Commits

Branch selector

 main

User selector

qedawkins

Datepicker

All time

Commit History

Commits on Oct 18, 2024

• [Codegen] Drop TransformStrategies (iree-org#18820)

  Show description for 012f8a6
  qedawkins

  authored
  012f8a6

  Copy full SHA for 012f8a6

Commits on Oct 15, 2024

• [Codegen] Replace LICM with a version that checks trip count (iree-org#18679)

  Show description for 3ccd4f1
  qedawkins

  authored
  3ccd4f1

  Copy full SHA for 3ccd4f1

Commits on Oct 13, 2024

• [Codegen][GPU] Add pass for fallback distribution (iree-org#18726)

  Show description for d7378bb
  qedawkins

  authored
  d7378bb

  Copy full SHA for d7378bb

Commits on Oct 9, 2024

• [Codegen][GPU] Do not promote fill matmul operands (iree-org#18722)

  Show description for 1b719b3
  qedawkins

  authored
  1b719b3

  Copy full SHA for 1b719b3

Commits on Oct 8, 2024

• [Codegen][GPU] Disable consumer fusion for multi use cases (iree-org#18723)

  Show description for 0e16a89
  qedawkins

  authored
  0e16a89

  Copy full SHA for 0e16a89

• [Codegen][GPU] Fix delinearized index order in forall resolution (iree-org#18724)

  qedawkins

  authored
  7fb28e0

  Copy full SHA for 7fb28e0

• [Codegen][GPU] Stop using redundant transfer hoisting pattern on memrefs (iree-org#18721)

  Show description for e8ff07e
  qedawkins

  authored
  e8ff07e

  Copy full SHA for e8ff07e

• [Codegen][GPU] Enable destination fusion for unit trip loops (iree-org#18674)

  Show description for f6a9b6b
  qedawkins

  authored
  f6a9b6b

  Copy full SHA for f6a9b6b

• [Codegen] Add option to disable copy vectorization (iree-org#18673)

  Show description for 5b0680d
  qedawkins

  authored
  5b0680d

  Copy full SHA for 5b0680d

Commits on Oct 7, 2024

• [Codegen][GPU] Improve loop fusion pattern verification (iree-org#18671)

  Show description for cc3b28f
  qedawkins

  authored
  cc3b28f

  Copy full SHA for cc3b28f

Commits on Oct 3, 2024

• [Codegen][GPU] Force linalg ops to always use the same vector size (iree-org#18654)

  Show description for 747081a
  qedawkins

  authored
  747081a

  Copy full SHA for 747081a

• [Codegen][GPU] Fuse into destinations for parallel tiling (iree-org#18666)

  Show description for 88153eb
  qedawkins

  authored
  88153eb

  Copy full SHA for 88153eb

• [Codegen][GPU] Fix alloc creation for dynamic outputs in loop fusion (iree-org#18665)

  Show description for 24ee841
  qedawkins

  authored
  24ee841

  Copy full SHA for 24ee841

Commits on Oct 2, 2024

• [Codegen] Allow vectorizing linalg.copy ops on memrefs (iree-org#18672)

  Show description for 206c1f2
  qedawkins

  authored
  206c1f2

  Copy full SHA for 206c1f2

Commits on Oct 1, 2024

• [Codegen] Add pass for unrolling annotated for loops (iree-org#18641)

  Show description for 451ef71
  qedawkins

  authored
  451ef71

  Copy full SHA for 451ef71

• [LinalgExt] Avoid rank reduction in Im2Col lowering pattern (iree-org#18637)

  Show description for 4d31d89
  qedawkins

  authored
  4d31d89

  Copy full SHA for 4d31d89

Commits on Sep 30, 2024

• [Codegen][GPU] Fix index delinearized index order (iree-org#18640)

  Show description for 5502ca4
  qedawkins

  authored
  5502ca4

  Copy full SHA for 5502ca4

Commits on Sep 27, 2024

• [Codegen][GPU] Make operand promotion controlled by lowering config (iree-org#18576)

