Code

Add sparse matrix to mp#860

Code

Open

Xewar313 wants to merge 68 commits intooneapi-src:mainoneapi-src/distributed-ranges:mainfrom Xewar313:benchmarkXewar313/distributed-ranges:benchmarkCopy head branch name to clipboard

+3,526-423

Commits

Commits on Aug 14, 2024

• Add initial implementation of sparse matrix in mp

  Xewar313

  committed
  00f1e39

  Copy full SHA for 00f1e39

• Fixed row shape calculation

  Xewar313

  committed
  36d7f38

  Copy full SHA for 36d7f38

Commits on Aug 19, 2024

• Extract matrix format from matrix implementation

  Xewar313

  committed
  dd57bb1

  Copy full SHA for dd57bb1

Commits on Aug 21, 2024

• Add initial gemv implementation

  Xewar313

  committed
  b84ecc0

  Copy full SHA for b84ecc0

• Move matrix related files from sp to general module

  Xewar313

  committed
  1c1dad7

  Copy full SHA for 1c1dad7

Commits on Aug 27, 2024

• Separated matrix format from mp sparse matrix implementation and added second format

  Xewar313

  committed
  2456379

  Copy full SHA for 2456379

Commits on Aug 28, 2024

• Improve matrix loading performance

  Xewar313

  committed
  bd63c2c

  Copy full SHA for bd63c2c

Commits on Sep 3, 2024

• Add sycl support to mp sparse matrixes

  Xewar313

  committed
  b75b7ed

  Copy full SHA for b75b7ed

Commits on Sep 4, 2024

• Added initialization from one node in mp sparse matrix

  Xewar313

  committed
  756fab1

  Copy full SHA for 756fab1

• Add concept requirement for gemv operation

  Xewar313

  committed
  0b498ad

  Copy full SHA for 0b498ad

• Initial improvement to matrix reading

  Xewar313

  committed
  bbf2acf

  Copy full SHA for bbf2acf

Commits on Sep 9, 2024

• Add small improvements to matrix loading

  Xewar313

  committed
  9eca244

  Copy full SHA for 9eca244

• Fix formatting

  Xewar313

  committed
  7b55a1b

  Copy full SHA for 7b55a1b

Commits on Sep 17, 2024

• Add sparse benchmark and broadcasted vector

  Xewar313

  committed
  bb2e02e

  Copy full SHA for bb2e02e

• Add benchmarking tools

  Xewar313

  committed
  18165da

  Copy full SHA for 18165da

Commits on Sep 18, 2024

• Add gemv benchmark to gbench

  Xewar313

  committed
  94f818e

  Copy full SHA for 94f818e

Commits on Sep 24, 2024

• Add reference gemv implementation

  Xewar313

  committed
  982a0e0

  Copy full SHA for 982a0e0

• Fixed gemv reference

  Xewar313

  committed
  47a8455

  Copy full SHA for 47a8455

Commits on Sep 25, 2024

• Fixed gemv benchmark implementation

  Xewar313

  committed
  a97a97b

  Copy full SHA for a97a97b

Commits on Sep 30, 2024

• Fix band csr generation

  Xewar313

  committed
  231a09a

  Copy full SHA for 231a09a

Commits on Oct 1, 2024

• Add support for slim matrix multiplication

  Xewar313

  committed
  6b8af49

  Copy full SHA for 6b8af49

• Fix benchmark and band csr generation

  Xewar313

  committed
  628aa07

  Copy full SHA for 628aa07

• Merge branch 'benchmark' of github.com:Xewar313/distributed-ranges into benchmark

  Xewar313

  committed
  2047ecf

  Copy full SHA for 2047ecf

Commits on Oct 2, 2024

• Add support to device based computing in distributed sparse matrix

  Xewar313

  committed
  4f12327

  Copy full SHA for 4f12327

• add broadcasted slim matrix device memory support

  Xewar313

  committed
  71bd336

  Copy full SHA for 71bd336

Commits on Oct 7, 2024

• Fix issue with inconsistent timing when using mp gemv

  Xewar313

  committed
  6f96929

  Copy full SHA for 6f96929

Commits on Oct 8, 2024

• Some fixes to sparse matrixes

  Xewar313

  committed
  08a2247

  Copy full SHA for 08a2247

Commits on Oct 9, 2024

• improve work division in csr eq distribution

  Xewar313

  committed
  6a4bd30

  Copy full SHA for 6a4bd30

Commits on Oct 16, 2024

• Add better work distribution to csr_row_distiribution and fix distributed_vector rma memory access

  Xewar313

  committed
  f93961b

  Copy full SHA for f93961b

Commits on Oct 18, 2024

• improve performance on less dense matrices and allow broadcasting bigger matrices

