I got some junk DDR5 64GB 5200 CL40, and I get ~26k in GB5, was a noticeable increase in OS responsiveness with it over my old garbage 64GB 3200 DDR4 CL16 which scored around 20k. Both just running on stock XMP.
TFLOPs are not the best metric for comparison since it doesnt account for driver/OS state. Geekbench 5 Metal score for 6900XT is 175k or so, for M1 Ultra it is around 100k. Even a 6800 non XT gets 120k, the M1 Ultra GPU has scaling issues presumably due to the link between the dies.
Hopefully...
Yep that is exactly in line with the expected drop from lowering the max boost, about the same as a 300W 12900K just with a way higher multicore :clap:
Your settings are limiting max single core boost to 5.2GHz down from 5.8GHz. So I would expect around 1900-2000 in GB5 by frequency scaling alone. You can also monitor the core its using in Activity monitor -> Window -> CPU history and make sure when you do the GB5 run its staying on the first...
How bad is bad for the single core?
For the single core, do you have CPUFriend + DataProvider kexts? Without that single core will be really bad (like half of normal, normal for a 13900K on MacOS should be around 1900 at ~300W).
That is awesome, glad it worked out.
Are you doing 5.2/4.2 all core? And how much were you able to undervolt?
Would be interested to see the single core score at 200W as well.
Might be redundant, but did you try the upper PCIE 4x slot on that board? TB3/4 cards are all 4x and wouldn't work on the lower 1x slot. Otherwise it's likely a power delivery issue, I would guess most Z690/Z790 with the additional power connector would work.
Yeah OCCT is fine, it uses similar algos to P95 anyway for the CPU / Ram tests. You can rule out CPU / Mobo / Ram / PSU right away.
All you have left is GPU, Disk, BIOS, or least likely IMO an incapability with OC and your chipset/mobo I say least likely since your board is basically part...
Can/have you run Prime95 Small FFT for an hour or so in Windows to confirm the machine is stable?
If that passes try an hour of Prime95 on Blend mode. If both those pass your system itself is most likely (you really need to do several hours of both to validate) stable and it would have to be...
Correct I never put my computer to sleep.
If you used CCC it would use APFS Snapshots to copy the data which would include any corruption already on the Adata drive. Also Samsung drives have issues with MacOS as far as I know.
That is similar to what I experienced, non specific corruption.
This setting under Energy Saver, stop the computer from entering sleep mode.
The first time the drive corrupted it was really obvious, almost every folder/file on the disk had incorrect permissions, in particular stuff in /etc. I...
I wish I could avoid it myself
- SSD issues
- Multimonitor issues (2 monitors with same IDs get corrupt settings after reboot from MacOS, reboot from Windows to MacOS fixes the issue, happened on my M1 Mac Mini too)
- Safari locks up machine since 12.5? Honestly not sure when it started as I...
MacOS 12.4 / 12.5, possible it was some application triggering the corruption, however I did 4 clean installs and ended up with 4 different corruption issues the hard lockups/resets were common among all the installs. Also I am using the exact same apps / config with the Sabrent drive and have...
The ADATA SSD uses the same controller as the Kingston KC2500 which I had corruption issues of MacOS (which caused hard lockups/random reboots). The same drive works perfectly in Windows/Linux. I would avoid using this drive in MacOS so you dont have compounding issues.
The Kingston Fury should...
Any chance you can take a screenshot of Intel Power Gadget, specifically the Power and Frequency graphs after R23 has run the Multicore test? Would be interested to see the W usage and Boost ceiling. Would mean you can compare after installing the ILM replacement as well.
This is my graphs...
With the 13900K it won't lower the temps as 100degC is basically the thermal target now just due to the sheer number of cores and their density, similar to the new AMD chips targetting 95degC.
What it will do is get the heat out of the CPU and into the AIO faster meaning it has a much higher...
I would try one of these first...
https://www.aliexpress.com/wholesale?trafficChannel=main&d=y&CatId=0&SearchText=thermalright+frame<ype=wholesale&SortType=total_tranpro_desc&page=1&groupsort=1
Check here for review/instructions
I dropped peak temps by 10degC+ on the 12700KF continuous R23...
Glad to hear its working now.
Enable Multicore enhancement in the BIOS, it unlocks the power limits and sets more aggressive voltage without it I get 37-38k.
On Windows I get 41.5k with and 39k without Multicore enhancement in R23.
You need good cooling though, mine peaks at 360W power for the...
Check here
https://dortania.github.io/OpenCore-Install-Guide/troubleshooting/extended/kernel-issues.html#stuck-on-eb-log-exitbs-start
For me personally it was this setting SetupVirtualMap
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.