codecov[bot]
commented
2 years ago Codecov Report
Merging #138 (688ce41) into master (b11fc8a) will increase coverage by
0.27%. The diff coverage is100.00%.
@@ Coverage Diff @@
## master #138 +/- ##
==========================================
+ Coverage 92.62% 92.90% +0.27%
==========================================
Files 12 12
Lines 610 662 +52
==========================================
+ Hits 565 615 +50
- Misses 45 47 +2 | Impacted Files | Coverage Δ | |
|---|---|---|
| src/dualize.jl | 96.42% <100.00%> (+0.42%) |
:arrow_up: |
| src/dual_model_variables.jl | 86.95% <0.00%> (-1.14%) |
:arrow_down: |
| src/utils.jl | 100.00% <0.00%> (ø) |
|
| src/dual_sets.jl | 100.00% <0.00%> (ø) |
|
| src/structures.jl | 100.00% <0.00%> (ø) |
|
| src/add_dual_cone_constraint.jl | 100.00% <0.00%> (ø) |
|
| src/objective_coefficients.jl | 96.84% <0.00%> (+0.13%) |
:arrow_up: |
| src/dual_equality_constraints.jl | 90.07% <0.00%> (+0.39%) |
:arrow_up: |
| src/MOI_wrapper.jl | 92.39% <0.00%> (+0.78%) |
:arrow_up: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact),ø = not affected,? = missing dataPowered by Codecov. Last update b11fc8a...688ce41. Read the comment docs.
kocvara
Getting the set and function repeatedly for each index of the same constraint is slower than getting it once for each constraint.
On the benchmark of https://github.com/jump-dev/MathOptInterface.jl/issues/1796, in terms of allocations:
and in terms of time:
Overall, for thetaG11, this gives a 40000x speedup in allocation and 14000x speedup in time. Moreover, it's now below the time for reading the SDPA file in geometric format so it's unclear that reading it in standard form would be a huge gain compared to reading it in geometric format + dualizing so it closes https://github.com/jump-dev/Dualization.jl/issues/134.
Closes https://github.com/jump-dev/Dualization.jl/issues/134