From 02b8e504f0a645230b850d6eb1ae95ed03ef2451 Mon Sep 17 00:00:00 2001 From: Nilesh M Negi Date: Wed, 26 Feb 2025 18:43:37 -0600 Subject: [PATCH] [BUILD] Fix generate.py for gfx950 (#1575) * [BUILD] Fix generate.py for gfx950 Signed-off-by: nilenegi * [BUILD] Cleaner way to check for gfx targets Signed-off-by: nilenegi --------- Signed-off-by: nilenegi --- src/device/generate.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/device/generate.py b/src/device/generate.py index c03abaf422..9f540eed31 100755 --- a/src/device/generate.py +++ b/src/device/generate.py @@ -164,7 +164,7 @@ def calc_unroll_for_local_arch(): # Homogeneous system is required to build for only 1 varient of unroll factor if len(gfx_targets) == 1: gfx_name, cu_count = gfx_targets[0] - if "gfx908" == gfx_name or ("gfx94" in gfx_name and cu_count > 80): + if "gfx908" == gfx_name or (gfx_name in ["gfx942", "gfx950"] and cu_count > 80): return 2 else: return 4 @@ -516,4 +516,4 @@ if is_msccl_kernels: out( "MSCCL_IMPL_KERNEL_ENTRY_FUNC_DEVREDOP_TYPE({redop}, {ty_cxx}, false);\n" .format(redop=redop, ty_cxx=ty_to_cxx[ty]) - ) \ No newline at end of file + )