Contribute
Register

Z690 Chipset Motherboards and Alder Lake CPU

Thanks for EFI. I have GPU RX 5500 XT, Z690-UD DDR4 and an i9-12900K. I run this EFI, and I have lrgacy shim 2. I change USB but the same problem. Maybe you know what I need to do? Thanks.
Try to disable USBInjectAll (thanks to @CaseySJ)
 
On an unrelated news, Intel is moving to kill AVX-512 in Alder Lake. One wonders why Intel even bothered to introduce this instruction set on the desktop with Rocket Lake just a few months ago.
Well, the wait wasn't long:
Remember that the MacPro and iMacPro are the sole real Macs with AVX-512, that all Intel hacks run fine without it, as long as AVX-512-aware application can correctly probe the hardware to know when to load or not to load their AVX-optimised code (Ryzen hacks fail the probing part), and that all Z690 users who enable their E-cores, including those who use the latest ProvideCurrentCpuInfo quirk to enable all cores and threads, already run (fine!) without AVX-512 anyway. In practice, this should be a non-change to most.
All previous speculations and testing in this thread about AVX-512 can be set aside and disregarded.
 
Not such a difference bt Mac and Windows on E-cores:

Captura de pantalla 2022-01-05 a las 20.49.08.png
Captura de pantalla 2022-01-05 a las 20.49.34.png
 
Why do some Cinebench R23 benchmarks for the 12900k result in incredibly high scores? IE;

28,265 (12900k)


Most of the benchmarks here seem to hover around a score of 19,000 in Cinebench R23. Even the post above in this thread with the 12700k resulted in a score of 13,239 but then there's this:

22,801 (12700k)


My 4790k (stock speeds) scores 5,100. With the scores some people are getting, I would assuming I'd get up to around 5-6X the speed of my current processor. But when looking at older processors like 11900k, it gets a score of around 16,000 in R23. Did Intel really improve performance in Alder Lake over Rocket Lake by up to 70%?

Geekbench scores also seem to have these disparities.

Why are benchmarks results all over the place?
 
Why do some Cinebench R23 benchmarks for the 12900k result in incredibly high scores? IE;

28,265 (12900k)


Most of the benchmarks here seem to hover around a score of 19,000 in Cinebench R23. Even the post above in this thread with the 12700k resulted in a score of 13,239 but then there's this:

22,801 (12700k)


My 4790k (stock speeds) scores 5,100. With the scores some people are getting, I would assuming I'd get up to around 5-6X the speed of my current processor. But when looking at older processors like 11900k, it gets a score of around 16,000 in R23. Did Intel really improve performance in Alder Lake over Rocket Lake by up to 70%?

Geekbench scores also seem to have these disparities.

Why are benchmarks results all over the place?
28730 win vs 27600 hack at 5.2 GHz
 

Attachments

  • 200.png
    200.png
    691.1 KB · Views: 63
Why do some Cinebench R23 benchmarks for the 12900k result in incredibly high scores? IE;

28,265 (12900k)


Most of the benchmarks here seem to hover around a score of 19,000 in Cinebench R23. Even the post above in this thread with the 12700k resulted in a score of 13,239 but then there's this:

22,801 (12700k)


My 4790k (stock speeds) scores 5,100. With the scores some people are getting, I would assuming I'd get up to around 5-6X the speed of my current processor. But when looking at older processors like 11900k, it gets a score of around 16,000 in R23. Did Intel really improve performance in Alder Lake over Rocket Lake by up to 70%?

Geekbench scores also seem to have these disparities.

Why are benchmarks results all over the place?
Geekbench and cinebench are different comparisons.

So yes! There is an improvement!
 
Well, the wait wasn't long:
Remember that the MacPro and iMacPro are the sole real Macs with AVX-512, that all Intel hacks run fine without it, as long as AVX-512-aware application can correctly probe the hardware to know when to load or not to load their AVX-optimised code (Ryzen hacks fail the probing part), and that all Z690 users who enable their E-cores, including those who use the latest ProvideCurrentCpuInfo quirk to enable all cores and threads, already run (fine!) without AVX-512 anyway. In practice, this should be a non-change to most.
All previous speculations and testing in this thread about AVX-512 can be set aside and disregarded.
No one stops us to keep the older BIOS, and there are also hacks to enable AVX-512 on motherboards that have not a BIOS switch like Gigabyte ones. I think having AVX-512 could be great for all the 12500 and 12400 CPUs that have no E cores at all. I just wonder what apps actually benefit from these in the macOS where as the only systems having them are iMac Pro and MacPro 7,1.
 
Has anyone tried how well Final Cut Pro and Photoshop work with Alder Lake?
 
I get after
smcrtc : start and stack
Look at page 102 of this thread for my history: I don't remember - what exactly helped from the tips.
Maybe something wrong in BIOS: have you enabled "Above 4G Decoding", for example?
 
Back
Top