[Kelutrel]

All right, cores #9 and #13 PBO stay at zero in the beginning.

I will set all other 14 cores to -10 and run my combo: CoreCycler + y-cruncher for single-core tests, and y-cruncher and Prime95 AVX2 for multi-core tests.

After testing, I’ll increase or decrease by 5 points until all cores are stable, After testing and adjusting all 14 cores, I may try something with cores #9 and #13.

Correct. As I added above, when you do this be sure that you have PBO Boost set to Disabled and PBO Scalar set to Auto, otherwise your cores 9 and 13 may need a positive offset to be stable. You may want to set "22-ZN4 ~ Kizuna" as test mode in CoreCycler if you have not yet done it.

 

[diablodbl]

Correct. As I added above, when you do this be sure that you have PBO Boost set to Disabled and PBO Scalar set to Auto, otherwise your cores 9 and 13 may need a positive offset to be stable. You may want to set "22-ZN4 ~ Kizuna" as test mode in CoreCycler if you have not yet done it.

PBO Boost you mean Boost Override? Should i not set +200mhz? When use it?

 

[Velome]

you have tested the new bios of asus 2505 small comparison I, that's not why my latency is really terrible on their latest bios which are in beta

 

[sblantipodi]

you have tested the new bios of asus 2505 small comparison I, that's not why my latency is really terrible on their latest bios which are in beta

View attachment 2679400 View attachment 2679401

Even your read speed is very very low.
Why only 70k MB/s?.

 

[Imprezzion]

Even your read speed is very very low.
Why only 70k MB/s?.

If it is 2000/2066 FCLK that read speed seems fine to me tho for 6200?

 

[sblantipodi]

If it is 2000/2066 FCLK that read speed seems fine to me tho for 6200?

sorry I was thinking at a dual CCD CPU but that is a single CCD one

 

[Kelutrel]

PBO Boost you mean Boost Override? Should i not set +200mhz? When use it?

You can set it to what you want but IF PBO Boost Override is not disabled THEN some of your cores (9 and 13 probably) will require a positive offset to be stable. An offset of zero may not be stable when using PBO Boost. In any case, on a 7950X3D the boost is minimal. Even if you set PBO Boost Override to 200MHz the boost will be 25-50MHz at most, and only on certain cores.

 

[ju-rek]

you have tested the new bios of asus 2505 small comparison I, that's not why my latency is really terrible on their latest bios which are in beta

On my 7800x3d they are good.

 

[diablodbl]

You can set it to what you want but IF PBO Boost Override is not disabled THEN some of your cores (9 and 13 probably) will require a positive offset to be stable. An offset of zero may not be stable when using PBO Boost. In any case, on a 7950X3D the boost is minimal. Even if you set PBO Boost Override to 200MHz the boost will be 25-50MHz at most, and only on certain cores.

For the single-core test, use CoreCycler + Y-cruncher (22-ZN4 ~ Kizuna ) (BKP, BBP, SFT, VST), for 20 minutes per core. Also, run Prime95 with AVX2 FFT 720-720 for 10 minutes per core, for a total or 30min per core test.
For multi-core testing, I use Y-cruncher with default settings (AVX2) for all iterations, running for 1 hour.

So far, so good; besides cores #9 and #13, which I haven't touched, all other cores are stable at -15.

Are there any other specific configurations that I should run?
Any specific config from corecycler/configs at master · sp00n/corecycler ?
What about change from 22-ZN4 ~ Kizuna to 19-ZN2 ~ Kagari?

 

[Kelutrel]

For the single-core test, use CoreCycler + Y-cruncher (22-ZN4 ~ Kizuna ) (BKP, BBP, SFT, VST), for 20 minutes per core. Also, run Prime95 with AVX2 FFT 720-720 for 10 minutes per core, for a total or 30min per core test.
For multi-core testing, I use Y-cruncher with default settings (AVX2) for all iterations, running for 1 hour.

So far, so good; besides cores #9 and #13, which I haven't touched, all other cores are stable at -15.

Are there any other specific configurations that I should run?
Any specific config from corecycler/configs at master · sp00n/corecycler ?
What about change from 22-ZN4 ~ Kizuna to 19-ZN2 ~ Kagari?

I don't know mate. I usually run CoreCycler for 20 seconds per core, for 20-30 minutes total, as I have seen that switching cores is what makes errors show up on my hardware. So a long test time on a single core is not as effective as a short test time switching continuously between cores. But that is all how I test for (or sometimes I do the same but with OCCT/AVX2), nothing heavy, 20 minutes and done. How you test for stability is up to you and your experience with your hardware. There is people that doesn't consider stability if Karhu has not reached at least a 30000%, so you see it's personal.

 

[soodksa434]

On my 7800x3d they are good.

View attachment 2679620

wow that's impressive is that stable on 7800x3d ?, also am interested in 8000mhz performacne could you benchmark small game just to compare with my tuned 6200mhz ram cl30 , you can test free benchmark game "final fantasy dawntrail benchmark" you can download that from their site , after you download set graphics settings to standard (laptop) run the test and after that press save score to show both score and fps, here are my result with 7800x3d and tuned 6200mhz mem cl30

 

[deceptXS]

Anyone know how much you can lap a 7800X3D before it has cooler incompatibility?

 

[Imprezzion]

Anyone know how much you can lap a 7800X3D before it has cooler incompatibility?

Depends on the cooler. My EK Nucleus AIO's mount / springs would compensate just fine. It can even be ran direct die without modifications to the mount height. AIO's that have standoffs with physical stops on them might have less room but still lapping doesn't remove that much material.

 

[deceptXS]

Depends on the cooler. My EK Nucleus AIO's mount / springs would compensate just fine. It can even be ran direct die without modifications to the mount height. AIO's that have standoffs with physical stops on them might have less room but still lapping doesn't remove that much material.

Great, I have a EK AIO Basic, just compared the mounts they're reasonably similar.

 

[ju-rek]

wow that's impressive is that stable on 7800x3d ?, also am interested in 8000mhz performacne could you benchmark small game just to compare with my tuned 6200mhz ram cl30

Processor at default settings?

 

[soodksa434]

Processor at default settings?

sorry i didn't provide full info , i used only curve optimizer -18 since its got unstable when i use both memory overclock and CO otherwise i can get CO -30 without memory overclock also i have fclk 2167 thats it really

 

[ju-rek]

sorry i didn't provide full info , i used only curve optimizer -18 since its got unstable when i use both memory overclock and CO otherwise i can get CO -30 without memory

Sorry, the first screenshot was on 6200CL26, the second one is on 8000CL34

 

[soodksa434]

View attachment 2679724

decent 5% boost on minimum fps thats not bad at all, now thinking about it you might get more performance with Windows 11 build but who knows at this point thanks for the test man

 

[GRABibus]

For the single-core test, use CoreCycler + Y-cruncher (22-ZN4 ~ Kizuna ) (BKP, BBP, SFT, VST), for 20 minutes per core. Also, run Prime95 with AVX2 FFT 720-720 for 10 minutes per core, for a total or 30min per core test.
For multi-core testing, I use Y-cruncher with default settings (AVX2) for all iterations, running for 1 hour.

So far, so good; besides cores #9 and #13, which I haven't touched, all other cores are stable at -15.

Are there any other specific configurations that I should run?
Any specific config from corecycler/configs at master · sp00n/corecycler ?
What about change from 22-ZN4 ~ Kizuna to 19-ZN2 ~ Kagari?

Kagari is much more difficult to pass than Kizuna.

 

[RedF]

sorry i didn't provide full info , i used only curve optimizer -18 since its got unstable when i use both memory overclock and CO otherwise i can get CO -30 without memory overclock also i have fclk 2167 thats it really View attachment 2679677

I can't imagine how tRCDRD 33 can be stable at 6200MT/s. That is 10.65 ns.