From ea7690558bca71ce3a9453adff4e0135352a352f Mon Sep 17 00:00:00 2001
From: jsmall-nvidia <jsmall@nvidia.com>
Date: Mon, 30 Mar 2020 19:23:09 -0400
Subject: CUDA version handling (#1301)

* render feature for CUDA compute model.

* Use SemanticVersion type.

* Enable CUDA wave tests that require CUDA SM 7.0.
Provide mechanism for DownstreamCompiler to specify version numbers.

* Enabled wave-equality.slang

* Make CUDA SM version major version not just a single digit.

* Fix assert.

* DownstreamCompiler::Version -> CapabilityVersion
---
 source/slang/hlsl.meta.slang | 6 ++++++
 1 file changed, 6 insertions(+)

(limited to 'source/slang/hlsl.meta.slang')

diff --git a/source/slang/hlsl.meta.slang b/source/slang/hlsl.meta.slang
index e2e745773..f82f7b5f4 100644
--- a/source/slang/hlsl.meta.slang
+++ b/source/slang/hlsl.meta.slang
@@ -2616,15 +2616,18 @@ __generic<T : __BuiltinType>
 __glsl_extension(GL_KHR_shader_subgroup_vote)
 __spirv_version(1.3)
 __target_intrinsic(glsl, "subgroupAllEqual($0)")
+__cuda_sm_version(7.0)
 __target_intrinsic(cuda, "_waveAllEqual($0)")
 bool WaveActiveAllEqual(T value);
 __generic<T : __BuiltinType, let N : int> 
 __glsl_extension(GL_KHR_shader_subgroup_vote)
 __spirv_version(1.3)
 __target_intrinsic(glsl, "subgroupAllEqual($0)")
+__cuda_sm_version(7.0)
 __target_intrinsic(cuda, "_waveAllEqualMultiple($0)")
 bool WaveActiveAllEqual(vector<T,N> value);
 __generic<T : __BuiltinType, let N : int, let M : int>
+__cuda_sm_version(7.0)
 __target_intrinsic(cuda, "_waveAllEqualMultiple($0)")
 bool WaveActiveAllEqual(matrix<T,N,M> value);
 
@@ -2796,14 +2799,17 @@ uint WavePrefixCountBits(bool value);
 
 __generic<T : __BuiltinType>
 __target_intrinsic(hlsl)
+__cuda_sm_version(7.0)
 __target_intrinsic(cuda, "_waveMatchScalar($0)")
 uint4 WaveMatch(T value);
 __generic<T : __BuiltinType, let N : int>
 __target_intrinsic(hlsl)
+__cuda_sm_version(7.0)
 __target_intrinsic(cuda, "_waveMatchMultiple($0)")
 uint4 WaveMatch(vector<T,N> value);
 __generic<T : __BuiltinType, let N : int, let M : int>
 __target_intrinsic(hlsl)
+__cuda_sm_version(7.0)
 __target_intrinsic(cuda, "_waveMatchMultiple($0)")
 uint4 WaveMatch(matrix<T,N,M> value);
 
-- 
cgit v1.2.3