maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/mlir/test/mlir-cpu-runner/linalg_integration_test.mlir

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

84 lines
4.0 KiB
MLIR
Raw Normal View History

Automated rollback of commit d60133f89bb08341718bb3132b19bc891f7d4f4d PiperOrigin-RevId: 282574110
2019-11-27 00:47:14 +08:00
// RUN: mlir-opt %s -convert-linalg-to-llvm | mlir-cpu-runner -e dot -entry-point-result=f32 -shared-libs=%linalg_test_lib_dir/libcblas%shlibext,%linalg_test_lib_dir/libcblas_interface%shlibext | FileCheck %s
Implement Linalg to loops lowering as a pattern This CL rewrites the linalg ops to loops transformations as patterns that can be targeted directly from Tablegen. Reliance on OpFolder is removed and to cope with it we introduce local folding patterns that are applied greedily. PiperOrigin-RevId: 282765550
2019-11-27 23:31:41 +08:00
// RUN: mlir-opt %s -convert-linalg-to-loops -convert-linalg-to-llvm | mlir-cpu-runner -e dot -entry-point-result=f32 -shared-libs=%linalg_test_lib_dir/libcblas%shlibext,%linalg_test_lib_dir/libcblas_interface%shlibext | FileCheck %s
Automated rollback of commit d60133f89bb08341718bb3132b19bc891f7d4f4d PiperOrigin-RevId: 282574110
2019-11-27 00:47:14 +08:00
// RUN: mlir-opt %s -convert-linalg-to-llvm | mlir-cpu-runner -e matmul -entry-point-result=f32 -shared-libs=%linalg_test_lib_dir/libcblas%shlibext,%linalg_test_lib_dir/libcblas_interface%shlibext | FileCheck %s
Implement Linalg to loops lowering as a pattern This CL rewrites the linalg ops to loops transformations as patterns that can be targeted directly from Tablegen. Reliance on OpFolder is removed and to cope with it we introduce local folding patterns that are applied greedily. PiperOrigin-RevId: 282765550
2019-11-27 23:31:41 +08:00
// RUN: mlir-opt %s -convert-linalg-to-loops -convert-linalg-to-llvm | mlir-cpu-runner -e matmul -entry-point-result=f32 -shared-libs=%linalg_test_lib_dir/libcblas%shlibext,%linalg_test_lib_dir/libcblas_interface%shlibext | FileCheck %s
[mlir] Move several static cl::opts to be pass options instead. This removes the reliance on global options, and also simplifies the pass registration. Differential Revision: https://reviews.llvm.org/D76552
2020-03-22 06:08:49 +08:00
// RUN: mlir-opt %s -linalg-tile="linalg-tile-sizes=2,3,4" -linalg-promote-subviews -convert-linalg-to-loops -convert-linalg-to-llvm | mlir-cpu-runner -e matmul -entry-point-result=f32 -shared-libs=%linalg_test_lib_dir/libcblas%shlibext,%linalg_test_lib_dir/libcblas_interface%shlibext | FileCheck %s
// RUN: mlir-opt %s -linalg-tile="linalg-tile-sizes=2,3,4" -linalg-promote-subviews -convert-linalg-to-llvm | mlir-cpu-runner -e matmul -entry-point-result=f32 -shared-libs=%linalg_test_lib_dir/libcblas%shlibext,%linalg_test_lib_dir/libcblas_interface%shlibext | FileCheck %s
Unify Linalg types by using strided memrefs This CL finishes the implementation of the Linalg + Affine type unification of the [strided memref RFC](https://groups.google.com/a/tensorflow.org/forum/#!topic/mlir/MaL8m2nXuio). As a consequence, the !linalg.view type, linalg::DimOp, linalg::LoadOp and linalg::StoreOp can now disappear and Linalg can use standard types everywhere. PiperOrigin-RevId: 272187165
2019-10-01 20:22:54 +08:00
[mlir] Change the syntax of AffineMapAttr and IntegerSetAttr to avoid conflicts with function types. Summary: The current syntax for AffineMapAttr and IntegerSetAttr conflict with function types, making it currently impossible to round-trip function types(and e.g. FuncOp) in the IR. This revision changes the syntax for the attributes by wrapping them in a keyword. AffineMapAttr is wrapped with `affine_map<>` and IntegerSetAttr is wrapped with `affine_set<>`. Reviewed By: nicolasvasilache, ftynse Differential Revision: https://reviews.llvm.org/D72429
2020-01-14 05:12:37 +08:00
#strided1D = affine_map<(d0) -> (d0)>
#strided2D = affine_map<(d0, d1)[s0] -> (d0 * s0 + d1)>
Pipe Linalg to LLVM via mlir-cpu-runner This CL adds support for functions in the Linalg dialect to run with mlir-cpu-runner. For this purpose, this CL adds BufferAllocOp, BufferDeallocOp, LoadOp and StoreOp to the Linalg dialect as well as their lowering to LLVM. To avoid collisions with mlir::LoadOp/StoreOp (which should really become mlir::affine::LoadOp/StoreOp), the mlir::linalg namespace is added. The execution uses a dummy linalg_dot function that just returns for now. In the future a proper library call will be used. -- PiperOrigin-RevId: 247476061
