From f5acb6b987609cc6c1d859120cd327c57c344803 Mon Sep 17 00:00:00 2001
From: venkataram-nv <vedavamadath@nvidia.com>
Date: Thu, 27 Jun 2024 16:09:33 -0700
Subject: Extend `countbits` intrinsic for vector types (#4499)

* Extend `countbits` intrinsic for vector types

This commit implements the overloading function for `countbits` function.
Because HLSL has following overloadings,
```
uint count_bits(uint value);
uint2 count_bits(uint2 value);
uint3 count_bits(uint3 value);
uint4 count_bits(uint4 value);
```
https://learn.microsoft.com/en-us/windows/win32/direct3dhlsl/countbits
---
 source/slang/glsl.meta.slang | 10 +---------
 source/slang/hlsl.meta.slang | 22 ++++++++++++++++++++++
 2 files changed, 23 insertions(+), 9 deletions(-)

(limited to 'source')

diff --git a/source/slang/glsl.meta.slang b/source/slang/glsl.meta.slang
index 523d9af90..f18e07fbd 100644
--- a/source/slang/glsl.meta.slang
+++ b/source/slang/glsl.meta.slang
@@ -1402,15 +1402,7 @@ __generic<let N:int>
 [require(cpp_cuda_glsl_hlsl_spirv, shader5_sm_5_0)]
 public vector<uint,N> bitCount(vector<uint,N> value)
 {
-    __target_switch
-    {
-    case glsl: __intrinsic_asm "bitCount";
-    case spirv: return spirv_asm {
-        result:$$vector<uint,N> = OpBitCount $value
-    };
-    default:
-        VECTOR_MAP_UNARY(uint, N, countbits, value);
-    }
+    return countbits(value);
 }
 
 [__readNone] 
diff --git a/source/slang/hlsl.meta.slang b/source/slang/hlsl.meta.slang
index 909bebb0d..0a440f986 100644
--- a/source/slang/hlsl.meta.slang
+++ b/source/slang/hlsl.meta.slang
@@ -6917,6 +6917,7 @@ vector<T,N> cospi(vector<T,N> x)
 
 // Population count
 [__readNone]
+[ForceInline]
 [require(cpp_cuda_glsl_hlsl_metal_spirv, shader5_sm_5_0)]
 uint countbits(uint value)
 {
@@ -6936,6 +6937,27 @@ uint countbits(uint value)
     }
 }
 
+__generic <let N : int>
+[__readNone]
+[ForceInline]
+[require(cpp_cuda_glsl_hlsl_metal_spirv, shader5_sm_5_0)]
+vector<uint, N> countbits(vector<uint, N> value)
+{
+    __target_switch
+    {
+    case hlsl:
+        __intrinsic_asm "countbits";
+    case glsl:
+        __intrinsic_asm "bitCount";
+    case metal:
+        __intrinsic_asm "popcount";
+    case spirv:
+        return spirv_asm {OpBitCount $$vector<uint, N> result $value};
+    default:
+        VECTOR_MAP_UNARY(uint, N, countbits, value);
+    }
+}
+
 // Cross product
 // TODO: SPIRV does not support integer vectors.
 __generic<T : __BuiltinFloatingPointType>
-- 
cgit v1.2.3