[MLIR][Python] add f8E5M2 and tests for np_to_memref

[PhrygianGates]

add f8E5M2 and tests for np_to_memref

[github-actions]

Thank you for submitting a Pull Request (PR) to the LLVM Project!

This PR will be automatically labeled and the relevant teams will be notified.

If you wish to, you can add reviewers by using the "Reviewers" section on this page.

If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using @ followed by their GitHub username.

If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers.

If you have further questions, they may be answered by the LLVM GitHub User Guide.

You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums.

[llvmbot]

@llvm/pr-subscribers-mlir

Author: None (PhrygianGates)

Changes

add f8E5M2 and tests for np_to_memref

---

Full diff: https://github.com/llvm/llvm-project/pull/106028.diff

2 Files Affected:

• (modified) mlir/python/mlir/runtime/np_to_memref.py (+12)
• (modified) mlir/test/python/execution_engine.py (+39-1)

diff --git a/mlir/python/mlir/runtime/np_to_memref.py b/mlir/python/mlir/runtime/np_to_memref.py
index 882b2751921bfd..8cca1e7ad4a9eb 100644
--- a/mlir/python/mlir/runtime/np_to_memref.py
+++ b/mlir/python/mlir/runtime/np_to_memref.py
@@ -37,6 +37,11 @@ class BF16(ctypes.Structure):
 
     _fields_ = [("bf16", ctypes.c_int16)]
 
+class F8E5M2(ctypes.Structure):
+    """A ctype representation for MLIR's Float8E5M2."""
+
+    _fields_ = [("f8E5M2", ctypes.c_int8)]
+
 
 # https://stackoverflow.com/questions/26921836/correct-way-to-test-for-numpy-dtype
 def as_ctype(dtp):
@@ -49,6 +54,8 @@ def as_ctype(dtp):
         return F16
     if ml_dtypes is not None and dtp == ml_dtypes.bfloat16:
         return BF16
+    if ml_dtypes is not None and dtp == ml_dtypes.float8_e5m2:
+        return F8E5M2
     return np.ctypeslib.as_ctypes_type(dtp)
 
 
@@ -65,6 +72,11 @@ def to_numpy(array):
     ), f"bfloat16 requires the ml_dtypes package, please run:\n\npip install ml_dtypes\n"
     if array.dtype == BF16:
         return array.view("bfloat16")
+    assert not (
+        array.dtype == F8E5M2 and ml_dtypes is None
+    ), f"float8_e5m2 requires the ml_dtypes package, please run:\n\npip install ml_dtypes\n"
+    if array.dtype == F8E5M2:
+        return array.view("float8_e5m2")
     return array
 
 
diff --git a/mlir/test/python/execution_engine.py b/mlir/test/python/execution_engine.py
index 8125bf3fb8fc92..ae0c691e1cbcd9 100644
--- a/mlir/test/python/execution_engine.py
+++ b/mlir/test/python/execution_engine.py
@@ -5,7 +5,7 @@
 from mlir.passmanager import *
 from mlir.execution_engine import *
 from mlir.runtime import *
-from ml_dtypes import bfloat16
+from ml_dtypes import bfloat16, float8_e5m2
 
 
 # Log everything to stderr and flush so that we have a unified stream to match
@@ -560,6 +560,44 @@ def testBF16Memref():
 
 run(testBF16Memref)
 
+# Test f8E5M2 memrefs
+# CHECK-LABEL: TEST: testF8E5M2Memref
+def testF8E5M2Memref():
+    with Context():
+        module = Module.parse(
+            """
+    module  {
+      func.func @main(%arg0: memref<1xf8E5M2>,
+                      %arg1: memref<1xf8E5M2>) attributes { llvm.emit_c_interface } {
+        %0 = arith.constant 0 : index
+        %1 = memref.load %arg0[%0] : memref<1xf8E5M2>
+        memref.store %1, %arg1[%0] : memref<1xf8E5M2>
+        return
+      }
+    } """
+        )
+
+        arg1 = np.array([0.5]).astype(float8_e5m2)
+        arg2 = np.array([0.0]).astype(float8_e5m2)
+
+        arg1_memref_ptr = ctypes.pointer(
+            ctypes.pointer(get_ranked_memref_descriptor(arg1))
+        )
+        arg2_memref_ptr = ctypes.pointer(
+            ctypes.pointer(get_ranked_memref_descriptor(arg2))
+        )
+
+        execution_engine = ExecutionEngine(lowerToLLVM(module))
+        execution_engine.invoke("main", arg1_memref_ptr, arg2_memref_ptr)
+
+        # test to-numpy utility
+        # CHECK: [0.5]
+        npout = ranked_memref_to_numpy(arg2_memref_ptr[0])
+        log(npout)
+
+
+run(testF8E5M2Memref)
+
 
 #  Test addition of two 2d_memref
 # CHECK-LABEL: TEST: testDynamicMemrefAdd2D

[github-actions]

✅ With the latest revision this PR passed the Python code formatter.

[PhrygianGates]

Hi experts! Since I'm not familiar with the LLVM PR merge workflow, does this mean I still need approvals to merge?

[makslevental]

Hi experts! Since I'm not familiar with the LLVM PR merge workflow, does this mean I still need approvals to merge?

It just meant someone had to approve a CI job for you. I've done it.

[github-actions]

@PhrygianGates Congratulations on having your first Pull Request (PR) merged into the LLVM Project!

Your changes will be combined with recent changes from other authors, then tested by our build bots. If there is a problem with a build, you may receive a report in an email or a comment on this PR.

Please check whether problems have been caused by your change specifically, as the builds can include changes from many authors. It is not uncommon for your change to be included in a build that fails due to someone else's changes, or infrastructure issues.

How to do this, and the rest of the post-merge process, is covered in detail here.

If your change does cause a problem, it may be reverted, or you can revert it yourself. This is a normal part of LLVM development. You can fix your changes and open a new PR to merge them again.

If you don't get any reports, no action is required from you. Your changes are working as expected, well done!