Sileni

My Vega 56 seems to be getting more WU than my 5700xt. They might prioritize certain architectures, or it might just be down to luck of the draw. You're helping just by contributing capacity, even if they don't saturate it. Just means the WU they do push out get started immediately.

Running a 2600x/5700xt system and haven't been able to get a work unit all morning. I'm getting "No WUs available for this configuration" on my CPU, and the below errors for the GPU. WorkServer connection failed on port 8080 trying 80 A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond. Anyone know if this is a server issue or if I need to fix anything?

They're commonly run in open air cases, which makes them far more likely to be bumped. Along with the fact that he has 3 identical models, and that they're 1060 6gb's, evidence points at mining.

I don't think the 3900x is really necessary; could cut that processor price in half and get a 3600. Could also reduce the cooler to a Hyper 212 or Noctua, probably not a good idea to overclock the 3600 long term past what those coolers can take anyway. Cases are personal preference, but I love my Meshify C, and it's significantly cheaper than what you have listed. An x570 board is only really necessary if you care about the PCIe 4.0. x470 boards are almost half the price. I like the x470 Taichi from ASRock. The 2080s could be reduced to a 2070s, but it would probably be a noticeable drop in performance. Beyond that, it's more pennies you're picking up. You could find cheaper fans (Arctic P12s come to mind), and you could reduce the NVMe size and get a larger hard drive for long term storage.

I've had some fun with a magic square program I wrote recently and would love to see how it runs on other processors. I've ran it on an R5 2600x and an R5 3500u, in both Linux and Windows. It tests every possible permutation of a set of numbers up to a given max to determine if they can form a magic square. It runs in parallel using Rayon with their base settings. Source can be found on my GitHub. It's pretty straightforward, so I'm not posting a compiled version yet. Let me know if you'd like to see one. Linux is significantly faster, probably because of how simple the inner function is. Currently don't have Linux results because I'm dual booted and saved my Linux results to an encrypted drive. Will post if there's interest. Results for a 3x3 square on the 2600x - Windows in the attached file. Compiled using cargo build --release Easiest way to run the program, and the way I ran the program for the results below, is ./msquare -ic //Linux %CD%/msquare -ic //Windows Then press Ctrl+C when you'd like to stop the run. Logging (the -c flag) adds a bit of time for obvious reasons. Mostly curious about this for comparison purposes, and because I'm learning Rust. Thanks! test.csv