My baseline results at 1500 MHz are reported inCurrently that is just numbers.
Use any benchmark to compare 3400 with default frequency, or other PIs.
Like pichart from this thread:
https://forums.raspberrypi.com/viewtopic.php?t=227177
https://fractal.math.unr.edu/~ejolson/p ... urrent.tgz
https://fractal.math.unr.edu/~ejolson/p ... irates.svg
https://forums.raspberrypi.com/viewtopi ... 1#p1699381
Since your baseline results will depend on the C compiler and whether 32-bit or 64-bit, I'd suggest running your own Pi chart tests at 1500, 1800, 2100 and then 3400 to see the differences. Note that RAM speed is a constraining factor for some of the tests but the prime sieve usually scales linearly with CPU clocks.
A scaling analysis depicting computation speed versus CPU clock speed is at
https://forums.raspberrypi.com/viewtopi ... 1#p1993511
Note the Lorenz 96 computation decreases in performance as clocks go beyond 1600 MHz. I've always thought that a little weird.
Statistics: Posted by ejolson — Sat Jun 22, 2024 6:06 am