ill test mine now how do i test and what do i need to look out for ?
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Poll: Ryzen 7950X3D, 7900X3D, 7800X3D
- Thread starter kindai
- Start date
you just get CoreCycler from https://github.com/sp00n/corecycler/releases, extract it into a directory. then u need to edit config.ini and set mode = AVX2 and FFTSize = Small. then you just run Run CoreCycler.bat, and wait. It will report any cores that error out.
Don't see this in the config though downloaded it ready copy your config over in dm and ill paste it in
Last edited:
I ran Karhu overnight and had no errors, I left the dram voltage as it was so I`ll try it with FCLK at 2100 and see how that holds up.Try these.
If stable, you can try FCLK at 2100. If you notice a drop in Karhu MBs, then lower FCLK.
Might need to put dram to 1.4v + VDD, not sure depends how good your kit is.
Make sure you have cpu cache set to enabled when using Karhu.
Yes, I remembered that from the earlier posts when you started helping people with their settings, which is much appreciated by the way.
No, default is Auto and Auto = disabled.
ohhhh so it won't boost clock the speeds? 5.7?
no, it will. i think you're thinking of CPB (core performance boost), that is on by default.
I’m sure it was off on mine. I had to manually change it to enable it.
oh just shows how much I know about overclocking ( not a lot ) ill keep everything on default except the ram and hope for best it gives haha
What results are people getting on curve optimiser with their 7800X3D ?
neg 20 ( all cores ) passed 12 hours of Y-cruncher, and neg 30 all cores has just passed a similar test / Ram deliberatly set low to rule that out
neg 20 ( all cores ) passed 12 hours of Y-cruncher, and neg 30 all cores has just passed a similar test / Ram deliberatly set low to rule that out
Last edited:
I’m on 7950x3d so not same exact. I’m currently testing idle after passing Y Cruncher for 12 hours. I have 25,28,30,30,30,30,30,30 on the 3d and all core 25 on second CCD.What results are people getting on curve optimiser with their 7800X3D ?
neg 20 ( all cores ) passed 12 hours of Y-cruncher, and neg 30 all cores has just passed a similar test / Ram deliberatly set low to rule that out
I have found that you can pass hours of testing only to find a WHEA in HWINFO64 when doing nothing on the desktop while making a cup of tea. I haven’t started the second CCD yet, but it will take time. Then I start memory.
Keep in mind that higher memory frequency/tighter timings often means slightly less curve optimiser margin in my testing, so make sure to rerun with your memory at normal frequency/timings too.
Here's my 7950X3D CO Settings.
Last edited:
Good point - thanksKeep in mind that higher memory frequency/tighter timings often means slightly less curve optimiser margin in my testing, so make sure to rerun with your memory at normal frequency/timings too.
Here's my 7950X3D CO Settings.
Was going to tune them seperately, then put them together and see how they behave.
Sadly just negative 18 for me.
Stable with 12+ hour runs of P95, Linpack and Y Cruncher, plus idled for 48 hours on desktop. Linpack is 12-hour stable at -30, but P95 instantly fails. Maybe because P95 uses all the memory and Linpack only 10gb?
Interesting... I wonder if my 64gb 6000 c30 is what's mandating a marginally higher vcore than others can manage.
I also get a total bluescreen catastrophe with Ryzen Master and have to use bios to tune CO, which is curious and annoying, but probably not an hardware issue? Windows 11 complains about unverifiable files on boot with so much as -1 set by Ryzen Master.
Last edited:
Ryzen Master tool (RMT) is a tool that allows you to change CPU/Memory BIOS options in Windows, albeit with a restart required for changing and applying most things.
It's really best to use one or the other, and me personally I'd much rather use the BIOS. I never use RMT for anything other than showing others my CO settings.
I'd advise everyone else to do the same, as it's quite easy to accidentally change something in RMT without realising, which is possibly what happened to you - is my best guess.
Last edited:
Idk, I tried a few times... literally just open RMT, go to the CO page, set it to -20, reboot: "blue screen" during startup saying something about Windows cannot verify this file and needs to be recovered. Not sure if it was an actual BSOD, or a message designed to look like one. Cleared CMOS, booted fine. Probably only really needed to reset CO in the bios but I hadn't found where it was hiding at the time
Repeat with -10; same outcome. -5; same outcome. Finally found the bios settings, set -20; booted just fine. Didn't pass stability testing, but at least now behaves like I expect an undervolted CPU to behave rather than a total block during the boot sequence.
It's not important, of course, but it is odd to me that RMT makes Windows 11 so very unhappy. But I'll just be leaving it alone now, this new build is so much cooler, quieter and faster than my 8700K that I'd really like to stop messing around tuning it and put it under my desk properly
Last edited:
I hear bad things about using RM for setting curve optimiser.Idk, I tried a few times... literally just open RMT, go to the CO page, set it to -20, reboot: "blue screen" during startup saying something about Windows cannot verify this file and needs to be recovered. Not sure if it was an actual BSOD, or a message designed to look like one. Cleared CMOS, booted fine. Probably only really needed to reset CO in the bios but I hadn't found where it was hiding at the time
Repeat with -10; same outcome. -5; same outcome. Finally found the bios settings, set -20; booted just fine. Didn't pass stability testing, but at least now behaves like I expect an undervolted CPU to behave rather than a total block during the boot sequence.
It's not important, of course, but it is odd to me that RMT makes Windows 11 so very unhappy. But I'll just be leaving it alone now, this new build is so much cooler, quieter and faster than my 8700K that I'd really like to stop messing around tuning it and put it under my desk properly
I’m finding that it is really just a case of testing one at a time fully, then also at idle for a couple of days. It is a long process but it will pay off in the end when you have a fully stable system for you to enjoy. My best 2 core have to be even lower than I first thought it to be stable from -28 to -25.