchmark Cinebench R23 "Multi" Scores are clickable, leading to the original post. NameCPUCoreClockSingleMultiCooling mirrormaxEPYC 7742 (x2)128C/256T@ 3160 MHz 100981 cbAir nepuEPYC 770264C/128T@ 2499 MHz 48844 cbAir Bret WeeksRyzen Threadripper...

www.techpowerup.com

Based on my own exsperience i just have with my 5950X and a Noctua NH-D15. I will say your claims is true. That´s what i have seen first hand now with my own setup.

Dude, 5.150 with an air-cooled 5950x is insane!!!! A friend of mine also has 5950x, paired with Kraken z73 and he can reach 5.1 as his best shot. Congrats!!
Anw, for the D15, personally I think it's kinda pity not using their stock fans but it seems like totally worth it

 

[Tomgang]

Dude, 5.150 with an air-cooled 5950x is insane!!!! A friend of mine also has 5950x, paired with Kraken z73 and he can reach 5.1 as his best shot. Congrats!!
Anw, for the D15, personally I think it's kinda pity not using their stock fans but it seems like totally worth it

yeah 5.15 GHz is pretty good i think. But it just shows that good contact is just as important to have good cooling capacity.

About fans. I all ready tried something simular with my old X58 setup. Meaning an I7 980X with Noctua NH-D14 with a pair of Noctua industrial IPPC 3000 RPM fans and thermal grizzly kryonaut paste none extreme in this case. With that setup, i manage to get my I7 980X up to 4.75 GHz all core clock at 1.55 volts and benchmark stable. I dit it on a cold winter day throw. But 4.75 GHz on that CPU on aircooling is really good as well. So you can say, that i just to my past exsperience over to my new setup and dit the same thing again and as you can see, result dit not dissapoint at all. The downside to this setup besides price, is the noise pennelty. 3000 RPM fans i not so silent. But i can live with it.

Two benchmark of my old machine with the cpu past 4.7 GHz with a simular air cooling setup i run now. Just to show that, it´s not just AMD cpu that benefit from it.

So yeah, a proper aircooling setup can defently compete with water cooling. As long you are prepared to accept more noise.

 

[mstenholm]

@Tomgang your HWINFO for the run is strange - CPU package power is way to low and the temperatures seems a bit of as well. If I’m wrong then you indee have a nice setup . I’m not questioning the R23 result at all, I just doubt that an OC will lower the max consumption from 136 W to 73.5 W

 

[Tomgang]

@Tomgang your HWINFO for the run is strange - CPU package power is way to low and the temperatures seems a bit of as well. If I’m wrong then you indee have a nice setup . I’m not questioning the R23 result at all, I just doubt that an OC will lower the max consumption from 136 W to 73.5 W

I think it has something to do with the dynamic oc switch. When it switch to the manuelt oc mode, it also bypass any power limits and i think it confused HWINFO. Also i had not HWINFO open while running the bench runs. It takes up cpu and lowers scores. So i first open the software after i finish a benchmark run. Hence why it doesent show max temperature or max power load. Cause while it was open, the cpu pretty much just sets ilde.

But i can tell you that in the first runs with ryzen master open to keep an eye on temp before i jumped in to the final runs with all software closed down. The CPU peaked at 84C with all power limits by passed. So yeah she got hot. But what to exspect from 16 cores running 4.65 GHz at full load with out power limits to hold it back. Then she sucks power.

 

[Mussels]

My 5800 peaks at 5.05 with boost, a 5950 should do that 100Mhz higher (with +200MHz in PBO) on single cores, at least - so thats exactly where it should be

 

[MarcusNumb]

My 5800 peaks at 5.05 with boost, a 5950 should do that 100Mhz higher (with +200MHz in PBO) on single cores, at least - so thats exactly where it should be

How is the temperature with your 5800x ?? I'm still wondering between 5800 and 5900, or maybe I should wait for the Zen3+ ?

 

[Mussels]

How is the temperature with your 5800x ?? I'm still wondering between 5800 and 5900, or maybe I should wait for the Zen3+ ?

around 60C

It's on custom water in winter, with tweaked PBO settings - did NOT behave like this on stock with an AIO

 

[MarcusNumb]

around 60C

It's on custom water in winter, with tweaked PBO settings - did NOT behave like this on stock with an AIO

Yeah I imagined that number must have come from a custom cooling. How was it with an AIO and what AIO was it ?

 

[gerardfraser]

Yeah I imagined that number must have come from a custom cooling. How was it with an AIO and what AIO was it ?

If interested my old 5800X on a cheap $40 USD MasterLiquid Lite ML240L RGB AIO

Single core hit 5170Mhz and Multi 4850Mhz Effective clocks while running some normal test and with AMD Curve optimizer on around 70°C . Some AMD 5800X are hard to tame.

Test done on release with old BIOS so I assume AMD Curve optimizer is way better in Aug 2021. Scores not the highest on AMD 5800X I done but the test were all done around same time. PC gaming could hit 5200Mhz for e-peen points.
5800x test

 

[Space Lynx]

ryzen has smaller dies than the intel chips, so they need better contact with a cooler, rather than just a large area
it's no shock that the premium coolers (yes, even air coolers) have better bases on them, and therefore work better

i'd take a Noctua over most AIO's in a heartbeat for Ryzen

NOCTUA TO THE MOON!!!!

If interested my old 5800X on a cheap $40 USD MasterLiquid Lite ML240L RGB AIO

Single core hit 5170Mhz and Multi 4850Mhz Effective clocks while running some normal test and with AMD Curve optimizer on around 70°C . Some AMD 5800X are hard to tame.

Test done on release with old BIOS so I assume AMD Curve optimizer is way better in Aug 2021. Scores not the highest on AMD 5800X I done but the test were all done around same time. PC gaming could hit 5200Mhz for e-peen points.
5800x test

R20 is not really that great of a test... run the Witcher 3 or any demanding CPU game at max settings for about 40 mins... then report back

or just run Aida64 system stability test for 15 mins. then report back those temps.

 

[gerardfraser]

NOCTUA TO THE MOON!!!!



R20 is not really that great of a test... run the Witcher 3 or any demanding CPU game at max settings for about 40 mins... then report back

or just run Aida64 system stability test for 15 mins. then report back those temps.

Well never Ran Witcher 3 on 5800X how about ,Shadow of the Tomb Raider is that a good enough game. Set to 5150Mhz for Boost. I do not own 5800X anymore ,I moved on.

@MarkcusNumb
3600x if interested hitting 4500Mhz Cinebench with same cheap cooler

 

[Space Lynx]

Well never Ran Witcher 3 on 5800X how about ,Shadow of the Tomb Raider is that a good enough game. Set to 5150Mhz for Boost. I do not own 5800X anymore ,I moved on.

@MarkcusNumb
3600x if interested hitting 4500Mhz Cinebench with same cheap cooler

bit off topic, but why did you get rid of your 5800x, honestly looks like you had a golden chip for it to run cool and boost high. even at @R-T-B has to wrangle his 5800x temps somewhat.

 

[gerardfraser]

bit off topic, but why did you get rid of your 5800x, honestly looks like you had a golden chip for it to run cool and boost high. even at @R-T-B has to wrangle his 5800x temps somewhat.

Well last two years CPU's I owned 2600X/2x3600X/3600XT/3800X/3800XT/5600X/5800X. Intel was on sale so I bought an Intel 10850K it was cheap. So I am crazy or just like changing gear. I have the 5600X and I might sell the 10850K and try out the 5600X for fun.

 

[Space Lynx]

Well last two years CPU's I owned 2600X/2x3600X/3600XT/3800X/3800XT/5600X/5800X. Intel was on sale so I bought an Intel 10850K it was cheap. So I am crazy or just like changing gear. I have the 5600X and I might sell the 10850K and try out the 5600X for fun.

as long as you are having fun that is all that matters I suppose.

