Hi people, I might be wrong but I think I saw once a Guide to OC the Ryzen 3900 from Wendell, but I can’t find it now, can someone provide me a link of it?..I don’t mean a vid in youtube, it was more like written guide, or a Thread here in the Forum…but like I’ve just said, I might have dream it and never existed…thanxs in advance!
Yeah a 3900X needs higher voltage to contain a certain clockspeed.
This is also one of the reasons i think they could run hotter.
1 Like
Completely forgot to ask here, but has anyone tested X570 boards with a PCIe Riser Cable connected to a Navi Card? I had some weird issues with my sapphire pulse 5700xt a while ago and dont know if it was the Gen3 cable I was using of if it was the card that had a fault in it.
Can’t say I have. I don’t even have PCI-E 4.0. I’m running x370!
Man i just had some funky behavior on my Aorus X570 PRO. Worked perfectly and rock solid since i dialed it in months ago and suddenly today no post, fans blazing full speed etc. Ram led staying lit.
Clear cmos does not work except now the CPU led is lit, reseating every ram stick does not work. Fine try one module… hey it works again. Ok one module died it happens, i insert the second module for dual channel. It works again so i replace the second module with the two other ones i removed and it all works fine.
So for shits n giggles i insert all modules and it posts again, i reload the profile i’ve been using and it’s all good.
Now i did run it at 3733 synced with mem voltage at 1.42V, SOC is at 1.1V, CPU is on normal/auto, VDDG 950mv and VDDP 900mv. But these are well within safety limits i’d reckon, hell they’re stock. Precision boost is on auto etc.
Either the motherboard is acting up or somehow the memory controller on the 3800X is messed up or it’s all just a fluke. Memory is Crucial E-die 4x8GB btw.
@Riotvan
I was reading a thread over at OCN , a couple cats reporting similar (not exact) experience also the recovery method, too. Close enough to make me think hmmm…
I’m hoping that it’s not a “thing”… I have four 3000 series chips all mild OC at least all of them with 1900 1:1:1 but pretty standard volts & powers otherwise.
Here’s the post
1 Like
Thanks. I guess we will have to keep an eye on it, after this my system is back to being rock solid again at the previous settings. It might just be a bug that can get triggered requiring a reset of some sort, maybe on the cpu management engine or something. Since clearing cmos did nothing.
There are some threads on Reddit about a bug(?) with AOC/PBO.
If you set EDC to 1 and turn off P-States - you get higher boosts…
Not sure how safe it is regarding vcore etc, temps went up a bit.
But I had to try at least - Woohoo Boost!
1 Like
More info here https://www.overclock.net/#/topics/1741052
Edit: ping @wendell what’s your take on this? (Does @ work like this here?)
edit2: working link https://www.overclock.net/forum/13-amd-general/1741052-edc-1-pbo-turbo-boost.html
1 Like
Neat. Looking into it.
Well, I thought my FCLK issues were because of silicon lottery… Turns out it was the stupid C6 C-state issue I got bit by.
And guess what? ASUS REMOVED the “Power Supply Current Idle” feature in AMD CBS in the newer BIOSes.
DO NOT buy an ASUS board for Linux Ryzen. The solution using Power Supply Current Idle cannot be done on ASUS boards.
2 Likes
If you’re right, it’s maybe do not buy an ASUS board for overclocking. My two ASUS Ryzen boards are running Linux and working great.
But the most exciting thing I’ve tried is going to 3,600 MHz on the RAM for the 3900X, and letting the board do an automatic all-core 3.7 GHz boost on the 1700X, which is honestly nothing exciting.
I’ve been having non-stop C6 idle state issues. I thought it was FCLK but it’s not. It only happens when the system is idle.
The Power Supply Current Idle is a workaround for the issue even at stock speeds, but ASUS removed the option. IT IS NOT AN OVERCLOCKING OPTION.
2 Likes
Well i can remember that there actually were issues,
with C6 powerstates and certain psu´s,
that could not really handle that very well.
But i think that was like years ago, and shouldn’t be,
an issue with modern psu’s anymore.
It’s not a PSU issue, it only happens in Linux with the up to date AGESA, and it doesn’t happen with the same AGESA in Windows. The tweaks for “Ryzen Balanced” are WAY too Windows specific. Linux users are being ignored for the “optimized” power plans.
1 Like
Yes i understand that it is not a psu issue.
But when it comes to the powerplans on linux,
isn ´t it possible that it might being distribution specific as well?
Unfortunately i don´t have Ryzen stuff to test around with.
But i cannot remember that i have read allot of complaints about this issue.
And not sure if wendell has ever experienced or adressed it?
It’s in the kernel bugtracker with no resolution in sight:
1 Like
Hmm interesting…
I’m running a Linux NAS system with a 1700X on a ASUS x370-Pro board with a pretty default Fedora 31 installation. And two other people who took my recommendation are doing the same thing. And none of us have that problem. I don’t know what distro the others are running, but one is Ubuntu I think.
And this system I am typing on right now is a 3900X with an ASUS x570. Uptime is over 7 days.
Anyway, my point is that while you may see this problem a lot, it is hardly universal because I’ve never seen it before.
1 Like
The random nature of it is what makes this so frustrating. If you look at the bugzilla tracker, nothing concrete has been established and it’s all guesses with no AMD engineers commenting.
This quote marks how grim the situation is:
I have mailed this issue and the link to this thread to many hardware reviews sites to get attention, and to AMD, and not a single one even bothered to reply let alone have a look at it.
If they even don’t want to recognize this as a bug, why fix it?
1 Like