There used to be a thread on the old MicroPython forum listing pystone_lowmem results: Benchmark comparison of MicroPython boards. Sadly, that little benchmark got kicked out of micropyhin-lib a couple of years ago
OP: I'm not sure what you're driving at. You're returning ~0.00000265 pystones/s, while the test thread returned 1086 pystones/second for the original Pico.
There's a perfbench tool is the MicroPython source repo: micropython/tests at master · micropython/micropython. Its output is rather opaque.
On an original Pico:On the fastest MicroPython board I own (Adafruit Metro M7 with 500 MHz MIMXRT1011DAE5A):I'd expect the Pico 2 to come somewhere between the Pico's 1417 pystone/s and the Metro M7's 7660 pystone/s.
(I found the old pystone_lowmem.py: it returns 1400 for the Pico and 6756 for the Metro M7. My name's not Roy, I have no idea what I'm doing ...)
OP: I'm not sure what you're driving at. You're returning ~0.00000265 pystones/s, while the test thread returned 1086 pystones/second for the original Pico.
There's a perfbench tool is the MicroPython source repo: micropython/tests at master · micropython/micropython. Its output is rather opaque.
On an original Pico:
Code:
$ ./run-perfbench.py 125 128 -p -d /dev/ttyACM1N=125 M=128 n_average=8perf_bench/bm_chaos.py: 223736.00 0.0303 223.48 0.0303perf_bench/bm_fannkuch.py: 106419.88 0.0202 56.38 0.0202perf_bench/bm_fft.py: 501973.12 0.0048 1529.96 0.0048perf_bench/bm_float.py: 102000.12 0.0399 2450.98 0.0399perf_bench/bm_hexiom.py: 283409.50 0.0202 35.28 0.0202perf_bench/bm_nqueens.py: 433759.38 0.0359 2305.43 0.0359perf_bench/bm_pidigits.py: 136814.38 0.0077 475.10 0.0077perf_bench/bm_wordcount.py: 251294.50 0.0264 31.84 0.0264perf_bench/core_import_mpy_multi.py: 151657.12 0.0660 329.69 0.0660perf_bench/core_import_mpy_single.py: 15518.00 0.1482 64.44 0.1484perf_bench/core_locals.py: 312529.00 0.0160 25.60 0.0160perf_bench/core_qstr.py: 29166.12 0.0262 137.15 0.0262perf_bench/core_str.py: 322572.12 0.0116 18.60 0.0116perf_bench/core_yield_from.py: 109750.75 0.0965 182.23 0.0963perf_bench/misc_aes.py: 103170.00 0.0077 310.17 0.0077perf_bench/misc_mandel.py: 231296.00 0.0190 1729.39 0.0190perf_bench/misc_pystone.py: 211626.88 0.0571 1417.59 0.0571perf_bench/misc_raytrace.py: 206510.75 0.0195 242.12 0.0195perf_bench/viper_call0.py: 90606.88 0.0020 331.10 0.0020perf_bench/viper_call1a.py: 92765.12 0.0018 323.40 0.0018perf_bench/viper_call1b.py: 124459.00 0.0018 241.04 0.0018perf_bench/viper_call1c.py: 123499.88 0.0028 242.92 0.0028perf_bench/viper_call2a.py: 94204.25 0.0015 318.46 0.0015perf_bench/viper_call2b.py: 141979.25 0.0014 211.30 0.0014Code:
$ ./run-perfbench.py 500 96 -p -d /dev/ttyACM1N=500 M=96 n_average=8perf_bench/bm_chaos.py: 54223.00 0.2330 922.12 0.2333perf_bench/bm_fannkuch.py: 241671.38 0.0983 28.96 0.0984perf_bench/bm_fft.py: 23270.12 0.1211 11001.25 0.1211perf_bench/bm_float.py: 9224.12 0.3993 16261.96 0.3987perf_bench/bm_hexiom.py: SKIP: no matching paramsperf_bench/bm_nqueens.py: 103902.12 0.1800 9624.47 0.1800perf_bench/bm_pidigits.py: 13435.38 0.4768 2605.12 0.4782perf_bench/bm_wordcount.py: 793602.12 0.1370 50.40 0.1366perf_bench/core_import_mpy_multi.py: 59473.62 0.3790 840.72 0.3799perf_bench/core_import_mpy_single.py: 4994.50 0.9130 200.24 0.9154perf_bench/core_locals.py: 929221.38 0.0068 43.05 0.0068perf_bench/core_qstr.py: 4282.12 0.4676 934.14 0.4673perf_bench/core_str.py: 350655.25 0.0999 85.55 0.0999perf_bench/core_yield_from.py: 13702.25 0.0370 1459.61 0.0370perf_bench/misc_aes.py: 9767.00 0.7617 1638.26 0.7623perf_bench/misc_mandel.py: SKIP: no matching paramsperf_bench/misc_pystone.py: 39160.50 0.2149 7660.82 0.2154perf_bench/misc_raytrace.py: SKIP: no matching paramsperf_bench/viper_call0.py: 21214.62 0.5918 1414.17 0.5886perf_bench/viper_call1a.py: 25235.38 1.2381 1188.99 1.2427perf_bench/viper_call1b.py: 43393.62 3.2312 692.04 3.1216perf_bench/viper_call1c.py: 42374.75 2.1171 708.30 2.2236perf_bench/viper_call2a.py: 27288.12 1.2628 1099.56 1.2859perf_bench/viper_call2b.py: 21816.62 2.0326 1375.64 1.9444(I found the old pystone_lowmem.py: it returns 1400 for the Pico and 6756 for the Metro M7. My name's not Roy, I have no idea what I'm doing ...)
Statistics: Posted by scruss — Fri Oct 11, 2024 9:24 pm