 

[Mussels]

Yeah I imagined that number must have come from a custom cooling. How was it with an AIO and what AIO was it ?

corsair H115i loved being around 90C on stock settings

EK 240mm AIO dropped that to the 75-80 range, then i went custom for another 5-10C off, then i found the PBO settings

 

[MarcusNumb]

corsair H115i loved being around 90C on stock settings

EK 240mm AIO dropped that to the 75-80 range, then i went custom for another 5-10C off, then i found the PBO settings

Wow, that's a lot hotter than I expected from a 5800x, I need to rethink about buying one

 

[R-T-B]

bit off topic, but why did you get rid of your 5800x, honestly looks like you had a golden chip for it to run cool and boost high. even at @R-T-B has to wrangle his 5800x temps somewhat.

Yep, I just let mine do stock boost. Manual OC for me is worthless.

 

[BryanNitro]

Hi guys, I just want to share my story after my 1st time overclocking my little 3600x CPU with 2 different coolers.
My specs are:
CPU: 3600x
Coolers: CoolerMaster MA410M and Noctua D15s Chromax.black
Ram: Gskill Trident Z RGB 3200 cl16 2x8Gb
Case: Lianli II Mesh RGB, 3 front intake, 1 rear ex, 2 bot intake.

My first cooler was CoolerMaster MA410M. With this one, I could only OC my 3600x to 4.2Ghz all core, with 1.175v, after that no matter what I tried, I could not even reach 4.250. The temperature during 30min Cinebench r23 peaked at 78C. I thought that I did not win any lottery

Then after 10 months using that cooler, I could not resist the Noctua D15s Chromax.black so I bought it, with the idea of upgrading 3600x to 5900x. The funny thing is with the D15s Chromax.black, now my 3600x is reaching 4.4ghz all core with 1.32v, around 83C on Cinebench r23 test. 4.2Ghz with 1.175v, the difference in temperatures are 14degrees (78 then and 64 now). This makes me kinda confuse about my silicon lottery, whether i won one or not at all, but I'm just too happy with my Noctua.

I also need to mention that before this, the last time I did something with a PC was 2005, when I was in high school, so you guys can imagine that all of these stuffs nowadays is like a whole new world to me. The fulfilling feeling after doing research and trying over and over again to reach some goals with my own PC is still the same, like the first time I finished my Lego.

Now I'm kinda curious about OC my Ram. I tried 3 times with the 1usmus software, but it did not work. My goal is to reach 3600.

Please feel free to share your opinions or advice.

Sup bud NITRO is I...

First thing first motherboards have WAY WAY WAY better OC profiles then any settings you can toggle, you just have to find them I like MSI advanced overclock overdrive profile 3 , look in amdoverclocking section in your bios test the profiles and pick the one you like the best. FOR FUN

This is what I've been overclocking the most...
RX6800 used that don't work very good 4 THEM...
5800x CPU

I hit 74.5 then they throttled me and took their cut but YAY... done it to every one XFX POWERCOLOR SAPPHIRE GIGABYTE ZOTAC NITRO CORN INSERT NAME HERE AND I still need to optimize the gpu frequency. I like all the little spins all the brands put into their designs...

 

[Mussels]

Wow, that's a lot hotter than I expected from a 5800x, I need to rethink about buying one

certain coolers suck for them, they dont put out that much heat

tweaking PBO is key to the same performance at drastically lower heat, if you dont have a cooler with great contact

 

[BryanNitro]

certain coolers suck for them, they dont put out that much heat

tweaking PBO is key to the same performance at drastically lower heat, if you dont have a cooler with great contact

doesn't matter she can get to 95c at 500+fps BFV so make it SCREAM

 

[MarcusNumb]

certain coolers suck for them, they dont put out that much heat

tweaking PBO is key to the same performance at drastically lower heat, if you dont have a cooler with great contact

Well im using the Noctua d15s chromax, I bought it for the ryzen 5000 so hopefully it is good enough

 

[BryanNitro]

Well im using the Noctua d15s chromax, I bought it for the ryzen 5000 so hopefully it is good enough

the biggie is ,,, just remember to have good air flow also break in that CPU run benchmarks and such they get hot but its also how they are designed to operate. I let it auto clock meaning just leave it and then run benchmarks and burn in software let it cool run it some more times check thermal paste re apply if needed do it some more until the temps load graph and operation level out.