  Show description for ff1b8b0
  qedawkins

  authored
  ff1b8b0

  Copy full SHA for ff1b8b0

Commits on Sep 26, 2024

• [Codegen][GPU] Change the location of barriers in forall fusion (iree-org#18542)

  Show description for 7db91ce
  qedawkins

  authored
  7db91ce

  Copy full SHA for 7db91ce

Commits on Sep 23, 2024

• [NFC][Codegen] Move LLVMCPUDropVectorUnitDims to Common (iree-org#18578)

  Show description for ac03e05
  qedawkins

  authored
  ac03e05

  Copy full SHA for ac03e05

Commits on Sep 22, 2024

• [Codegen][GPU] Use I64ArrayAttr for tile sizes for simpler printing (iree-org#18575)

  Show description for 5a6bd8d
  qedawkins

  authored
  5a6bd8d

  Copy full SHA for 5a6bd8d

Commits on Sep 19, 2024

• [Codegen][GPU] Add support for bufferizing iree_gpu.barrier_region (iree-org#18497)

  Show description for 73ffafb
  qedawkins

  authored
  73ffafb

  Copy full SHA for 73ffafb

• [Codegen][GPU] Add pass to combine adjacent barrier_region ops (iree-org#18541)

  Show description for 75d5aab
  qedawkins

  authored
  75d5aab

  Copy full SHA for 75d5aab

• [Codegen][GPU] Allow iree_gpu.barrier_region to take multiple operands/results (iree-org#18490)

  Show description for c9eca66
  qedawkins

  authored
  c9eca66

  Copy full SHA for c9eca66

• [Codegen][GPU] Allow odd workgroup sizes when resolving non-warp foralls (iree-org#18549)

  Show description for 04144f6
  qedawkins

  authored
  04144f6

  Copy full SHA for 04144f6

Commits on Sep 16, 2024

• [Codegen][LLVMGPU] Default to private memory space for scalar dispatches (iree-org#18523)

  Show description for f2eaa2a
  qedawkins

  authored
  f2eaa2a

  Copy full SHA for f2eaa2a

Commits on Sep 11, 2024

• [Codegen][GPU] Add pass to resolve scf.forall ops (iree-org#18394)

  Show description for 5a6521c
  qedawkins

  authored
  5a6521c

  Copy full SHA for 5a6521c

Commits on Sep 9, 2024

• [Codegen] Add patterns for folding away no-op slices (iree-org#18419)

  Show description for 97710b3
  qedawkins

  authored
  97710b3

  Copy full SHA for 97710b3

Commits on Sep 7, 2024

• [Codegen][GPU] Improve forall hoisting pattern for single trip loops (iree-org#18418)

  Show description for 6c9aad0
  qedawkins

  authored
  6c9aad0

  Copy full SHA for 6c9aad0

Commits on Sep 3, 2024

• [Codegen][GPU] Enable forall hoisting with other ops in loop (iree-org#18395)

  Show description for 9cbe380
  qedawkins

  authored
  9cbe380

  Copy full SHA for 9cbe380

Commits on Aug 28, 2024

• [Codegen][GPU] Defer ConvertToDPS until after pack/reshape propagation (iree-org#18375)

  Show description for aa364e4
  qedawkins

  authored
  aa364e4

  Copy full SHA for aa364e4

Commits on Aug 27, 2024

• [Codegen][GPU] Add support for WMMA_I32_16x16x16_I8 (iree-org#18372)

  Show description for 7050033
  qedawkins

  authored
  7050033

  Copy full SHA for 7050033

• [Codegen][GPU] Use alloca for private allocations on GPU (iree-org#18371)

  Show description for 271a14b
  qedawkins

  authored
  271a14b

  Copy full SHA for 271a14b

Commits on Aug 23, 2024

• [Codegen][GPU] Handle dynamic and unaligned cases in DerivedThreadConfig (iree-org#18281)

  Show description for eb694f1
  qedawkins

  authored
  eb694f1

  Copy full SHA for eb694f1

Commits on Aug 20, 2024

• Move LinalgQuantized* passes to GlobalOptimization (iree-org#18287)

  Show description for cea581f
  qedawkins

  authored
  cea581f

  Copy full SHA for cea581f

Pagination

Previous

Next