2019-05-10 03:34:04 +08:00
External library name mangling support for linalg. This CL introduces the ability to generate the external library name for Linalg operations. The problem is that neither mlir or C support overloading and we want a simplified form of name mangling that is still reasonable to read. This CL creates the name of the external call that Linalg expects from the operation name and the type of its arguments. The interface library names are updated and use new cases are added for FillOp. PiperOrigin-RevId: 262556833
2019-08-09 22:33:34 +08:00
// Creates and returns a 1-D buffer of size %s filled with the value %f
Update Linalg to use std.view Now that a view op has graduated to the std dialect, we can update Linalg to use it and remove ops that have become obsolete. As a byproduct, the linalg buffer and associated ops can also disappear. PiperOrigin-RevId: 279073591
2019-11-07 22:32:39 +08:00
func @alloc_filled_f32(%s : index, %f : f32) -> memref<?xi8> {
Pipe Linalg to LLVM via mlir-cpu-runner This CL adds support for functions in the Linalg dialect to run with mlir-cpu-runner. For this purpose, this CL adds BufferAllocOp, BufferDeallocOp, LoadOp and StoreOp to the Linalg dialect as well as their lowering to LLVM. To avoid collisions with mlir::LoadOp/StoreOp (which should really become mlir::affine::LoadOp/StoreOp), the mlir::linalg namespace is added. The execution uses a dummy linalg_dot function that just returns for now. In the future a proper library call will be used. -- PiperOrigin-RevId: 247476061
2019-05-10 03:34:04 +08:00
%c0 = constant 0 : index
%c1 = constant 1 : index
Update Linalg to use std.view Now that a view op has graduated to the std dialect, we can update Linalg to use it and remove ops that have become obsolete. As a byproduct, the linalg buffer and associated ops can also disappear. PiperOrigin-RevId: 279073591
2019-11-07 22:32:39 +08:00
%c4 = constant 4 : index
%s4 = muli %s, %c4: index
%buf = alloc(%s4) {alignment = 256} : memref<?xi8>
%V = view %buf[%s][] : memref<?xi8> to memref<?xf32, #strided1D>
Unify Linalg types by using strided memrefs This CL finishes the implementation of the Linalg + Affine type unification of the [strided memref RFC](https://groups.google.com/a/tensorflow.org/forum/#!topic/mlir/MaL8m2nXuio). As a consequence, the !linalg.view type, linalg::DimOp, linalg::LoadOp and linalg::StoreOp can now disappear and Linalg can use standard types everywhere. PiperOrigin-RevId: 272187165
2019-10-01 20:22:54 +08:00
linalg.fill(%V, %f) : memref<?xf32, #strided1D>, f32
Update Linalg to use std.view Now that a view op has graduated to the std dialect, we can update Linalg to use it and remove ops that have become obsolete. As a byproduct, the linalg buffer and associated ops can also disappear. PiperOrigin-RevId: 279073591
2019-11-07 22:32:39 +08:00
return %buf : memref<?xi8>
Pipe Linalg to LLVM via mlir-cpu-runner This CL adds support for functions in the Linalg dialect to run with mlir-cpu-runner. For this purpose, this CL adds BufferAllocOp, BufferDeallocOp, LoadOp and StoreOp to the Linalg dialect as well as their lowering to LLVM. To avoid collisions with mlir::LoadOp/StoreOp (which should really become mlir::affine::LoadOp/StoreOp), the mlir::linalg namespace is added. The execution uses a dummy linalg_dot function that just returns for now. In the future a proper library call will be used. -- PiperOrigin-RevId: 247476061
2019-05-10 03:34:04 +08:00
}
External library name mangling support for linalg. This CL introduces the ability to generate the external library name for Linalg operations. The problem is that neither mlir or C support overloading and we want a simplified form of name mangling that is still reasonable to read. This CL creates the name of the external call that Linalg expects from the operation name and the type of its arguments. The interface library names are updated and use new cases are added for FillOp. PiperOrigin-RevId: 262556833
2019-08-09 22:33:34 +08:00
// Test for linalg.dot.
