Compiler optimizations

[Bisaloo]

No description provided.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

❌ Patch coverage is 83.92857% with 9 lines in your changes missing coverage. Please review.
✅ Project coverage is 90.96%. Comparing base (b846220) to head (139b4ed).

Files with missing lines	Patch %	Lines
src/float16_conversion.h	61.11%	7 Missing ⚠️
src/bit64_conversion.h	93.93%	2 Missing ⚠️
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #21      +/-   ##
==========================================
+ Coverage   90.11%   90.96%   +0.84%     
==========================================
  Files           6        6              
  Lines         354      354              
==========================================
+ Hits          319      322       +3     
+ Misses         35       32       -3     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 466cb87 is merged into main:

• ✔️pkg_load: 3.65ms -> 3.68ms [-0.81%, +2.3%]
• ✔️read_float16: 586µs -> 594µs [-0.04%, +2.9%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 7169cdf is merged into main:

• ✔️pkg_load: 6.02ms -> 6.04ms [-1.73%, +2.41%]
• ✔️read_bigendian: 850µs -> 857µs [-0.39%, +2.15%]
• ✔️read_bool: 878µs -> 870µs [-3.81%, +2%]
• ✔️read_empty: 834µs -> 839µs [-0.45%, +1.68%]
• ✔️read_float16: 876µs -> 878µs [-2.66%, +3.07%]
• ✔️read_float32: 858µs -> 861µs [-0.97%, +1.6%]
• ✔️read_float64: 870µs -> 869µs [-1.38%, +1.25%]
• ✔️read_int16: 860µs -> 862µs [-1.59%, +1.98%]
• ❗🐌read_int32: 855µs -> 871µs [+0.54%, +3.12%]
• ✔️read_int64: 1.29ms -> 1.3ms [-0.49%, +2.01%]
• ✔️read_int8: 878µs -> 879µs [-1.41%, +1.63%]
• ✔️read_npz: 2.54ms -> 2.58ms [-0.68%, +3.96%]
• ✔️read_string: 766µs -> 778µs [-2.07%, +5.09%]
• ❗🐌read_structured: 1.59ms -> 1.62ms [+0.48%, +3.74%]
• ✔️read_uint16: 834µs -> 837µs [-0.81%, +1.43%]
• ❗🐌read_uint32: 1.1ms -> 1.11ms [+0.14%, +2.44%]
• ✔️read_uint64: 1.25ms -> 1.26ms [-0.64%, +2.09%]
• ✔️read_uint8: 830µs -> 834µs [-0.51%, +1.54%]
• ✔️read_unicode: 923µs -> 930µs [-1.23%, +2.74%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[Bisaloo]

On 1000x1000 arrays:

float16

✔ Installing branch 'compiler-optimizations'
✔ Installing branch 'main'
✔ Running `expr` across variants
# A tibble: 2 × 14
  branch            expression    min median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time result   memory     time       gc      
  <chr>             <bch:expr> <bch:> <bch:>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm> <list>   <list>     <list>     <list>  
1 compiler-optimiz… "read_npy… 1.02ms 3.52ms      283.    9.68MB     71.3  7989  2011      28.2s <dbl[…]> <Rprofmem> <bench_tm> <tibble>
2 main              "read_npy… 1.31ms 3.67ms      271.    9.68MB     68.1  7989  2011      29.5s <dbl[…]> <Rprofmem> <bench_tm> <tibble>

uint32

✔ Installing branch 'compiler-optimizations'
✔ Installing branch 'main'
✔ Running `expr` across variants
# A tibble: 2 × 14
  branch            expression    min median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time result   memory     time       gc      
  <chr>             <bch:expr> <bch:> <bch:>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm> <list>   <list>     <list>     <list>  
1 compiler-optimiz… "read_npy… 3.51ms 3.85ms      247.    7.84MB     74.6  7684  2316      31.1s <int[…]> <Rprofmem> <bench_tm> <tibble>
2 main              "read_npy… 3.71ms 4.12ms      231.    7.84MB     69.5  7684  2316      33.3s <int[…]> <Rprofmem> <bench_tm> <tibble>

