neurotix

It came with the numpad board (was Yunzii something from Amazon)

Got the CK-721 set up next to the wife's macaron numpad. She'll be using it for work all next week. Thanks Cooler Master and Avacado!

In bigred On the wife's 3060.

Don't have a picture, but I got a tiny little GT 1030 DDR4 as a temporary card until I get my 4090 back. This thing has no idea what's waiting for it, as I am going to oc it and run Fire Strike Ultra and Time Spy Extreme on it. Anyone remember good idea, bad idea from Animaniacs and Mr. Skullhead? Not sure if I should do it or not but I'm bored usually and not into games atm and the thing might get me cups/hwboints, especially with my CPU, so..

@damric Can you please add my prizes to the prize list? We have a lot of newcomers, maybe they'd want my fans and cable. Additionally I can provide up to 8 of these Corsair stock radiator fans:

I was so close to passing you up, too! Lol

Bad news: Yesterday night, my rig started not waking up from monitor sleep while folding. My monitor would just say no signal. I tried with folding not running, same issue. It is freezing in games in Windows. I'm afraid folding has killed my beloved 4090 Gaming X Trio, but it's done 2.3 billion points for the team. We've tried a lot of things and the issue persists. This is bizarre as I once folded on an OC'ed HD 6870 for a year straight 24/7 and got the badge for it on OCN. I've folded for months at a time on a ton of other cards and never had them die, but this really is probably the only reason/explanation for it. Just realized there's a rally expansion for Forza Horizon 5 and was planning on playing it this week, I guess that's not happening. Anyway TL;DR: folding has appeared to kill my 4090, am getting a GT730 tomorrow to use temporarily for graphic design, MSI approved my RMA and they should have my card on Friday, we shipped it. From now on, will not be folding on my 4090 if they replace it, its just a waiting game now (2 weeks after they receive the card).

Nah add them to the prize list please. I'm not setting up a bench rig, I'd need to buy a psu, the spare coolers I have can't be mounted on AM3+ (my spare FX-8350 setup), plus I'd need to buy old GPUs. We've spent a lot from savings recently but our goal is to have $25k by 2025 to go to Japan to the 2025 Osaka World's Fair (look it up, it's really interesting) Edit: also my card was approved for RMA by MSI so that's good news. Double edit: sorry for hijacking the thread with my dead GPU woes. Carry on.

I am contributing 3 prizes for this foldathon if anyone wants them. Delta high speed fan, giant Yate Loon 230mm fan, and an approximately 30ft Cat 6e 10Gbps flat Ethernet cable. @damric

Maybe it was the OC. Though there's no way to overvolt the card in Linux and it was only at 2970MHz.

Thanks pio. Yeah, maybe it has the same problem and got really hot and fried with these weird new 16xxx units I've gotten the last 3 days.

Yes, I had it clocked to 2970MHz core, anything higher than 3000 caused Linux to lock up. Though the card can do 3135/12210 in Superposition. Yeah I have the thick 3 slot cooler and triple fan MSI cooler. The thing that is though, even overclocked, with 80% fan, the card never passed 52c, except when we had 18601 units back in January, those would cause it to heat up to 68c. I can tell you guys this much: if I get MSI to replace or repair my 4090, or if not and we have to buy a new one, I am never ever running folding@home on it ever again. I'm steaming.

OUT bigred. Remove me from the foldathon please. Yesterday night, my rig started not waking up from monitor sleep while folding. My monitor would just say no signal. I tried with folding not running, same issue. It is freezing in games in Windows. I'm afraid folding has killed my beloved 4090 Gaming X Trio, but it's done 2.3 billion points for the team. We've tried a lot of things and the issue persists. This is bizarre as I once folded on an OC'ed HD 6870 for a year straight 24/7 and got the badge for it on OCN. I've folded for months at a time on a ton of other cards and never had them die, but this really is probably the only reason/explanation for it. So I will be buying a GT 710 at Best Buy later, just so I have video and my system is usable as a server, as well as for graphic design. Just realized there's a rally expansion for Forza Horizon 5 and was planning on playing it this week, I guess that's not happening. Anyway TL;DR: folding has appeared to kill my 4090, will get a new temporary card and starting the RMA process with MSI.

Ignore

So I found a bunch of different power plans. One was a custom one from OCN that someone made and it included like 5 but none of them helped. However, I found this: GitHub - neon-dev/Ryzen-Optimized-Power-Plan: Further optimized version of 1usmus' great power plan (https://www.techpowerup.com/review/1usmus-power-plan-for-amd-ryzen-new-developments/) GITHUB.COM Further optimized version of 1usmus' great power plan... @Sir Beregond , you might want to try this out. With this one on, my cores of my 2nd CCD were boosting to 5425MHz. And that was without a PBO offset of +200. I am going to test it more during this week, but I did this: /img/logo.png neurotix`s 3DMark Vantage - Extreme score: 132362 marks with a GeForce RTX 4090 HWBOT.ORG The GeForce RTX 4090 @ 3105/1519MHzscores getScoreFormatted in the 3DMark Vantage - Extreme benchmark. neurotixranks #47... Unfortunately, the 2nd CCD cores still aren't boosting past 5000MHz in 3dmark11 Performance. I doubt they would in older benchmarks like 3dmark06 or 3dmark05. I believe the low resolution is the problem, the AMD 3d V-Cache service or the Windows scheduler is deciding because it's so low resource intensive now that it doesn't need to boost clocks. 3dmark Vantage Extreme runs at 1920x1200, which is where I was seeing boosts of 5425MHz across like 6 cores at once on the 2nd CCD.

