I've tried compiling the standard SETI source in order to optimize it for my architecture, but instead of being faster as I'd assumed it would be, the compiled app is several times slower. I've enabled all the CPU instructions in ./configure and edited the Makefiles so that only -march=skylake-avx512 appears in the flags; what could be the problem? I assume the app that BOINC automatically downloads is -march=x86_64 and compiled from the same source that I'm using. I don't understand why my binary is so slow. It's been suggested to me that I compile a modified (optimized) SETI, but I can't get that source to compile on my system. I feel that I shouldn't need to use a modified version considering that the BOINC automatic version is presumably compiled from the standard source, although I guess that could be an invalid assumption. Any advice would be greatly appreciated.

We don't have any official Linux developers for the science applications. Most of our apps are very old from 2016. Only TBar has done any recent compilations of the Linux science apps and he has found that compiling on any distribution newer than 15.04 troublesome.

Richard always makes a disclaimer that he is a Windows guy and says he doesn't know much about Linux. Inferred is the message that we Linux users have to fend for ourselves.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Thanks for the info.

Looks like RueiKe has had some success finally with compiling over in the I've Built a Couple OSX CUDA Apps... thread.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)