fix: ensure moffat derivs are not nan and increase performance for `trunc=0`
#188Comparing
debug-moffat-perf
(ecba9da
) withmain
(51df286
)×3.7
- Improvement
- 4
- Untouched
- 32
Unknown Walltime execution environment detected
Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data. For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.Benchmarks
36 total
test_benchmark_moffat_conv[run]
tests/jax/test_benchmarks.py×3.7
10.4 s2.8 s
test_benchmark_moffat_conv[run]
tests/jax/test_benchmarks.py×2.3
681.8 ms301.2 ms
test_benchmark_moffat_conv_grad[run]
tests/jax/test_benchmarks.py×2.2
13.6 s6.2 s
test_benchmark_moffat_conv_grad[run]
tests/jax/test_benchmarks.py+42%
1,279.6 ms899.7 ms
test_benchmark_moffat_init[run]
tests/jax/test_benchmarks.py+14%
135.9 µs118.8 µs
test_benchmark_interpimage_flux_frac[run]
tests/jax/test_benchmarks.py+13%
2.5 ms2.2 ms
test_benchmark_spergel_xvalue[run]
tests/jax/test_benchmarks.py+9%
1.6 s1.5 s
test_benchmark_moffat_init[run]
tests/jax/test_benchmarks.py+8%
491.8 µs454 µs
test_benchmark_invert_ab_noraise[run]
tests/jax/test_benchmarks.py+7%
131.4 µs123.1 µs
test_benchmarks_lanczos_interp[xval-no_conserve_dc-run]
tests/jax/test_benchmarks.py+6%
113.2 µs106.9 µs
test_benchmarks_lanczos_interp[xval-conserve_dc-run]
tests/jax/test_benchmarks.py+4%
122.8 µs118.1 µs
test_benchmarks_lanczos_interp[kval-conserve_dc-run]
tests/jax/test_benchmarks.py+3%
46.3 µs45 µs
test_benchmark_spergel_calcfluxrad[run]
tests/jax/test_benchmarks.py+2%
1.1 ms1.1 ms
test_benchmark_spergel_conv[run]
tests/jax/test_benchmarks.py+2%
279.5 ms273.5 ms
test_benchmark_interpimage_flux_frac[run]
tests/jax/test_benchmarks.py+2%
436.1 µs427.6 µs
test_benchmark_spergel_init[run]
tests/jax/test_benchmarks.py+2%
6 µs5.9 µs
test_benchmark_rng_discard[run]
tests/jax/test_benchmarks.py+1%
109.9 µs108.4 µs
test_benchmark_gaussian_init[run]
tests/jax/test_benchmarks.py0%
6 µs6 µs
test_benchmarks_lanczos_interp[kval-no_conserve_dc-run]
tests/jax/test_benchmarks.py0%
44.7 µs44.5 µs
test_benchmarks_metacal[run]
tests/jax/test_benchmarks.py0%
179.6 ms179.2 ms
test_benchmarks_lanczos_interp[xval-conserve_dc-run]
tests/jax/test_benchmarks.py0%
1 ms1 ms
test_benchmark_spergel_kvalue[run]
tests/jax/test_benchmarks.py0%
122.1 ms121.9 ms
test_benchmarks_interpolated_image[run]
tests/jax/test_benchmarks.py0%
5.1 ms5.1 ms
test_benchmark_spergel_conv[run]
tests/jax/test_benchmarks.py0%
2.6 s2.6 s
test_benchmark_spergel_xvalue[run]
tests/jax/test_benchmarks.py0%
33.6 s33.6 s
Commits
Click on a commit to change the comparison rangeBase
main
51df286
×4
fix: ensure moffat derivs are not nan
bb69aa4
3 days ago
by beckermr+0.11%
[pre-commit.ci] auto fixes from pre-commit.com hooks
7ef38d4
3 days ago
by pre-commit-ci[bot]-10.83%
fix: do not use lazy property
ea7bde8
3 days ago
by beckermr-0.32%
fix: match the maxk for truncation in other parts of the code
2d7b9bf
3 days ago
by beckermr+0.22%
Merge branch 'debug-moffat-perf' of https://github.com/GalSim-developers/JAX-GalSim into debug-moffat-perf
aa433e4
3 days ago
by beckermr+0.38%
fix: adjust tests again
33f12a6
3 days ago
by beckermr-22.38%
feat: use aymptotic expansion with untruncated profiles
68d4334
2 days ago
by beckermrfix: use slope properly
8db2728
2 days ago
by beckermrfix: nan derivs needs mask at k=0
be849d3
2 days ago
by beckermr×4
fix: wrong way to truncate
51421d7
2 days ago
by beckermr-0.59%
fix: wrong value at k = 0
3b4cf52
2 days ago
by beckermrperf: use fewer points
0fc251d
2 days ago
by beckermr0%
fix: enable derivs for moffats with truncation
99df4fc
2 days ago
by beckermr×4.7
fix: performance regression for moffat init and test derivs with trunc
ecba9da
2 days ago
by beckermr