Closed jrdegreeff closed 2 years ago
We should remove the results data. It would be interesting to setup Benchmarking reports.
Agreed. I'll add them to the .gitignore now. Separate benchmarking reports would be really cool to set up.
Merging #43 (df83189) into main (b8b4315) will decrease coverage by
1.93%
. The diff coverage is97.56%
.
@@ Coverage Diff @@
## main #43 +/- ##
===========================================
- Coverage 100.00% 98.06% -1.94%
===========================================
Files 11 11
Lines 154 155 +1
===========================================
- Hits 154 152 -2
- Misses 0 3 +3
Impacted Files | Coverage Δ | |
---|---|---|
src/types/empirical_potential.jl | 94.11% <96.77%> (-5.89%) |
:arrow_down: |
src/nnlist.jl | 94.73% <100.00%> (-5.27%) |
:arrow_down: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update b8b4315...df83189. Read the comment docs.
First round of performance enhancements. These changes separate the loop iteration from the core calculations and include many serial and parallel optimizations. Measured a 20x speedup for the 1000 atom lj cluster on my laptop. These changes set the stage for GPU acceleration which will come in the next PR. Also made some simplifications to the neighbor list structure which should ease transition to new neighbor list interface.