Add a lowering for Linalg matmul to LLVM This CL adds a lowering to LLVM for MamulOp and a corresponding integration test. View descriptor manipulation is moved from MLIR's LLVM dialect to C++ code compiled on the side. To this end a separation is introduced between `cblas.cpp` and `cblas_interface.cpp`, the latter operating on view types whose ABI correspond to the LLVM signature generated by MLIR. An intermediary step is introduced that allocates a new descriptor on the MLIR side for the purpose of passing it to LLVM. The reason for this extra step is that the ABI for by-value ViewType objects wants aligned descriptors, e.g.: ``` extern "C" void linalg_dot_impl(ViewType<float, 1> X, ViewType<float, 1> Y, BaseViewType<float> Z) { ... } ``` produces LLVM IR with the signature: ``` %struct.ViewType = type { %struct.BaseViewType, [1 x i64], [1 x i64] } %struct.BaseViewType = type { float*, i64 } define void @linalg_dot_impl(%struct.ViewType* byval align 8, %struct.ViewType* byval align 8, float*, i64) tensorflow/mlir#0 { ... } ``` We don't seem to be able to make such aligned allocations in the MLIR -> LLVM converter atm. Going through a level of indirection allows the test to pass. The temporary tradeoff is that the MLIR shims have to be written by hand. They will disappear in the future. PiperOrigin-RevId: 252670672
2019-06-12 03:09:50 +08:00
func @dot() -> f32 {
Pipe Linalg to LLVM via mlir-cpu-runner This CL adds support for functions in the Linalg dialect to run with mlir-cpu-runner. For this purpose, this CL adds BufferAllocOp, BufferDeallocOp, LoadOp and StoreOp to the Linalg dialect as well as their lowering to LLVM. To avoid collisions with mlir::LoadOp/StoreOp (which should really become mlir::affine::LoadOp/StoreOp), the mlir::linalg namespace is added. The execution uses a dummy linalg_dot function that just returns for now. In the future a proper library call will be used. -- PiperOrigin-RevId: 247476061
2019-05-10 03:34:04 +08:00
%c0 = constant 0 : index
%c1 = constant 1 : index
%c16 = constant 16 : index
Pipe Linalg to a cblas call via mlir-cpu-runner This CL extends the execution engine to allow the additional resolution of symbols names that have been registered explicitly. This allows linking static library symbols that have not been explicitly exported with the -rdynamic linking flag (which is deemed too intrusive). -- PiperOrigin-RevId: 247969504
2019-05-14 01:59:04 +08:00
%f10 = constant 10.00000e+00 : f32
%f1 = constant 1.00000e+00 : f32
Pipe Linalg to LLVM via mlir-cpu-runner This CL adds support for functions in the Linalg dialect to run with mlir-cpu-runner. For this purpose, this CL adds BufferAllocOp, BufferDeallocOp, LoadOp and StoreOp to the Linalg dialect as well as their lowering to LLVM. To avoid collisions with mlir::LoadOp/StoreOp (which should really become mlir::affine::LoadOp/StoreOp), the mlir::linalg namespace is added. The execution uses a dummy linalg_dot function that just returns for now. In the future a proper library call will be used. -- PiperOrigin-RevId: 247476061
2019-05-10 03:34:04 +08:00
%f2 = constant 2.00000e+00 : f32
Update Linalg to use std.view Now that a view op has graduated to the std dialect, we can update Linalg to use it and remove ops that have become obsolete. As a byproduct, the linalg buffer and associated ops can also disappear. PiperOrigin-RevId: 279073591
2019-11-07 22:32:39 +08:00
%bA = call @alloc_filled_f32(%c16, %f2) : (index, f32) -> (memref<?xi8>)
%bB = call @alloc_filled_f32(%c16, %f1) : (index, f32) -> (memref<?xi8>)
%bC = call @alloc_filled_f32(%c1, %f10) : (index, f32) -> (memref<?xi8>)
Add a lowering for Linalg matmul to LLVM This CL adds a lowering to LLVM for MamulOp and a corresponding integration test. View descriptor manipulation is moved from MLIR's LLVM dialect to C++ code compiled on the side. To this end a separation is introduced between `cblas.cpp` and `cblas_interface.cpp`, the latter operating on view types whose ABI correspond to the LLVM signature generated by MLIR. An intermediary step is introduced that allocates a new descriptor on the MLIR side for the purpose of passing it to LLVM. The reason for this extra step is that the ABI for by-value ViewType objects wants aligned descriptors, e.g.: ``` extern "C" void linalg_dot_impl(ViewType<float, 1> X, ViewType<float, 1> Y, BaseViewType<float> Z) { ... } ``` produces LLVM IR with the signature: ``` %struct.ViewType = type { %struct.BaseViewType, [1 x i64], [1 x i64] } %struct.BaseViewType = type { float*, i64 } define void @linalg_dot_impl(%struct.ViewType* byval align 8, %struct.ViewType* byval align 8, float*, i64) tensorflow/mlir#0 { ... } ``` We don't seem to be able to make such aligned allocations in the MLIR -> LLVM converter atm. Going through a level of indirection allows the test to pass. The temporary tradeoff is that the MLIR shims have to be written by hand. They will disappear in the future. PiperOrigin-RevId: 252670672
2019-06-12 03:09:50 +08:00
Update Linalg to use std.view Now that a view op has graduated to the std dialect, we can update Linalg to use it and remove ops that have become obsolete. As a byproduct, the linalg buffer and associated ops can also disappear. PiperOrigin-RevId: 279073591
2019-11-07 22:32:39 +08:00
%A = view %bA[%c16][] : memref<?xi8> to memref<?xf32, #strided1D>
%B = view %bB[%c16][] : memref<?xi8> to memref<?xf32, #strided1D>
%C = view %bC[][] : memref<?xi8> to memref<f32>
Pipe Linalg to a cblas call via mlir-cpu-runner This CL extends the execution engine to allow the additional resolution of symbols names that have been registered explicitly. This allows linking static library symbols that have not been explicitly exported with the -rdynamic linking flag (which is deemed too intrusive). -- PiperOrigin-RevId: 247969504
2019-05-14 01:59:04 +08:00
Unify Linalg types by using strided memrefs This CL finishes the implementation of the Linalg + Affine type unification of the [strided memref RFC](https://groups.google.com/a/tensorflow.org/forum/#!topic/mlir/MaL8m2nXuio). As a consequence, the !linalg.view type, linalg::DimOp, linalg::LoadOp and linalg::StoreOp can now disappear and Linalg can use standard types everywhere. PiperOrigin-RevId: 272187165
2019-10-01 20:22:54 +08:00
linalg.dot(%A, %B, %C) : memref<?xf32, #strided1D>, memref<?xf32, #strided1D>, memref<f32>
%res = load %C[] : memref<f32>
Pipe Linalg to a cblas call via mlir-cpu-runner This CL extends the execution engine to allow the additional resolution of symbols names that have been registered explicitly. This allows linking static library symbols that have not been explicitly exported with the -rdynamic linking flag (which is deemed too intrusive). -- PiperOrigin-RevId: 247969504
2019-05-14 01:59:04 +08:00
Update Linalg to use std.view Now that a view op has graduated to the std dialect, we can update Linalg to use it and remove ops that have become obsolete. As a byproduct, the linalg buffer and associated ops can also disappear. PiperOrigin-RevId: 279073591
2019-11-07 22:32:39 +08:00
dealloc %bC : memref<?xi8>
dealloc %bB : memref<?xi8>
dealloc %bA : memref<?xi8>
Pipe Linalg to a cblas call via mlir-cpu-runner This CL extends the execution engine to allow the additional resolution of symbols names that have been registered explicitly. This allows linking static library symbols that have not been explicitly exported with the -rdynamic linking flag (which is deemed too intrusive). -- PiperOrigin-RevId: 247969504
2019-05-14 01:59:04 +08:00
Pipe Linalg to LLVM via mlir-cpu-runner This CL adds support for functions in the Linalg dialect to run with mlir-cpu-runner. For this purpose, this CL adds BufferAllocOp, BufferDeallocOp, LoadOp and StoreOp to the Linalg dialect as well as their lowering to LLVM. To avoid collisions with mlir::LoadOp/StoreOp (which should really become mlir::affine::LoadOp/StoreOp), the mlir::linalg namespace is added. The execution uses a dummy linalg_dot function that just returns for now. In the future a proper library call will be used. -- PiperOrigin-RevId: 247476061
2019-05-10 03:34:04 +08:00
return %res : f32
}
External library name mangling support for linalg. This CL introduces the ability to generate the external library name for Linalg operations. The problem is that neither mlir or C support overloading and we want a simplified form of name mangling that is still reasonable to read. This CL creates the name of the external call that Linalg expects from the operation name and the type of its arguments. The interface library names are updated and use new cases are added for FillOp. PiperOrigin-RevId: 262556833
2019-08-09 22:33:34 +08:00
// Test for linalg.matmul.
Add a lowering for Linalg matmul to LLVM This CL adds a lowering to LLVM for MamulOp and a corresponding integration test. View descriptor manipulation is moved from MLIR's LLVM dialect to C++ code compiled on the side. To this end a separation is introduced between `cblas.cpp` and `cblas_interface.cpp`, the latter operating on view types whose ABI correspond to the LLVM signature generated by MLIR. An intermediary step is introduced that allocates a new descriptor on the MLIR side for the purpose of passing it to LLVM. The reason for this extra step is that the ABI for by-value ViewType objects wants aligned descriptors, e.g.: ``` extern "C" void linalg_dot_impl(ViewType<float, 1> X, ViewType<float, 1> Y, BaseViewType<float> Z) { ... } ``` produces LLVM IR with the signature: ``` %struct.ViewType = type { %struct.BaseViewType, [1 x i64], [1 x i64] } %struct.BaseViewType = type { float*, i64 } define void @linalg_dot_impl(%struct.ViewType* byval align 8, %struct.ViewType* byval align 8, float*, i64) tensorflow/mlir#0 { ... } ``` We don't seem to be able to make such aligned allocations in the MLIR -> LLVM converter atm. Going through a level of indirection allows the test to pass. The temporary tradeoff is that the MLIR shims have to be written by hand. They will disappear in the future. PiperOrigin-RevId: 252670672
2019-06-12 03:09:50 +08:00
func @matmul() -> f32 {
%c0 = constant 0 : index
%c1 = constant 1 : index
%c6 = constant 6 : index
%c7 = constant 7 : index
%c10 = constant 10 : index
%c16 = constant 16 : index
%c100 = constant 100 : index
%c160 = constant 160 : index
%f1 = constant 1.00000e+00 : f32
%f2 = constant 2.00000e+00 : f32
%f10 = constant 10.00000e+00 : f32
Update Linalg to use std.view Now that a view op has graduated to the std dialect, we can update Linalg to use it and remove ops that have become obsolete. As a byproduct, the linalg buffer and associated ops can also disappear. PiperOrigin-RevId: 279073591
2019-11-07 22:32:39 +08:00
%bA = call @alloc_filled_f32(%c160, %f2) : (index, f32) -> (memref<?xi8>)
%bB = call @alloc_filled_f32(%c160, %f1) : (index, f32) -> (memref<?xi8>)
%bC = call @alloc_filled_f32(%c100, %f10) : (index, f32) -> (memref<?xi8>)
Add a lowering for Linalg matmul to LLVM This CL adds a lowering to LLVM for MamulOp and a corresponding integration test. View descriptor manipulation is moved from MLIR's LLVM dialect to C++ code compiled on the side. To this end a separation is introduced between `cblas.cpp` and `cblas_interface.cpp`, the latter operating on view types whose ABI correspond to the LLVM signature generated by MLIR. An intermediary step is introduced that allocates a new descriptor on the MLIR side for the purpose of passing it to LLVM. The reason for this extra step is that the ABI for by-value ViewType objects wants aligned descriptors, e.g.: ``` extern "C" void linalg_dot_impl(ViewType<float, 1> X, ViewType<float, 1> Y, BaseViewType<float> Z) { ... } ``` produces LLVM IR with the signature: ``` %struct.ViewType = type { %struct.BaseViewType, [1 x i64], [1 x i64] } %struct.BaseViewType = type { float*, i64 } define void @linalg_dot_impl(%struct.ViewType* byval align 8, %struct.ViewType* byval align 8, float*, i64) tensorflow/mlir#0 { ... } ``` We don't seem to be able to make such aligned allocations in the MLIR -> LLVM converter atm. Going through a level of indirection allows the test to pass. The temporary tradeoff is that the MLIR shims have to be written by hand. They will disappear in the future. PiperOrigin-RevId: 252670672
2019-06-12 03:09:50 +08:00
Fix ViewOp to have at most one offset operand As described in the documentation, ViewOp is expected to take an optional dynamic offset followed by a list of dynamic sizes. However, the ViewOp parser did not include a check for the offset being a single value and accepeted a list of values instead. Furthermore, several tests have been exercising the wrong syntax of a ViewOp, passing multiple values to the dyanmic stride list, which was not caught by the parser. The trailing values could have been erronously interpreted as dynamic sizes. This is likely due to resyntaxing of the ViewOp, with the previous syntax taking the list of sizes before the offset. Update the tests to use the syntax with the offset preceding the sizes. Worse, the conversion of ViewOp to the LLVM dialect assumed the wrong order of operands with offset in the trailing position, and erronously relied on the permissive parsing that interpreted trailing dynamic offset values as leading dynamic sizes. Fix the lowering to use the correct order of operands. PiperOrigin-RevId: 283532506
2019-12-03 22:22:31 +08:00
%A = view %bA[][%c10, %c16] : memref<?xi8> to memref<?x?xf32, #strided2D>
%B = view %bB[][%c16, %c10] : memref<?xi8> to memref<?x?xf32, #strided2D>
%C = view %bC[][%c10, %c10] : memref<?xi8> to memref<?x?xf32, #strided2D>
Add a lowering for Linalg matmul to LLVM This CL adds a lowering to LLVM for MamulOp and a corresponding integration test. View descriptor manipulation is moved from MLIR's LLVM dialect to C++ code compiled on the side. To this end a separation is introduced between `cblas.cpp` and `cblas_interface.cpp`, the latter operating on view types whose ABI correspond to the LLVM signature generated by MLIR. An intermediary step is introduced that allocates a new descriptor on the MLIR side for the purpose of passing it to LLVM. The reason for this extra step is that the ABI for by-value ViewType objects wants aligned descriptors, e.g.: ``` extern "C" void linalg_dot_impl(ViewType<float, 1> X, ViewType<float, 1> Y, BaseViewType<float> Z) { ... } ``` produces LLVM IR with the signature: ``` %struct.ViewType = type { %struct.BaseViewType, [1 x i64], [1 x i64] } %struct.BaseViewType = type { float*, i64 } define void @linalg_dot_impl(%struct.ViewType* byval align 8, %struct.ViewType* byval align 8, float*, i64) tensorflow/mlir#0 { ... } ``` We don't seem to be able to make such aligned allocations in the MLIR -> LLVM converter atm. Going through a level of indirection allows the test to pass. The temporary tradeoff is that the MLIR shims have to be written by hand. They will disappear in the future. PiperOrigin-RevId: 252670672
2019-06-12 03:09:50 +08:00
