• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Question Zen 4 builders thread

Page 21 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
AdamK47 said:
I managed to get the unreleased 0703 BIOS for the Crosshair Extreme from someone on Overclock.net today. Been testing my Corsair Dominator Platinum 6200 memory with it. DDR5 6000 is currently stable with average timings using Prime95 512K+ FFT no AVX torture testing. 34-38-38-72 with tweaked sub-timings. CL32 is a no go. Will error within 4 hours. CL34 went for 8 hours so far.

The sub-timings do translate to what I was accustomed to with Intel. Wording is a bit different. Not going too crazy with them. Just going to do a middle ground from what I did with Intel and the JEDEC default I see in the BIOS.

Played with fan curves in the Asus BIOS. Default is of course too deafening at load. Went with all fans at 70% at 95 degrees, 60% at 80 degrees, and 50% at 20 degrees. Tied to the CPU. Much more tolerable.

I have precision boost overdrive enabled. Gives me a few hundred more points in Cinebench R20.

FCLK seems stable at 2100MHz. Tried 2200MHz. It seemed to run fine, but benchmark scores went down. Perhaps there is error correction overhead on an unstable clock. Going to test 2100MHz further.
Click to expand...

It would be extremely if AMD would provide “safe, yet fast” values at common speeds from 5600 and up for all timings and subtimings, so those of us without EXPO support have a better chance of success.
 
Asrock B650 lineup:
videocardz.com

ASRock AMD B650 motherboard specs leaked, B650E Taichi is E-ATX with 27 phase VRM - VideoCardz.com

ASRock’s mid-range B650 motherboard series exposed The Taiwanese company is preparing at least seven models of B650 motherboards, including three of the “Extreme” variant. We have already shared pictures of three B650 motherboards from ASRock, but now we have a full list. We can now confirm...
videocardz.com videocardz.com

I was looking at the B650E PG Riptide, how much does only having 17 power phases matters when only running 8 cores and not planning any OC?
 
biostud said:
I was looking at the B650E PG Riptide, how much does only having 17 power phases matters when only running 8 cores and not planning any OC?
Click to expand...
According to Bulidzoid not much, unless it’s a 30 amp power phase(which it is probably not). And especially if you’re not planning on doing any OC. Your 7700x will be well fed by those 17 phases.
 
Woke up today after another 8 hours of stability testing at DDR5 6000. Prime95 16 core 2 threads each (32 thread) no-AVX 512K+ FFT with 24GB of the 32GB allocated. No errors. That has been the hardest one to pass memory-wise and it looks to be stable.
Chose the middle ground on the memory timings to play it safe. May tweak them down further. I probably don't need 1.4V on the memory. May also try lowering that.

rm.png
 
Markfw said:
I can't get linux to install... It was installed yesterday, but I had to redo windows.
Click to expand...

Haven’t tried Linux yet — I’m going to attempt to put Manjaro on my old 1TB nvme though tomorrow perhaps. Just reinstalled Windows 11 on my new sn850x and still no sleep mode available, so it wasn’t an old Windows install problem. That seems like the only thing that isn’t working .. hopefully they fix it with a BIOS update soon. I’m on 1.07 and might try the 1.08 beta soon.
 
repoman0 said:
Haven’t tried Linux yet — I’m going to attempt to put Manjaro on my old 1TB nvme though tomorrow perhaps. Just reinstalled Windows 11 on my new sn850x and still no sleep mode available, so it wasn’t an old Windows install problem. That seems like the only thing that isn’t working .. hopefully they fix it with a BIOS update soon. I’m on 1.07 and might try the 1.08 beta soon.
Click to expand...
Linux 19 would not install or 20. But 21 is installed there just fine right now ! This is mint.
 
Sounds like the no sleep mode bug is due to selecting the EXPO memory profile — so say the good people on the ASRock Reddit. Seems like they have employees watching that Reddit and reporting to engineers.
 
AdamK47 said:
Woke up today after another 8 hours of stability testing at DDR5 6000. Prime95 16 core 2 threads each (32 thread) no-AVX 512K+ FFT with 24GB of the 32GB allocated. No errors. That has been the hardest one to pass memory-wise and it looks to be stable.
Chose the middle ground on the memory timings to play it safe. May tweak them down further. I probably don't need 1.4V on the memory. May also try lowering that.

View attachment 68595
Click to expand...
I find y-cruncher all stresstests enabled to be the hardest to pass with memory OC on Zen4.. Easy to use and finding errors much faster than all other stresstest in my experience.
Latest version have also been updated for Zen4 AVX512
 
igor_kavinski said:
Then it matters even less. That will sip power if it's anything like the 5800X3D. I don't think the V-cache will be able to clock as well as the cores on a 7700X.
Click to expand...

I expect it will clock the same or marginally better than the 7700X.

I am less sure about 7900X3D and 7950X3D parts though.
 
In typical AMD fashion, goodlike CCD0 is binned together with useless junk CCD1.
1664818118177.png

CCD0 can almost run PBO CO -30 allcore, while 3 cores on junk CCD1 have zero headroom left
1664818292135.png

CCD0 singlecore clockspeed is around ~5780mhz in prime95
CCD1 singlecore clockspeed is around ~5450mhz in prime95
1664818462079.png
Wish me better luck on the other 7950x i get in a few days ^^

*edit*
Have done a further comparison in Cinebench r23 with static allcore OC @ 1200mv core LLC5. (maximum PPT was ~187w and max load temp was 68 degrees)
CCD0 managed 5500mhz
CCD1 could only run 5200mhz.

= 40.5k points at lowish vcore
1664827300776.png
Seems like its a 300mhz delta between the CCD's both in ST and MT workloads
 
Last edited:
Those are at load clock speeds in Prime95? If so, that's actually great. At 100% load across all cores, with the previous test I mentioned, I would see CCD0 hover between 5.2 to 5.4GHz. CCD1 between 5.0GHz and 5.2GHz.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top