int64

✔ Installing branch 'compiler-optimizations'
✔ Installing branch 'main'
✔ Running `expr` across variants
# A tibble: 2 × 14
  branch            expression    min median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time result   memory     time       gc      
  <chr>             <bch:expr> <bch:> <bch:>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm> <list>   <list>     <list>     <list>  
1 compiler-optimiz… "read_npy… 6.94ms 7.65ms      131.    11.7MB     43.7  7501  2499      57.2s <int[…]> <Rprofmem> <bench_tm> <tibble>
2 main              "read_npy… 6.85ms 7.72ms      129.    11.7MB     43.0  7501  2499      58.1s <int[…]> <Rprofmem> <bench_tm> <tibble>

[Bisaloo]

uint32

✔ Installing branch 'compiler-optimizations'
✔ Installing branch 'main'
✔ Running `expr` across variants
# A tibble: 2 × 14
  branch                 expression                                       min   median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time result   memory     time       gc      
  <chr>                  <bch:expr>                                  <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm> <list>   <list>     <list>     <list>  
1 compiler-optimizations "read_npy(\"inst/extdata/test_large.npy\")"   1.11ms   1.24ms      691.    7.84MB    208.   7684  2316      11.1s <int[…]> <Rprofmem> <bench_tm> <tibble>
2 main                   "read_npy(\"inst/extdata/test_large.npy\")"   3.65ms   3.87ms      245.    7.84MB     73.9  7684  2316      31.3s <int[…]> <Rprofmem> <bench_tm> <tibble>

uint64

✔ Installing branch 'compiler-optimizations'
✔ Installing branch 'main'
✔ Running `expr` across variants
# A tibble: 2 × 14
  branch                 expression                                       min   median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time result   memory     time       gc      
  <chr>                  <bch:expr>                                  <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm> <list>   <list>     <list>     <list>  
1 compiler-optimizations "read_npy(\"inst/extdata/test_large.npy\")"   4.93ms   5.59ms      178.    11.7MB     59.3  7501  2499      42.2s <int[…]> <Rprofmem> <bench_tm> <tibble>
2 main                   "read_npy(\"inst/extdata/test_large.npy\")"   6.97ms    7.8ms      128.    11.7MB     42.8  7501  2499      58.4s <int[…]> <Rprofmem> <bench_tm> <tibble>

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 139b4ed is merged into main:

• ✔️pkg_load: 5.13ms -> 5.13ms [-1.63%, +1.39%]
• ❗🐌read_bigendian: 802µs -> 819µs [+1.1%, +3.18%]
• ✔️read_bool: 822µs -> 834µs [-1.99%, +4.76%]
• ✔️read_empty: 783µs -> 790µs [-0.36%, +2.17%]
• ❗🐌read_float16: 836µs -> 849µs [+0.58%, +2.57%]
• ✔️read_float32: 828µs -> 836µs [-0.2%, +2%]
• ✔️read_float64: 839µs -> 828µs [-5.69%, +3.05%]
• ✔️read_int16: 838µs -> 845µs [-0.23%, +1.87%]
• ❗🐌read_int32: 833µs -> 838µs [+0.01%, +1.19%]
• ✔️read_int64: 1.24ms -> 1.25ms [-0.08%, +2.84%]
• ❗🐌read_int8: 850µs -> 865µs [+0.56%, +3.15%]
• ❗🐌read_npz: 2.4ms -> 2.44ms [+0.88%, +2.98%]
• ✔️read_string: 722µs -> 737µs [-1.63%, +5.83%]
• ❗🐌read_structured: 1.48ms -> 1.5ms [+0.31%, +2.87%]
• ❗🐌read_uint16: 771µs -> 782µs [+0.66%, +2.17%]
• ❗🐌read_uint32: 1.02ms -> 1.03ms [+0.78%, +1.86%]
• ❗🐌read_uint64: 1.15ms -> 1.16ms [+0.12%, +1.51%]
• ❗🐌read_uint8: 771µs -> 783µs [+0.48%, +2.67%]
• ✔️read_unicode: 829µs -> 840µs [-0.13%, +2.67%]
  Further explanation regarding interpretation and methodology can be found in the documentation.