Unify Linalg types by using strided memrefs This CL finishes the implementation of the Linalg + Affine type unification of the [strided memref RFC](https://groups.google.com/a/tensorflow.org/forum/#!topic/mlir/MaL8m2nXuio). As a consequence, the !linalg.view type, linalg::DimOp, linalg::LoadOp and linalg::StoreOp can now disappear and Linalg can use standard types everywhere. PiperOrigin-RevId: 272187165
2019-10-01 20:22:54 +08:00
linalg.matmul(%A, %B, %C) : memref<?x?xf32, #strided2D>, memref<?x?xf32, #strided2D>, memref<?x?xf32, #strided2D>
%res = load %C[%c6, %c7] : memref<?x?xf32, #strided2D>
Add a lowering for Linalg matmul to LLVM This CL adds a lowering to LLVM for MamulOp and a corresponding integration test. View descriptor manipulation is moved from MLIR's LLVM dialect to C++ code compiled on the side. To this end a separation is introduced between `cblas.cpp` and `cblas_interface.cpp`, the latter operating on view types whose ABI correspond to the LLVM signature generated by MLIR. An intermediary step is introduced that allocates a new descriptor on the MLIR side for the purpose of passing it to LLVM. The reason for this extra step is that the ABI for by-value ViewType objects wants aligned descriptors, e.g.: ``` extern "C" void linalg_dot_impl(ViewType<float, 1> X, ViewType<float, 1> Y, BaseViewType<float> Z) { ... } ``` produces LLVM IR with the signature: ``` %struct.ViewType = type { %struct.BaseViewType, [1 x i64], [1 x i64] } %struct.BaseViewType = type { float*, i64 } define void @linalg_dot_impl(%struct.ViewType* byval align 8, %struct.ViewType* byval align 8, float*, i64) tensorflow/mlir#0 { ... } ``` We don't seem to be able to make such aligned allocations in the MLIR -> LLVM converter atm. Going through a level of indirection allows the test to pass. The temporary tradeoff is that the MLIR shims have to be written by hand. They will disappear in the future. PiperOrigin-RevId: 252670672
2019-06-12 03:09:50 +08:00
Update Linalg to use std.view Now that a view op has graduated to the std dialect, we can update Linalg to use it and remove ops that have become obsolete. As a byproduct, the linalg buffer and associated ops can also disappear. PiperOrigin-RevId: 279073591
2019-11-07 22:32:39 +08:00
dealloc %bC : memref<?xi8>
dealloc %bB : memref<?xi8>
dealloc %bA : memref<?xi8>
Add a lowering for Linalg matmul to LLVM This CL adds a lowering to LLVM for MamulOp and a corresponding integration test. View descriptor manipulation is moved from MLIR's LLVM dialect to C++ code compiled on the side. To this end a separation is introduced between `cblas.cpp` and `cblas_interface.cpp`, the latter operating on view types whose ABI correspond to the LLVM signature generated by MLIR. An intermediary step is introduced that allocates a new descriptor on the MLIR side for the purpose of passing it to LLVM. The reason for this extra step is that the ABI for by-value ViewType objects wants aligned descriptors, e.g.: ``` extern "C" void linalg_dot_impl(ViewType<float, 1> X, ViewType<float, 1> Y, BaseViewType<float> Z) { ... } ``` produces LLVM IR with the signature: ``` %struct.ViewType = type { %struct.BaseViewType, [1 x i64], [1 x i64] } %struct.BaseViewType = type { float*, i64 } define void @linalg_dot_impl(%struct.ViewType* byval align 8, %struct.ViewType* byval align 8, float*, i64) tensorflow/mlir#0 { ... } ``` We don't seem to be able to make such aligned allocations in the MLIR -> LLVM converter atm. Going through a level of indirection allows the test to pass. The temporary tradeoff is that the MLIR shims have to be written by hand. They will disappear in the future. PiperOrigin-RevId: 252670672
2019-06-12 03:09:50 +08:00
return %res : f32
}
// All tests return this value
Fix MacOS test: use %shlibext in lit command line to expand to .dylib on MacOS and .so on Linux -- PiperOrigin-RevId: 249113478
2019-05-21 04:25:35 +08:00
// CHECK: 4.2{{0+}}e+01