llvm-project.git/mlir/lib/Interfaces/VectorInterfaces.cpp, branch main Unnamed repository; edit this file 'description' to name the repository. Reland "[mlir][Vector] Re-define masking semantics in vector.transfer ops"" 2022-11-29T03:36:54+00:00 Diego Caballero diegocaballero@google.com 2022-11-18T18:35:32+00:00 eb7e2998d135ac30198cb7e6709db3bdc155f2d0 This relands commit 847b5f82a4a34218bf16d6f83f1b7c32df3117ba. Differential Revision: https://reviews.llvm.org/D138079 
This relands commit 847b5f82a4a34218bf16d6f83f1b7c32df3117ba.

Differential Revision: https://reviews.llvm.org/D138079
Revert "[mlir][Vector] Re-define masking semantics in vector.transfer ops" 2022-11-18T01:18:11+00:00 Diego Caballero diegocaballero@google.com 2022-11-18T01:18:11+00:00 847b5f82a4a34218bf16d6f83f1b7c32df3117ba This reverts commit 6c59c5cd08c879c9d1cfa653711613244a7c39bf. 
This reverts commit 6c59c5cd08c879c9d1cfa653711613244a7c39bf.
[mlir][Vector] Re-define masking semantics in vector.transfer ops 2022-11-18T01:05:42+00:00 Diego Caballero diegocaballero@google.com 2022-11-16T00:43:50+00:00 6c59c5cd08c879c9d1cfa653711613244a7c39bf Masking hasn't been widely used in vector transfer ops and the semantics of the mask operand were a bit loose. This patch states that the mask operand in a vector transfer op is applied to the read/write part of the operation and, therefore, its shape should match the shape of the elements read/written from/into the memref/tensor regardless of any permutation/broadcasting also applied by the transfer operation. Reviewers: nicolasvasilache Differential Revision: https://reviews.llvm.org/D138079 
Masking hasn't been widely used in vector transfer ops and the semantics
of the mask operand were a bit loose. This patch states that the mask
operand in a vector transfer op is applied to the read/write part of the
operation and, therefore, its shape should match the shape of the
elements read/written from/into the memref/tensor regardless of any
permutation/broadcasting also applied by the transfer operation.

Reviewers: nicolasvasilache

Differential Revision: https://reviews.llvm.org/D138079
[mlir][Vector] Thread 0-d vectors through vector.transfer ops 2021-12-01T16:49:43+00:00 Nicolas Vasilache nicolas.vasilache@gmail.com 2021-12-01T15:48:17+00:00 c537a943342be66d0876c6440a2df317b572c092 This revision adds 0-d vector support to vector.transfer ops. In the process, numerous cleanups are applied, in particular around normalizing and reducing the number of builders. Reviewed By: ThomasRaoux, springerm Differential Revision: https://reviews.llvm.org/D114803 
This revision adds 0-d vector support to vector.transfer ops.
In the process, numerous cleanups are applied, in particular around normalizing
and reducing the number of builders.

Reviewed By: ThomasRaoux, springerm

Differential Revision: https://reviews.llvm.org/D114803
[mlir] Unrolled progressive-vector-to-scf. 2021-05-13T04:08:48+00:00 Matthias Springer springerm@google.com 2021-05-13T03:57:45+00:00 9b77be5583d2da03f2ccd7319d33a2daedf8b1b3 Instead of an SCF for loop, these pattern generate fully unrolled loops with no temporary buffer allocations. Differential Revision: https://reviews.llvm.org/D101981 
Instead of an SCF for loop, these pattern generate fully unrolled loops with no temporary buffer allocations.

Differential Revision: https://reviews.llvm.org/D101981
[mlir] Fix masked vector transfer ops with broadcasts 2021-05-13T03:46:03+00:00 Matthias Springer springerm@google.com 2021-05-13T02:57:10+00:00 c52cbe63e42fff8c1a95921effd35d7bb59301d3 Broadcast dimensions of a vector transfer op have no corresponding dimension in the mask vector. E.g., a 2-D TransferReadOp, where one dimension is a broadcast, can have a 1-D `mask` attribute. This commit also adds a few additional transfer op integration tests for various combinations of broadcasts, masking, dim transposes, etc. Differential Revision: https://reviews.llvm.org/D101745 
Broadcast dimensions of a vector transfer op have no corresponding dimension in the mask vector. E.g., a 2-D TransferReadOp, where one dimension is a broadcast, can have a 1-D `mask` attribute.

This commit also adds a few additional transfer op integration tests for various combinations of broadcasts, masking, dim transposes, etc.

Differential Revision: https://reviews.llvm.org/D101745
Revert "[mlir] Fix masked vector transfer ops with broadcasts" 2021-05-13T02:55:00+00:00 Matthias Springer springerm@google.com 2021-05-13T02:55:00+00:00 6555e53ab0f2bca3dc30f5ab3a2d6872d50b6ff8 This reverts commit c9087788f7e41285445729127dd07ff7f82e3fc0. Accidentally pushed old version of the commit. 
This reverts commit c9087788f7e41285445729127dd07ff7f82e3fc0.

Accidentally pushed old version of the commit.
[mlir] Fix masked vector transfer ops with broadcasts 2021-05-13T02:37:36+00:00 Matthias Springer springerm@google.com 2021-05-03T09:03:19+00:00 c9087788f7e41285445729127dd07ff7f82e3fc0 Broadcast dimensions of a vector transfer op have no corresponding dimension in the mask vector. E.g., a 2-D TransferReadOp, where one dimension is a broadcast, can have a 1-D `mask` attribute. This commit also adds a few additional transfer op integration tests for various combinations of broadcasts, masking, dim transposes, etc. Differential Revision: https://reviews.llvm.org/D101745 
Broadcast dimensions of a vector transfer op have no corresponding dimension in the mask vector. E.g., a 2-D TransferReadOp, where one dimension is a broadcast, can have a 1-D `mask` attribute.

This commit also adds a few additional transfer op integration tests for various combinations of broadcasts, masking, dim transposes, etc.

Differential Revision: https://reviews.llvm.org/D101745
[mlir] NFC: fix trivial typos 2020-10-28T19:05:22+00:00 Kazuaki Ishizaki ishizaki@jp.ibm.com 2020-10-28T19:03:15+00:00 41b09f4efff1a9cd82af2d7d7eeb9916a88332e5 fix typos in comments and documents Reviewed By: jpienaar Differential Revision: https://reviews.llvm.org/D90089 
fix typos in comments and documents

Reviewed By: jpienaar

Differential Revision: https://reviews.llvm.org/D90089
[mlir][Vector] NFC - Improve VectorInterfaces 2020-07-20T12:24:22+00:00 Nicolas Vasilache ntv@google.com 2020-07-20T12:15:31+00:00 47cbd9f92282e3a19f161053cfbf77a7691de43e This revision improves and makes better use of OpInterfaces for the Vector dialect. Differential Revision: https://reviews.llvm.org/D84053 
This revision improves and makes better use of OpInterfaces for the Vector dialect.

Differential Revision: https://reviews.llvm.org/D84053