Bohan Hou
859498dc01
## Summary This PR adds the initial TIRx support needed for low-level programming of Blackwell-class GPU architectures. As part of the ongoing TIRx refactor, it introduces TVMScript support for directly scripting advanced hardware features without relying on scheduling as the primary programming interface. The change keeps existing `s_tir` script support intact while making direct scripting a first-class path for TIRx programs. ## Main Changes - Add TIRx operator dispatch and layout infrastructure. - Add TVMScript support for new low-level TIRx operations. - Add analysis, transform, and lowering support for TIRx IR nodes. - Add CUDA/Blackwell-oriented codegen and intrinsic coverage. - Add Python and C++ integration points for TIRx scripting and runtime support. ## Validation - `pre-commit run --all-files` - `ninja -C build -j32` - `CUDA_VISIBLE_DEVICES=2 pytest tests/python/tirx/ -n 16` - `1723 passed, 47 skipped, 32 warnings` - `CUDA_VISIBLE_DEVICES=2 python -m pytest -v tests/python/all-platform-minimal-test` - `37 passed, 105 skipped` - `TVM_TEST_TARGETS=llvm python -m pytest -v tests/python/tirx-analysis tests/python/tirx-base tests/python/tirx-transform -n 16` - `664 passed, 25 skipped, 9 xfailed, 1 xpassed` ## Local CI Notes Some full CI-equivalent jobs were not locally reproducible because this machine is missing parts of the Apache TVM CI environment, including `llvm-config-15/17`, Vulkan, ROCm, Maven, Sphinx, Doxygen, Emscripten, and ARM/QEMU cross-toolchain components. Metal-specific tests were skipped locally because no Metal runtime is available.
87 lines
3.3 KiB
Python
87 lines
3.3 KiB
Python
# Licensed to the Apache Software Foundation (ASF) under one
|
|
# or more contributor license agreements. See the NOTICE file
|
|
# distributed with this work for additional information
|
|
# regarding copyright ownership. The ASF licenses this file
|
|
# to you under the Apache License, Version 2.0 (the
|
|
# "License"); you may not use this file except in compliance
|
|
# with the License. You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing,
|
|
# software distributed under the License is distributed on an
|
|
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
|
# KIND, either express or implied. See the License for the
|
|
# specific language governing permissions and limitations
|
|
# under the License.
|
|
|
|
import tvm
|
|
from tvm.script import tirx as T
|
|
|
|
|
|
def test_meta_programming_matmul():
|
|
def matmul_generator(M: int, N: int, K: int, dtype: str):
|
|
@T.prim_func(s_tir=True)
|
|
def matmul(a: T.handle, b: T.handle, c: T.handle) -> None:
|
|
A = T.match_buffer(a, [M, K], dtype=dtype)
|
|
B = T.match_buffer(b, [N, K], dtype=dtype)
|
|
C = T.match_buffer(c, [M, N], dtype=dtype)
|
|
|
|
for i, j, k in T.grid(M, N, K):
|
|
with T.sblock():
|
|
vi, vj, vk = T.axis.remap("SSR", [i, j, k])
|
|
with T.init():
|
|
C[vi, vj] = T.float32(0)
|
|
C[vi, vj] = C[vi, vj] + A[vi, vk] * B[vj, vk]
|
|
|
|
return matmul
|
|
|
|
@T.prim_func(s_tir=True)
|
|
def matmul_128_128_128_fp16(a: T.handle, b: T.handle, c: T.handle) -> None:
|
|
A = T.match_buffer(a, [128, 128], dtype="float16")
|
|
B = T.match_buffer(b, [128, 128], dtype="float16")
|
|
C = T.match_buffer(c, [128, 128], dtype="float16")
|
|
|
|
for i, j, k in T.grid(128, 128, 128):
|
|
with T.sblock():
|
|
vi, vj, vk = T.axis.remap("SSR", [i, j, k])
|
|
with T.init():
|
|
C[vi, vj] = T.float32(0)
|
|
C[vi, vj] = C[vi, vj] + A[vi, vk] * B[vj, vk]
|
|
|
|
f = matmul_generator(128, 128, 128, "float16").with_attr("global_symbol", "main")
|
|
tvm.ir.assert_structural_equal(f, matmul_128_128_128_fp16.with_attr("global_symbol", "main"))
|
|
|
|
|
|
def test_meta_programming_uncaptured_var():
|
|
def generate_erf(dtype):
|
|
@T.prim_func(s_tir=True)
|
|
def main(A: T.Buffer((1,), dtype), C: T.Buffer((1,), dtype)):
|
|
for i in range(1):
|
|
with T.sblock("C"):
|
|
C[i] = T.erf(A[i])
|
|
|
|
return main
|
|
|
|
@T.prim_func(s_tir=True)
|
|
def fp32(A: T.Buffer((1,), "float32"), C: T.Buffer((1,), "float32")):
|
|
for i in range(1):
|
|
with T.sblock("C"):
|
|
C[i] = T.erf(A[i])
|
|
|
|
@T.prim_func(s_tir=True)
|
|
def fp16(A: T.Buffer((1,), "float16"), C: T.Buffer((1,), "float16")):
|
|
for i in range(1):
|
|
with T.sblock("C"):
|
|
C[i] = T.erf(A[i])
|
|
|
|
f1 = generate_erf("float32").with_attr("global_symbol", "main")
|
|
tvm.ir.assert_structural_equal(f1, fp32.with_attr("global_symbol", "main"))
|
|
f2 = generate_erf("float16").with_attr("global_symbol", "main")
|
|
tvm.ir.assert_structural_equal(f2, fp16.with_attr("global_symbol", "main"))
|
|
|
|
|
|
if __name__ == "__main__":
|
|
test_meta_programming_matmul()
|
|
test_meta_programming_uncaptured_var()
|