I compiled gcc on the pi5 at 2900, no over-voltage_delta, using jahboater's old gcc_compile script (it still builds gcc-10.2.0) see here, with a few added timing lines of my own:Rounding some seconds, I'll call that 1 hour 9 minutes. The fastest I had run this before, same pi5, same nvme drive, was 1 hour and 21 or 22 minutes.
So, now, having gone back and shown that 2900 is solid using longer tests and other stressors, how do I attempt to climb higher again? "3000 no over_voltage_delta" was stable under pi-under-pressure but not under the lorenz test. See the last test results here and here. I think I'll see whether there's some over_voltage_delta that makes 3000 stable under lorenz. If not, I'll see what I can do with 2950.
Code:
pi@pi5:~/Documents/temptests $ grep CST pi5_2900.out Getting source at Fri Jan 9 12:23:37 CST 2026 ...Getting prerequisites at Fri Jan 9 12:24:03 CST 2026 ...Configuring at Fri Jan 9 12:24:06 CST 2026 ...Making at Fri Jan 9 12:24:08 CST 2026 ...Done at Fri Jan 9 13:32:53 CST 2026 ...So, now, having gone back and shown that 2900 is solid using longer tests and other stressors, how do I attempt to climb higher again? "3000 no over_voltage_delta" was stable under pi-under-pressure but not under the lorenz test. See the last test results here and here. I think I'll see whether there's some over_voltage_delta that makes 3000 stable under lorenz. If not, I'll see what I can do with 2950.
Statistics: Posted by tinker2much — Fri Jan 09, 2026 8:38 pm