Remove explicit `--iree-mlir-to-vm-bytecode-module` flag usage. (#9445)
That flag is deprecated in favor of the (default) `--output-format=vm-bytecode`.
`iree_bytecode_module()` now always passes `--output-format=vm-bytecode` explicitly, while samples/docs/scripts depend on the implicit default mode. In the future (see discussion on https://github.com/google/iree/pull/9431), we can have `iree_c_module()` always pass `--output-format=vm-c`.
Also in this commit:
* Passing comment updates
* A few changes to `--iree-hal-target-backends=cpu`
diff --git a/build_tools/bazel/iree_check_test.bzl b/build_tools/bazel/iree_check_test.bzl
index e60d5d9..1cd20b6 100644
--- a/build_tools/bazel/iree_check_test.bzl
+++ b/build_tools/bazel/iree_check_test.bzl
@@ -35,13 +35,14 @@
driver: driver to run the module with. This can be omitted to test only
compilation, but consider omiting the driver as a hacky abuse of the
rule since compilation on its own not use iree-check-module.
- compiler_flags: additional flags to pass to the compiler. Bytecode translation and backend
- flags are passed automatically.
- runner_args: additional runner_args to pass to iree-check-module. The driver and input file
- are passed automatically.
- tags: additional tags to apply to the generated test. A tag "driver=DRIVER" is added
- automatically.
- target_cpu_features: currently unimplemented (must be empty), will eventually allow specifying target CPU features.
+ compiler_flags: additional flags to pass to the compiler. Bytecode output
+ format and backend flags are passed automatically.
+ runner_args: additional runner_args to pass to iree-check-module. The
+ driver and input file are passed automatically.
+ tags: additional tags to apply to the generated test. A tag
+ "driver=DRIVER" is added automatically.
+ target_cpu_features: currently unimplemented (must be empty), will
+ eventually allow specifying target CPU features.
timeout: timeout for the generated tests.
**kwargs: any additional attributes to pass to the underlying native_test.
"""
@@ -54,7 +55,6 @@
name = bytecode_module_name,
src = src,
flags = [
- "--iree-mlir-to-vm-bytecode-module",
"--mlir-print-op-on-diagnostic=false",
"--iree-hal-target-backends=%s" % target_backend,
] + compiler_flags,
@@ -99,16 +99,20 @@
driver: driver to run the module with. This can be omitted to test only
compilation, but consider omiting the driver as a hacky abuse of the
rule since compilation on its own not use iree-check-module.
- compiler_flags: additional flags to pass to the compiler. Bytecode translation and backend
- flags are passed automatically.
- runner_args: additional runner_args to pass to the underlying iree-check-module tests. The
- driver and input file are passed automatically. To use different runner_args per test,
- create a separate suite or iree_check_test.
- target_cpu_features: currently unimplemented (must be empty), will eventually allow specifying target CPU features.
- tags: tags to apply to the generated tests. Note that as in standard test suites, manual
- is treated specially and will also apply to the test suite itself.
+ compiler_flags: additional flags to pass to the compiler. Bytecode output
+ format and backend flags are passed automatically.
+ runner_args: additional runner_args to pass to the underlying
+ iree-check-module tests. The driver and input file are passed
+ automatically. To use different runner_args per test, create a
+ separate suite or iree_check_test.
+ target_cpu_features: currently unimplemented (must be empty), will
+ eventually allow specifying target CPU features.
+ tags: tags to apply to the generated tests. Note that as in standard test
+ suites, manual is treated specially and will also apply to the test
+ suite itself.
timeout: timeout for the generated tests.
- **kwargs: any additional attributes to pass to the underlying tests and test suite.
+ **kwargs: any additional attributes to pass to the underlying tests and
+ test suite.
"""
# We haven't implemented this so far because we have been using target_cpu_features so far only
@@ -168,18 +172,23 @@
Args:
name: name of the generated test suite.
srcs: source mlir files containing the module.
- target_backends_and_drivers: backend/driver pairs to compile and run the module, respectively.
- compiler_flags: additional flags to pass to the compiler. Bytecode translation and backend
- flags are passed automatically.
- runner_args: additional runner_args to pass to the underlying iree-check-module tests. The
- driver and input file are passed automatically. To use different runner_args per test,
- create a separate suite or iree_check_test.
- tags: tags to apply to the generated tests. Note that as in standard test suites, manual
- is treated specially and will also apply to the test suite itself.
- target_cpu_features_variants: list of target cpu features variants. Currently unimplemented, so each
- entry must be either "default" or start with "aarch64:" so as Bazel builds are currently x86-only,
- we know that it is correct to ignore this.
- **kwargs: any additional attributes to pass to the underlying tests and test suite.
+ target_backends_and_drivers: backend/driver pairs to compile and run the
+ module, respectively.
+ compiler_flags: additional flags to pass to the compiler. Bytecode output
+ format and backend flags are passed automatically.
+ runner_args: additional runner_args to pass to the underlying
+ iree-check-module tests. The driver and input file are passed
+ automatically. To use different runner_args per test, create a
+ separate suite or iree_check_test.
+ tags: tags to apply to the generated tests. Note that as in standard test
+ suites, manual is treated specially and will also apply to the test
+ suite itself.
+ target_cpu_features_variants: list of target cpu features variants.
+ Currently unimplemented, so each entry must be either "default" or
+ start with "aarch64:" so as Bazel builds are currently x86-only, we
+ know that it is correct to ignore this.
+ **kwargs: any additional attributes to pass to the underlying tests and
+ test suite.
"""
for target_cpu_features in target_cpu_features_variants:
