update to 4.5 (#3228)

CHANGELOG.md

@@ -48,6 +48,7 @@
 * Use 64-bit adds for SM100 MMA descriptor offsets and reduce move instructions for improved code generation.
 * Add [example 95](https://github.com/NVIDIA/cutlass/tree/main/examples/95_blackwell_gemm_green_context) to support green context SM partition
   - Enables launching GEMM on stream with partial SM allocation.
+* Add [Snake](https://github.com/NVIDIA/cutlass/blob/main/test/unit/epilogue/thread/activation.cu#L409) activation functor for EVT.
 * Fix some kernel issues:
   - Fix l2_capacity=0 handling in Blackwell SM100/SM120 kernel templates
   - Fix CUTLASS clang build issues

README.md

@@ -89,6 +89,7 @@ To get started quickly - please refer :
 * Use 64-bit adds for SM100 MMA descriptor offsets and reduce move instructions for improved code generation.
 * Add [example 95](https://github.com/NVIDIA/cutlass/tree/main/examples/95_blackwell_gemm_green_context) to support green context SM partition
   - Enables launching GEMM on stream with partial SM allocation.
+* Add [Snake](https://github.com/NVIDIA/cutlass/blob/main/test/unit/epilogue/thread/activation.cu#L409) activation functor for EVT.
 * Fix some kernel issues:
   - Fix l2_capacity=0 handling in Blackwell SM100/SM120 kernel templates
   - Fix CUTLASS clang build issues
@@ -97,9 +98,7 @@ To get started quickly - please refer :
   - Remove `PipelineStorage` shadowing in SM100 complex epilogue
   - Fix build issue in SM90 epilogue fusion visitor TMA warpspecialized
 * Fix some profiler issues:
-  - Add missing reference kernels for blockwise GEMM profiler
-* Various improvements and fixes from the community and CUTLASS team. Thanks to everyone who submitted PRs!
-* Optimal code generation with CUDA toolkit versions 13.2.
+  - Add missing reference kernels for blockwise GEMM profiler.
 
 Note: CUTLASS 4.x builds are known to be down on Windows platforms for all CUDA toolkits.
 CUTLASS team is working on a fix.

include/cutlass/version.h

@@ -35,8 +35,8 @@
 #include <string>
 
 #define CUTLASS_MAJOR 4
-#define CUTLASS_MINOR 4
-#define CUTLASS_PATCH 2
+#define CUTLASS_MINOR 5
+#define CUTLASS_PATCH 0
 
 #ifdef CUTLASS_VERSIONS_GENERATED
 #include "cutlass/version_extended.h"

python/cutlass_cppgen/__init__.py

@@ -133,7 +133,7 @@ def get_option_registry():
         this._option_registry = OptionRegistry(device_cc())
     return this._option_registry
 
-this.__version__ = '4.4.2'
+this.__version__ = '4.5.0'
 
 from cutlass_cppgen.backend import create_memory_pool
 from cutlass_cppgen.emit.pytorch import pytorch

python/setup_cutlass.py

@@ -51,7 +51,7 @@ setup_pycute.perform_setup()
 
 setup(
     name='cutlass_cppgen',
-    version='4.4.2',
+    version='4.5.0',
     description='CUTLASS Pythonic Interface',
     package_dir={'': '.'},
     packages=[

python/setup_library.py

@@ -36,7 +36,7 @@ from setuptools import setup
 def perform_setup():
     setup(
         name='cutlass_library',
-        version='4.4.2',
+        version='4.5.0',
         description='CUTLASS library generation scripts',
         packages=['cutlass_library']
     )

python/setup_pycute.py

@@ -36,7 +36,7 @@ from setuptools import setup
 def perform_setup():
     setup(
         name='pycute',
-        version='4.4.2',
+        version='4.5.0',
         description='Python implementation of CuTe',
         packages=['pycute'],
     )