Yeah I know, I even uninstalled the 3D V-Cache services and tried and it still didn't boost the 2nd CCD high. I really loved this setup and the ability to toggle the V-Cache cores for gaming (Forza Horizon 5 benefits greatly) and then toggle to the 2nd CCD and see boosts of 5.5Ghz, now its like its stuck in Balanced power plan or something. This was working that way when I still had the Ryzen Balanced and Ryzen High Performance power plans on the same Win10 install that had my 3900x, which the chipset driver installed a long time ago. I've tried looking for an installer that would reinstall those power plans and found something on Google Drive but link was dead. Enabled the Ultimate Performance power plan and that didn't work either. Id you manage to dig up a link or shell script that can add the Ryzen power plans back, let me know I think I'll have to wait for a new bios/AGESA and possibly newer chipset drivers to fix it. I might contact AMD support and Asus over this issue. So frustrating, I did 70k in 3dm11 performance and 134k in Vantage Extreme.

Awesome. About time.

Yeah I was getting really good scores in 3dmark11 performance, 3dmark Vantage Extreme (both give globals) and was getting decent scores in 06 and 05 as well. Now for some reason, probably the AMD 3D V-Cache optimizer service that gets installed by the chipset driver, my second CCD cores won't boost like they were. Thanks Luke.

Picture of my rig with my GMMK instead of my E-Yooso Z-88. Also benched the processor. /img/logo.png neurotix`s y-cruncher - Pi-1b score: 19sec 196ms with a Ryzen 9 7900X3D HWBOT.ORG The Ryzen 9 7900X3D @ 4921.3MHzscores getScoreFormatted in the y-cruncher - Pi-1b benchmark... /img/logo.png neurotix`s GPUPI for CPU - 1B score: 59sec 350ms with a Ryzen 9 7900X3D HWBOT.ORG The Ryzen 9 7900X3D @ 4999.8MHzscores getScoreFormatted in the GPUPI for CPU - 1B benchmark... /img/logo.png neurotix`s HWBOT x265 Benchmark - 4k score: 33.7 fps with a Ryzen 9 7900X3D HWBOT.ORG The Ryzen 9 7900X3D @ 4905MHzscores getScoreFormatted in the HWBOT x265 Benchmark - 4k... Unfortunately, since last week I have encountered a problem with my 7900X3D, for some reason the second CCD won't boost past 5100MHz or so on High Performance power plan, whereas last week it was sitting at 5325 all core minimum and hitting 5525 then going back down etc. Tried everything including reinstalling win10 and all my programs, nothing has worked and I don't know why. The 2nd CCD boosting high is the difference between 60k 3dm11 performance, whereas last week before this started I was doing 70k which is actually worth submitting. I think I will probably have to wait for a fix from Asus/AMD but basically Balanced power plan would park cores in CCD2 and focus on boosting CCD1 with the V-Cache. Changing it to high performance power plan clocked down the first CCD with V-Cache, and I'd see very high boosting on CCD2. For some reason now in high performance power plan my first CCD sits at 4950MHz and the 2nd one sits with all cores at 5100MHz and they don't boost past that. Any help appreciated. @damric

Gl bro. Work hard. At least you can do it when you aren't busy. My wife basically works on my pc now, I really can't because of pain. I love her so much. Been nagging her to join the site so its not just a white guy sausage fest.

In bigred

Thanks so much. I'll try it tomorrow.

Ok, so I finally found an answer to this on Reddit. I had wanted to lower CPU_VDDIO by a bit to a lot to try to reduce my temperatures. Basically, it's supposed to be 1.35-1.4v. However, I tried giving it 1.35v and my board was overvolting it to 1.375v. I was able to boot into Windows and pass memtestCL built into Ryzen Dram Calculator to 2400% coverage (100% for each thread). I did not run stressapptest in Linux however, because upon rebooting right after memory training I was getting POST code of 46 and my system failed to boot. I also noticed that on auto, the board is giving 1.456v to that voltage, and my memory voltage is 1.435v so it's giving nearly the same overvolt to CPU_VDDIO based on my memory voltage. I will probably try setting it to 1.4v tomorrow and run through some tests, or possibly try 1.375 cuz the board will overvolt that to 1.4v and see if that works. The other DRAM power delivery stuff I asked about, I have not touched and still need an answer to, but I figure my memory is running very well already so why increase heat. I might play with those settings and see if I get lower ns or something in AIDA64. However, with the crappy cooler I'm forced to use and the fact that I fold on it in Linux all the time and the system is usually saying its 64-68c just from the heat of my GPU rising and heating the memory, I will probably leave the dram power delivery settings on auto most of the time and maybe only increase them for benching. So there's some answer. I'd be curious to know what Nikado7's Taichi board is giving this voltage as well as Fluxmaven's Crosshair Gene.