From a0676b6ead197527947724e5959ca747d8fde058 Mon Sep 17 00:00:00 2001
From: manish <trmanish@gmail.com>
Date: Wed, 9 Jul 2025 16:56:40 -0700
Subject: [PATCH] [Packaging] Shrink wheel ~35 % via nvcc --compress-mode=size

Signed-off-by: manish <trmanish@gmail.com>
---
 CMakeLists.txt | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/CMakeLists.txt b/CMakeLists.txt
index 770b4ba30..ee03aa751 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -133,6 +133,11 @@ if(BUILD_CUDA)
 
     string(APPEND CMAKE_CUDA_FLAGS " --use_fast_math")
 
+    # Compress embedded fatbins when we build with CUDA ≥ 12.4
+   if (CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL "12.4")
+      string(APPEND CMAKE_CUDA_FLAGS " --compress-mode=size")
+   endif()
+
     if(PTXAS_VERBOSE)
         # Verbose? Outputs register usage information, and other things...
         string(APPEND CMAKE_CUDA_FLAGS " -Xptxas=-v")