  Xewar313

  committed
  e421523

  Copy full SHA for e421523

• Reversed change to eq distribution

  Xewar313

  committed
  3fa4a68

  Copy full SHA for 3fa4a68

Commits on Oct 25, 2024

• update some examples and benchmarks

  Xewar313

  committed
  0a1a4dc

  Copy full SHA for 0a1a4dc

Commits on Nov 4, 2024

• Improved communication in eq distribution

  Xewar313

  committed
  2beec18

  Copy full SHA for 2beec18

Commits on Nov 5, 2024

• Improve equ format on very sparse matrices

  Xewar313

  committed
  5edd0ba

  Copy full SHA for 5edd0ba

• Merge branch 'main' into benchmark

  Xewar313

  committed
  5800f99

  Copy full SHA for 5800f99

• Fix test compilation

  Xewar313

  committed
  cae67ef

  Copy full SHA for cae67ef

• Reformat changes in mp matrix oneapi-src#1

  Xewar313

  committed
  28519e0

  Copy full SHA for 28519e0

Commits on Nov 8, 2024

• Fix and improve tests

  Xewar313

  committed
  f2c2fbe

  Copy full SHA for f2c2fbe

• Add tests for sparse gemm in mp

  Xewar313

  committed
  3a00d61

  Copy full SHA for 3a00d61

• Reformat changes in mp matrix oneapi-src#2

  Xewar313

  committed
  2ec4b21

  Copy full SHA for 2ec4b21

Commits on Nov 12, 2024

• Fix compilation on borealis

  Xewar313

  committed
  05f5c63

  Copy full SHA for 05f5c63

Commits on Nov 13, 2024

• fix compilation

  Xewar313

  committed
  06a6628

  Copy full SHA for 06a6628

• Fix issues with very small and very big matrices

  Xewar313

  committed
  aa706f7

  Copy full SHA for aa706f7

• Merge branch 'main' into benchmark

  Xewar313

  committed
  6e0e9d2

  Copy full SHA for 6e0e9d2

• Fix compilation on older OneDpl

  Xewar313

  committed
  8f1a2b7

  Copy full SHA for 8f1a2b7

• Fix style

  Xewar313

  committed
  28e023e

  Copy full SHA for 28e023e

• Merge onedpl fix

  Xewar313

  committed
  e4ee8c7

  Copy full SHA for e4ee8c7

• Some fixes with verions

  Xewar313

  committed
  55185dc

  Copy full SHA for 55185dc

Commits on Nov 20, 2024

• Add local to csr_eq_segment

  Xewar313

  committed
  b7704ea

  Copy full SHA for b7704ea

Commits on Nov 22, 2024

• Add proper local method

  Xewar313

  committed
  4acbad6

  Copy full SHA for 4acbad6

Commits on Nov 25, 2024

• Add problem to review

  Xewar313

  committed
  1f84ba7

  Copy full SHA for 1f84ba7

• Moved local view to distribution

  Xewar313

  committed
  ba20ee3

  Copy full SHA for ba20ee3

• Add new example of not working code

  Xewar313

  committed
  bad5606

  Copy full SHA for bad5606

Commits on Nov 27, 2024

• Fix issue with lambda copy

  Xewar313

  committed
  8e7f1fe

  Copy full SHA for 8e7f1fe

• Make local work with shared memory

  Xewar313

  committed
  3503271

  Copy full SHA for 3503271

Commits on Nov 29, 2024

• Fix device memory when using local in row distribution

  Xewar313

  committed
  44a6e78

  Copy full SHA for 44a6e78

Commits on Dec 2, 2024

• Fix local in eq distribution

  Xewar313

  committed
  2bf503e

  Copy full SHA for 2bf503e

• Fix formatting

  Xewar313

  committed
  dc89bc8

  Copy full SHA for dc89bc8

• Reverse change in dr::transform_view

  Xewar313

  committed
  7e7f2d2

  Copy full SHA for 7e7f2d2

• Fix benchmark when default vector size is small

  Xewar313

  committed
  dd1d6ed

  Copy full SHA for dd1d6ed

Commits on Dec 3, 2024

• Fix issue when distributed vector is too small

  Xewar313

  committed
  e42cfa2

  Copy full SHA for e42cfa2

• Improve performance of eq distribution gather

  Xewar313

  committed
  2318a46

  Copy full SHA for 2318a46

• Remove unneccessary comment

  Xewar313

  committed
  4cfb110

  Copy full SHA for 4cfb110

Commits on Dec 9, 2024

• Add test for reduce and fix type error in sparse matrix local

  Xewar313

  committed
  04191d7

  Copy full SHA for 04191d7

• Add broadcast_vector tests

  Xewar313

  committed
  adad4f7

  Copy full SHA for adad4f7

• Fix formatting

  Xewar313

  committed
  f17243b

  Copy full SHA for f17243b

Commits on Jan 13, 2025

• Corrected gemv matrix creation

  Xewar313

  committed
  f1639b0

  Copy full SHA for f1639b0

• Fix formatting

  Xewar313

  committed
  3dfdac0

  Copy full SHA for 3dfdac0