From bbd54d3dfbcabc959f0ff98bb94be1059a3397f8 Mon Sep 17 00:00:00 2001
From: ClementLinCF <162283536+ClementLinCF@users.noreply.github.com>
Date: Mon, 13 Jan 2025 12:50:32 +0800
Subject: [PATCH] [CK_TILE] Adjust kBlockSize of reduce example for better perf
 (#1779)

* Observed a 2x perf improvement with kBlockSize = 256
* Using 512 threads may lead to redundant computations

[ROCm/composable_kernel commit: 0b8f117f1ae765f0e490368de7f0c5d7591b17b6]
---
 example/ck_tile/05_reduce/reduce.cpp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/example/ck_tile/05_reduce/reduce.cpp b/example/ck_tile/05_reduce/reduce.cpp
index 005541dc62..602661f779 100644
--- a/example/ck_tile/05_reduce/reduce.cpp
+++ b/example/ck_tile/05_reduce/reduce.cpp
@@ -52,7 +52,7 @@ bool run(const ck_tile::ArgParser& arg_parser)
     // using WarpTile   = ck_tile::sequence<1, 512>;
     // using Vector = ck_tile::sequence<1, 8>;
 
-    constexpr ck_tile::index_t kBlockSize  = 512;
+    constexpr ck_tile::index_t kBlockSize  = 256;
     constexpr ck_tile::index_t kBlockPerCu = 1;
     ck_tile::index_t kGridSize             = (m / BlockTile::at(ck_tile::number<0>{}));
     std::cout << "grid size " << kGridSize << std::endl;