Joined
·
15 Posts
I will probably disable the E-core as my main purpose is to play game with it. Need to spend more time test out the setting. Did you delid your 12900? Is it worth it?
Overclocking ADL - 12900k etc results, bins and discussion
5741 - 5760 of 5906 Posts
Joined
·
656 Posts
Why tm5 extreme is only 27m? At least 1 hour. And y cruncher for 7000c30 is kinda low. Not sure because u run 5.2 core and 2600 ring or what.
Joined
·
2,003 Posts
@sugi0lover knows much more about the differences of Stock vs Relid Liquid Metal vs Direct Die, but it essentially goes from Stock > -10C > -17C (total values). So it's worth it.
Joined
·
1,498 Posts
I'd imagine he's trying to test the ram and making sure his ring clock isn't causing any crashes, and forgot to set it back before running the tests lol
Probably part of the reason his Y Cruncher is looking weird.
59s is pretty normal score for 52 p core clocks on y cruncher with e cores enabled.
You need to disable e cores and use avx to get sub 56s.
Ring at 2600 I would guess is because he is not running static clocks and some power saving feature enabled in Windows. Mine will show 3600 when not using static clocks on some runs using High Performance Power Plan.
You need to disable e cores and use avx to get sub 56s.
Ring at 2600 I would guess is because he is not running static clocks and some power saving feature enabled in Windows. Mine will show 3600 when not using static clocks on some runs using High Performance Power Plan.
Joined
·
656 Posts
5.3/4.3/4.5. I remember sugi with his golden chip hit 55s at 5.5ghz.
![Blue Font Screenshot Software Electric blue]()
Joined
·
946 Posts
that depends entirely on which profile you run. The 1usmus v3 heats the memory up a lot hotter than the old Anita extreme or even absolute. I generally run tm5 for 30 min to run the memory hot and then HCI to over 1500% which is at least 12 hours.
y-chruncher was cpu all stock with other things running, like xtu, hwinfo, etc.
Joined
·
946 Posts
Nice run. I remember that run even at 5.5 he was running a stripped out OS to get 55s.
Joined
·
946 Posts
Ring was stock; so 3800 ghz on the 7000c30, and 4.5 on the 6800c30 run. (It is in the SS, not sure where he got 2600)
TBH, I wasn’t really going for speed, just using it as a stability checker. Pretty sure I had hwinfo, xtu, EVGA’s in OS OC tool, note pad, Aida 64 etc. all open and full fat Win 11 pro.
I shocked it was below 60s to be honest….
Joined
·
618 Posts
Nope, I don't have any stripped OS installed. I have two normal windows. One for Overclock test, so it's okay if that windows got crashed. The other windows is for normal usage once the setup is stable on the test Window.
Here is a run using Arctic 420 AIO and Karhu stable mem oc and hitting 55s, it is throttling so with better cooling would do better. OS regular Win 10 iso and install.
In the y cruncher thread @sugi0lover is hitting 52s.
www.overclock.net
In the y cruncher thread @sugi0lover is hitting 52s.
Punish it with y-cruncher
www.overclock.net
Attachments
Joined
·
946 Posts
Are you running with avx512 for those times? Really impressive. Nice job
Joined
·
618 Posts
Thanks! With E cores on, avx512 can't be active, so it was run without avx512. For E cores off, 52s record, avx512 can be active, so it was enabled for that e cores off run.
Joined
·
15 Posts
Ohh this guy has the same problem as me. Seems like Asus motherboard might bugged out if you change CPU.
Joined
·
407 Posts
Max turbo limit simply means that the highest frequency the processor can go has been reached. If it's lower than what you configured, possibly it could have to do with AVX instructions (used during Cinebench) which by default on the 12900K can go to 51x maximum. I'm not sure about the KS model. You might have to check out AVX offsets on your motherboard to work around this limit. Perhaps you're better off asking other users who have this or similar Asus motherboards in this regard.
Joined
·
2,003 Posts
@mike7877 @bscool @Groove2013 Stabilized 4,000 MHz Gear 1 with 4x8 GB Samsung B-die Single Rank, on the MSI Edge:
![]()
It was not easy. VDIMM could probably go lower; VCCSA is most likely tacked at that value (and that's with a binned DDR4 IMC).
Two mixed Samsung B-die kits from different manufacturers (Galax and Team Group), so the RTLs are displaced.
Those are pretty much the tightest timings I can slap on these DIMMs in 4x8 configuration. In 2x8, I can tighten them even further for both kits.
tCL is basically impossible at 15, and many of the other primaries/secondaries can be tightened further, but they will fail y-cruncher.
(Haven't tested higher multipliers or with E-cores on because it's only running with an NH-D15 for now... Water loop will come in a few weeks.)
If anyone's been hoping to run quad DIMMs on Gear 1, there is certainly hope.
I don't know whether the MSI Edge is Daisy Chain or T-Topology, but it's certainly feeling like a T-Topo to me right now. Or good enough to be one.
And for those curious, on Gear 2, it can boot 4x8 with Samsung B-die up to 4,400 MHz. 4,533 will POST but fail to reach Windows. With 2x8, max is 5,000+ MHz.
I expect even better results with my Micron B-die 4x16 GB kit once I actually get around to testing it.
It was not easy. VDIMM could probably go lower; VCCSA is most likely tacked at that value (and that's with a binned DDR4 IMC).
Two mixed Samsung B-die kits from different manufacturers (Galax and Team Group), so the RTLs are displaced.
Those are pretty much the tightest timings I can slap on these DIMMs in 4x8 configuration. In 2x8, I can tighten them even further for both kits.
tCL is basically impossible at 15, and many of the other primaries/secondaries can be tightened further, but they will fail y-cruncher.
(Haven't tested higher multipliers or with E-cores on because it's only running with an NH-D15 for now... Water loop will come in a few weeks.)
If anyone's been hoping to run quad DIMMs on Gear 1, there is certainly hope.
I don't know whether the MSI Edge is Daisy Chain or T-Topology, but it's certainly feeling like a T-Topo to me right now. Or good enough to be one.
And for those curious, on Gear 2, it can boot 4x8 with Samsung B-die up to 4,400 MHz. 4,533 will POST but fail to reach Windows. With 2x8, max is 5,000+ MHz.
I expect even better results with my Micron B-die 4x16 GB kit once I actually get around to testing it.
Attachments
Joined
·
1,658 Posts
Yeah it's always at 5.5G...never drops below. I never configured loadline stuff thats maybe why the its not showing the right watts...I set llc to 6 and left the rest at auto...
Also avx offsets are 0
Does anyone know what's the right settings for LLC4 or LLC6? Ia ac ia dc?
5741 - 5760 of 5906 Posts
Join the discussion
