-
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.
**AMD 5700 X3D Owners CLub**
- Thread starter something daft already!!
- Start date
Associate
Yes UEFI version used needs to support 5700X3D. Motherboard support page will have information from which UEFI version the 5700X3D is supported.
AGESA is one piece of the jigsaw. The UEFI can contain various other updates which maybe needed. Sometimes the UEFI contains "Auto rule" changes, which might apply to the new HW, which is from the motherboard makers side not AMD AGESA. You've got AMD firmware and motherboard manufacturer firmware rolled into one package.
Hello again,
Just one more little question since never actually done AMD before.
If Im dropping in new CPU, and I have never updated previous CPU/bios.
Do I update old CPU/bios first , then drop in new CPU then update to latest bios ?
Just one more little question since never actually done AMD before.
If Im dropping in new CPU, and I have never updated previous CPU/bios.
Do I update old CPU/bios first , then drop in new CPU then update to latest bios ?
Last edited:
Update the bios first with the old CPU still in.
Then after that you can put the new CPU in and it should work.
Associate
No problem
. Yeah some manufacturers slower than others, but all seem to be updating as and when needed.Enjoy your new HW
.You can do as xiphrex states, but the Crosshair VII Hero also supports flashback.
You don't even need any HW on motherboard other than power supply and can flash latest BIOS.
Personally I'd prep USB with latest UEFI, turn system off, swap CPU, press clear CMOS button on rear IO, flash UEFI using flashback.
ASUS flashback guide link.
Last edited:
Soldato
The AGESA for 5700X3D is newerer than the 5800X3D, I have MSI and it was quite clear which BIOS they patched it in - which was a while back AFAIK. I just stuck the latest non Beta on.
Soldato
Sorry I’m sure it’s already been asked a million times but literally just a case of updating bios, plop new cpu in, load into windows and make sure all is as it should? Then go back to bios set CO to -20 (or 30) enable xmp on ram and load back in and do a bit of stress testing?
Associate
Pretty much - that's all I did anyway. As a few others have mentioned, it took my rig a couple of false starts to boot up cleanly but once it did - no issues. I still haven't enabled xmp but have set to -30 in the bios and all good.
Thats what I did.
I enabled XMP and tested that first, then went back and put CO on -30. Did a bit of stress testing, it worked fine. Have had no issues since.
Feel like if overclocking was unlocked on this that I could get it to boost higher.
I am a tinkerer and always get the latest BIOS apart from Betas, have been burnt by those before. My board just took the 5700x3d right away, no issues. I remember how annoying it was when Windows 11 came out and I kept getting annoying and seemingly random crashes until they released a new BIOS update. Rock solid since. Just did a cable mod for it and now it looks even better, to me. My partner could not tell, but she is only a woman after all, bless her.
Soldato
Well that was as straight forward as it could have been. Plonked it in, couple of clicks on the bios to set CO and it’s all trundling along nicely by the looks of things.
I actually hadnt updated my bios since since 2021 because of the random issues that each new AGESA or whatever update brought. So since my PC was running okay I didnt take the risk.
But now Ive updated I regret not doing it sooner - PC seems more stable, cold boot up issues have resolved as well.
Associate
Post 192 had the data from CO tweaking of my kit. Post 194 had info -30 all cores had been unstable.
Set CO profile per core, which at wall plug power meter ~17W lower power usage vs stock, CB23 multicore test load. Then this CO profile passed ~3 days uptime of mixed testing.
Next I checked profile for POST to POST stability variance and resuming from sleep, bumping up VDDG CCD & IOD resolved stability issue.
To round up my testing of CO profiles I rerun some tests, testing ZIP link, organise by time to see order of testing.
Some graphs:-
Stock vs -10 all cores vs -20 all cores, link.
Stock vs -30 all cores vs CO per core, link.
Stock vs -30 all cores vs CO per core +4, link.
Link to Generic Log Viewer, used to create graphs, great app IMO.
Even though CO -30 all cores results in maximum lowering of power/temperature, stability may not be ideal. Per core CO tuning can give similar results, with enhanced stability.
Security updates are also delivered via FW, personally I always go for latest drivers/FW.
Set CO profile per core, which at wall plug power meter ~17W lower power usage vs stock, CB23 multicore test load. Then this CO profile passed ~3 days uptime of mixed testing.
Next I checked profile for POST to POST stability variance and resuming from sleep, bumping up VDDG CCD & IOD resolved stability issue.
To round up my testing of CO profiles I rerun some tests, testing ZIP link, organise by time to see order of testing.
| Stock | ~100.5W |
| -10 all cores | ~92.9W |
| -20 all cores | ~86W |
| -30 all cores | ~78.5W |
| CO Profile per core | |
| -21 -21 -23 -24 -23 -23 -24 -24 | ~83.3W |
| -22 -22 -24 -25 -24 -24 -25 -25 | ~82.4W |
| -23 -23 -25 -26 -25 -25 -26 -26 | ~82.1W |
| -24 -24 -26 -27 -26 -26 -27 -27 | ~81.4W |
| -25 -25 -27 -28 -27 -27 -28 -28 | ~81.1W |
Some graphs:-
Stock vs -10 all cores vs -20 all cores, link.
Stock vs -30 all cores vs CO per core, link.
Stock vs -30 all cores vs CO per core +4, link.
Link to Generic Log Viewer, used to create graphs, great app IMO.
Even though CO -30 all cores results in maximum lowering of power/temperature, stability may not be ideal. Per core CO tuning can give similar results, with enhanced stability.
Security updates are also delivered via FW, personally I always go for latest drivers/FW.
Last edited:
Interesting.
So the -30 OC all cores was stable in stress testing except with Kahru RAM test.
I miss the days when you could run Toast or Prime95 and if it passed a few hours or overnight you knew it was stable. Then came Ryzen and it could pass one stress test but then bluescreen/reboot on idle, or fail another stress test.
Is there a consensus on which stress tester is more likely to pick up on instability the quickest?
Associate
The -30 all cores passed Y-Cruncher 20min (all tests), 9hrs CoreCycler P95, turned system off as was writing post 192 on another system. Resumed testing on new POST of system, passed 1hr RealBench stress test, then I had a fail in Kahru RAM Test, post 194 had link to screenie. I could only assume the combination of -30 CO and RAM profile wasn't working together for each POST of system.
CPU's boost behaviour and number of cores was so different when you could run few tests and know you have high stability. CPU boost behaviour is based on more factors now, application loading CPU, instruction set in use, number of cores in use, temperature, voltage, power, etc. I can't wait to get a 9800X3D, see the TPU review for clocks on different test loads.
I can't say I have found one load test that does it all. Resuming from sleep and testing is always the quickest way to see instability. Motherboards POST differently if power has been switched off to them, if standby power was there, if you do a restart and resuming for sleep. In the past I have sat and watched Q-Codes displays to see POST differences.
Last edited:
I had so many issues trying to get my RAM timings stable after updating my UEFI once on my X470 Taichi Ultimate that I stopped doing it. I only did it once after that, to add support for my 5800X3D. I didn't go as tight with the timings since installing that chip because of lack of time coupled with it not mattering as much with the X3D chips. Now it's been stable the whole time I won't update the UEFI again unless absolutely necessary!
Associate
I can't say my RAM profiles and or CO profiles have failed between different UEFI versions, that was 5900X/5800X3D. Only used one UEFI with 5700X3D, doubt I'll have a differing experience.
TBH the Crucial Ballistix Sport LT I've had a lengthy while, even when I swapped boards I didn't have to slacken timings, 2022 link, 2020 link.
TBH the Crucial Ballistix Sport LT I've had a lengthy while, even when I swapped boards I didn't have to slacken timings, 2022 link, 2020 link.
Last edited: