You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by "Lunderberg (via GitHub)" <gi...@apache.org> on 2023/03/13 13:43:54 UTC
[GitHub] [tvm] Lunderberg commented on a diff in pull request #14280: [TensorIR][Transform] Enable warp shuffling for `LowerWarpMemory`
Lunderberg commented on code in PR #14280:
URL: https://github.com/apache/tvm/pull/14280#discussion_r1133951480
##########
tests/python/unittest/test_tir_transform_lower_warp_memory.py:
##########
@@ -347,5 +348,105 @@ def test_lower_warp_memory_divide_by_factor():
tvm.tir.transform.LowerWarpMemory()(mod)["f_kernel0"]
+@T.prim_func
+def func(a: T.handle, b: T.handle) -> None:
+ A = T.match_buffer(a, [32], "float32")
+ B = T.match_buffer(b, [32], "float32")
+ for i in range(32):
+ with T.block("warp_shuffle"):
+ vi = T.axis.spatial(32, i)
+ B[vi] = A[(vi % 4) * 8 + vi // 4] + T.float32(1)
+
+
+def test_warp_shuffle_transform():
+ @tvm.script.ir_module
+ class Before:
+ @T.prim_func
+ def main(A: T.handle("float32", "global"), B: T.handle("float32", "global")):
+ blockIdx_x = T.env_thread("blockIdx.x")
+ threadIdx_x = T.env_thread("threadIdx.x")
+ T.func_attr(
Review Comment:
It looks like the test case only requires the `"target"` attribute, and only requires `"kind"` and `"thread_warp_size"` within that. Can we remove the extra attributes from the unit test?
##########
tests/python/unittest/test_tir_transform_lower_warp_memory.py:
##########
@@ -347,5 +348,105 @@ def test_lower_warp_memory_divide_by_factor():
tvm.tir.transform.LowerWarpMemory()(mod)["f_kernel0"]
+@T.prim_func
+def func(a: T.handle, b: T.handle) -> None:
+ A = T.match_buffer(a, [32], "float32")
+ B = T.match_buffer(b, [32], "float32")
+ for i in range(32):
+ with T.block("warp_shuffle"):
+ vi = T.axis.spatial(32, i)
+ B[vi] = A[(vi % 4) * 8 + vi // 4] + T.float32(1)
+
+
+def test_warp_shuffle_transform():
+ @tvm.script.ir_module
+ class Before:
+ @T.prim_func
+ def main(A: T.handle("float32", "global"), B: T.handle("float32", "global")):
+ blockIdx_x = T.env_thread("blockIdx.x")
+ threadIdx_x = T.env_thread("threadIdx.x")
+ T.func_attr(
+ {
+ "calling_conv": 2,
+ "global_symbol": "main",
+ "target": T.target(
+ {
+ "host": {"keys": ["cpu"], "kind": "llvm", "tag": ""},
+ "keys": ["cuda", "gpu"],
+ "kind": "cuda",
+ "max_num_threads": 1024,
+ "tag": "",
+ "thread_warp_size": 32,
+ }
+ ),
+ "tir.device_thread_axis": [
+ T.iter_var(blockIdx_x, [0, 1], "ThreadIndex", "blockIdx.x"),
+ T.iter_var(threadIdx_x, [0, 32], "ThreadIndex", "threadIdx.x"),
+ ],
+ "tir.is_global_func": 1,
+ "tir.noalias": 1,
+ }
+ )
+ T.launch_thread(blockIdx_x, 1)
+ A_warp = T.allocate([32], "float32", "warp")
+ B_warp = T.allocate([32], "float32", "warp")
+ T.launch_thread(threadIdx_x, 32)
+ A_warp_1 = T.Buffer((32,), data=A_warp, scope="warp")
+ A_1 = T.Buffer((32,), data=A)
Review Comment:
Instead of having a separate `A: T.handle` and `A_1: T.Buffer`, the buffer could be declared as a parameter `A_1: T.Buffer(32)`. It does result in slightly different TIR, as it follows the style from before `MakePackedAPI` is applied, but for a unit test would help to emphasize the change being tested.
##########
tests/python/unittest/test_tir_transform_lower_warp_memory.py:
##########
@@ -347,5 +348,105 @@ def test_lower_warp_memory_divide_by_factor():
tvm.tir.transform.LowerWarpMemory()(mod)["f_kernel0"]
+@T.prim_func
+def func(a: T.handle, b: T.handle) -> None:
+ A = T.match_buffer(a, [32], "float32")
+ B = T.match_buffer(b, [32], "float32")
+ for i in range(32):
+ with T.block("warp_shuffle"):
+ vi = T.axis.spatial(32, i)
+ B[vi] = A[(vi % 4) * 8 + vi // 4] + T.float32(1)
+
+
+def test_warp_shuffle_transform():
+ @tvm.script.ir_module
+ class Before:
+ @T.prim_func
+ def main(A: T.handle("float32", "global"), B: T.handle("float32", "global")):
+ blockIdx_x = T.env_thread("blockIdx.x")
+ threadIdx_x = T.env_thread("threadIdx.x")
+ T.func_attr(
+ {
+ "calling_conv": 2,
+ "global_symbol": "main",
+ "target": T.target(
+ {
+ "host": {"keys": ["cpu"], "kind": "llvm", "tag": ""},
+ "keys": ["cuda", "gpu"],
+ "kind": "cuda",
+ "max_num_threads": 1024,
+ "tag": "",
+ "thread_warp_size": 32,
+ }
+ ),
+ "tir.device_thread_axis": [
+ T.iter_var(blockIdx_x, [0, 1], "ThreadIndex", "blockIdx.x"),
+ T.iter_var(threadIdx_x, [0, 32], "ThreadIndex", "threadIdx.x"),
+ ],
+ "tir.is_global_func": 1,
+ "tir.noalias": 1,
+ }
+ )
+ T.launch_thread(blockIdx_x, 1)
+ A_warp = T.allocate([32], "float32", "warp")
+ B_warp = T.allocate([32], "float32", "warp")
+ T.launch_thread(threadIdx_x, 32)
+ A_warp_1 = T.Buffer((32,), data=A_warp, scope="warp")
+ A_1 = T.Buffer((32,), data=A)
+ A_warp_1[threadIdx_x] = A_1[threadIdx_x]
+ B_warp_1 = T.Buffer((32,), data=B_warp, scope="warp")
+ T.tvm_storage_sync("warp")
+ B_warp_1[threadIdx_x] = A_warp_1[threadIdx_x % 4 * 8 + threadIdx_x // 4] + T.float32(1)
Review Comment:
Could we add a comment here, indicating that this line is the one that should be updated correctly?
##########
tests/python/unittest/test_tir_transform_lower_warp_memory.py:
##########
@@ -347,5 +348,105 @@ def test_lower_warp_memory_divide_by_factor():
tvm.tir.transform.LowerWarpMemory()(mod)["f_kernel0"]
+@T.prim_func
+def func(a: T.handle, b: T.handle) -> None:
+ A = T.match_buffer(a, [32], "float32")
+ B = T.match_buffer(b, [32], "float32")
+ for i in range(32):
+ with T.block("warp_shuffle"):
+ vi = T.axis.spatial(32, i)
+ B[vi] = A[(vi % 4) * 8 + vi // 4] + T.float32(1)
+
+
+def test_warp_shuffle_transform():
Review Comment:
The test looks reasonable as-is, though there's also a `tvm.testing.CompareBeforeAfter` that you could use to further reduce the boilerplate.
```python
class TestWarpShuffleTransform(tvm.testing.CompareBeforeAfter):
transform = tvm.tir.transform.LowerWarpMemory()
def before(A: T.handle("float32", "global"), B: T.handle("float32", "global")):
...
def expected(A: T.handle("float32", "global"), B: T.handle("float32", "global")):
...
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org