tictoc

I did remove the beta flag for now on the 980, but since I am fairly active testing beta projects on the beta team, I try to keep at least one GPU running beta tasks. The error is a CUDA kernel error, but it is fairly random and after a restart I can complete the task that had errored out.

980 crashed again. I'm not sure if it's a hardware issue or an issue with the WUs. The crash has only happened on a specific set of beta WUs. Back up and running now. I have reported the error, so we'll see if anyone else has hit the same problem.

I am running F@H on Arch. Right now I have a 980 and a 2080S folding on a headless machine, running a minimal Arch install. The current AUR package has a few quirks to tighten up security, but it does work without issue. SNAPs will never see the light of day on any of my machines, so I've always ran the package from the AUR. I've had at least one machine folding on Arch for at least the last five years. I actually put together a little guide for one of my teammates on how to setup a minimal headless Arch install for folding. You can find that guide in a bit of an odd location: https://forums.extremehw.net/topic/1574-unable-to-add-attachment-to-pm-or-post-and-kicked-out-of-editor/?do=findComment&comment=18925

My 980 crashed at some point in time over the weekend. I didn't even realize it was down, since it's in the same machine as the 2080S, which just kept on folding. I backed the clocks down a bit, and it's back up and running now.

Evil Linux h4x0|2 Thanks for looking into it E, and archiving similar files is an easy workaround.

Here is a screenshot of the error when attaching the file.

I have been tempted to upgrade. I have a machine running one of the first batch Ryzen 7 1700s with the segfault bug. I never RMA'd, so from time to time I hit the bug. That machine mostly sits on my test bench. The bug doesn't come up too often, but every time it crashes, I have to determine if it's a real crash or the faulty CPU. Too bad I also need a new motherboard to run a 5800x, which is $370 at Microcenter right now.

Sorry I forgot to report back. I did try it again, but the result was more or less the same. I will give it another test in this post. *Test 1 with Firefox* With the text pasted into a code block in a spoiler, I was sent to a captcha page, completed the captcha, and then it just kicked me back without submitting the reply. All of the content was still in the editor. *Test 2 with Firefox* Attaching the .txt file to this post resulted in the same -200 error *Test 3 with Chromium* Attaching the file failed with the -200 error. Pasting the text into a code block did work after completing a captcha.

In Q_Cruncher

Testing out a new home router setup. My first custom built router was running Debian Etch, and now we're completing the circle with VyOS (which is built on a Debian base).

Power went out at my house about 4 hours ago. No eta on when it will be back up. Unfotunately, all of my systems are not on circuits that are hooked up to the generator, so I'll have some downtime until power is back.

I certainly can. It won't be until later this afternoon, but I'll give it a go and let you know the results.

--Update-- I also cannot add the guide to a post. Blocked with the same screen, but upon returning to the thread the content is still in the editor. While attempting to attach a .txt file in a PM, the attacment errored out with a -200 error code forcing me to close the message and then open it again to get rid of the failed attachment. This worked without issue a few days ago. After the attachment failed I than pasted the text into the editor, and after spending quite some time reformatting, I was greeted with the below screen after hitting send. Next I fired up a different browser, tried pasting the text into a code block, but I got hit with the same screen, just with a different Cloud Flare Ray ID: 69c425165eddfeaa OS: Arch Linux Browsers: Chromium and Firefox Extensions: Disabled while trying to send the PM

Heads up to anyone folding on AMD GPUs and running beta tasks. There are some very low ppd WUs that started going out yesterday. -Edit- The same tasks are also on the low end for NVIDIA, but just lower than average, not the -50% ppd that I was seeing on my Radeon VII.

The reports I read seem to support this idea. Memory usage goes up with every window that is opened, but then when the windows are closed, Windows fails to reclaim all the memory.

Outside of this issue, the overall performance is pretty poor on Threadripper as well. Windows still can't seem to sort out their scheduler for high core count CPUs. That L3 cache issue is pretty terrible, especially since one of the reasons for the improved performance on Zen 3 is the whack ton of L3 cache that it has.

Nice. I bet there is still some headroom to push it even a bit higher. I would personally be comfortable pushing 100W on the 6-pin connector, since I doubt you're running a cheapo PSU with 20 gauge wires. The GPU should certainly have no problem handling a bit more power with half way decent cooling.

That's definitely a bug, and a true memory leak. How did a bug like that ever make it through QC and into the release version? Bugs happen, but this seems like a pretty obvious one. I'll have to fire up my Win 11 VM, and see if it's the same for one of the earlier betas.

Inadvertent double. Accidentally quoted myself on my phone when I meant edit.

It has been quite some time since I modded a Keplar BIOS. You will need to edit the Def table value. I'll have to grab a copy of the BIOS to poke at it to try and figure out which table is which. For starters the second table looks like PCIe power, so bump the PCIe rail Def up to 75000, flash and then see what the results are. Looking at your first post, and the fact that you are only hitting 70% of TDP, I suspect the third table is the Power Target. I would slowly bump up the Def value there, and test to see what the results are. -EDIT- Def is just short for Default

Is this even a bug? explorer.exe is heavily integrated into everything, and from a non-Windows user, this just looks like typical OS caching. If things are never evicted from memory it would be a problem, but unused memory is just wasted memory.

I can poke around on this later this week and this weekend. Is the KBT screenshot your modded BIOS?

I'll shoot you the log for that machine once the stats site is back up. The only thing running F@H on that machine is the Radeon VII, so it's easy enough to track any missed units.

I will be temping in the AMD slot with a Radeon VII. It's not a great card for the ETF. The average ppd is quite a bit lower than what it used to be, but it was the easiest thing for me to get up and running in a short amount of time. I do have the OC pretty much as high as you can possibly go outside of sub-ambient cooling. Right now it is clocked at 2125core and 1225mem, and running on modified Power Play Tables to get those clocks stable.

The K5000 is the same GPU as the GTX 680, but it runs a much lower stock clock to fit in a smaller TDP budget (120W). Upping the power limit and giving it proper cooling allows for pretty nuts OCs that are in the range of 40% over stock.