Merge branch 'natevm-bitfield-instructions' into master

shader-slang · Dec 8, 2024 · 7b220e3 · 7b220e3
2 parents 0d5636c + b5e3ca5
commit 7b220e3
Show file tree

Hide file tree

Showing 14 changed files with 706 additions and 175 deletions.
diff --git a/source/slang/core.meta.slang b/source/slang/core.meta.slang
@@ -2716,6 +2716,19 @@ T reinterpret(U value);
 
 /// Use an otherwise unused value
 /// This can be used to silence the warning about returning before initializing an out paramter.
+// Bitfield extract / insert
+__generic<T>
+[__readNone]
+[__unsafeForceInlineEarly]
+__intrinsic_op($(kIROp_BitfieldInsert))
+T bitfieldInsert(T base, T insert, int offset, int bits);
+
+__generic<T>
+[__readNone]
+[__unsafeForceInlineEarly]
+__intrinsic_op($(kIROp_BitfieldExtract))
+T bitfieldExtract(T value, int offset, int bits);
+
 __generic<T>
 [__readNone]
 [ForceInline]

diff --git a/source/slang/glsl.meta.slang b/source/slang/glsl.meta.slang
@@ -1150,168 +1150,6 @@ public void imulExtended(highp vector<int,N> x, highp vector<int,N> y, out highp
     }
 }
 
-[__readNone]
-[ForceInline]
-[require(cpp_cuda_glsl_hlsl_spirv, GLSL_400)]
-public int bitfieldExtract(int value, int offset, int bits)
-{
-    __target_switch
-    {
-    case glsl: __intrinsic_asm "bitfieldExtract";
-    case spirv: return spirv_asm {
-        result:$$int = OpBitFieldSExtract $value $offset $bits
-    };
-    default:
-        return int(uint(value >> offset) & ((1u << bits) - 1));
-    }
-}
-
-__generic<let N:int>
-[__readNone]
-[ForceInline]
-[require(cpp_cuda_glsl_hlsl_spirv, GLSL_400)]
-public vector<int,N> bitfieldExtract(vector<int,N> value, int offset, int bits)
-{
-    __target_switch
-    {
-    case glsl: __intrinsic_asm "bitfieldExtract";
-    case spirv: return spirv_asm {
-        result:$$vector<int,N> = OpBitFieldSExtract $value $offset $bits
-    };
-    default:
-        vector<int,N> result;
-        [ForceUnroll]
-        for (int i = 0; i < N; ++i)
-        {
-            result[i] = bitfieldExtract(value[i], offset, bits);
-        }
-        return result;
-    }
-}
-
-[__readNone]
-[ForceInline]
-[require(cpp_cuda_glsl_hlsl_spirv, GLSL_400)]
-public uint bitfieldExtract(uint value, int offset, int bits)
-{
-    __target_switch
-    {
-    case glsl: __intrinsic_asm "bitfieldExtract";
-    case spirv: return spirv_asm {
-        result:$$uint = OpBitFieldUExtract $value $offset $bits
-    };
-    default:
-        return (value >> offset) & ((1u << bits) - 1);
-    }
-}
-
-__generic<let N:int>
-[__readNone]
-[ForceInline]
-[require(cpp_cuda_glsl_hlsl_spirv, GLSL_400)]
-public vector<uint,N> bitfieldExtract(vector<uint,N> value, int offset, int bits)
-{
-    __target_switch
-    {
-    case glsl: __intrinsic_asm "bitfieldExtract";
-    case spirv: return spirv_asm {
-        result:$$vector<uint,N> = OpBitFieldUExtract $value $offset $bits
-    };
-    default:
-        vector<uint,N> result;
-        [ForceUnroll]
-        for (int i = 0; i < N; ++i)
-        {
-            result[i] = bitfieldExtract(value[i], offset, bits);
-        }
-        return result;
-    }
-}
-
-[__readNone]
-[ForceInline]
-[require(cpp_cuda_glsl_hlsl_spirv, GLSL_400)]
-public uint bitfieldInsert(uint base, uint insert, int offset, int bits)
-{
-    __target_switch
-    {
-    case glsl: __intrinsic_asm "bitfieldInsert";
-    case spirv: return spirv_asm {
-        result:$$uint = OpBitFieldInsert $base $insert $offset $bits
-    };
-    default:
-        uint clearMask = ~(((1u << bits) - 1u) << offset);
-        uint clearedBase = base & clearMask;
-        uint maskedInsert = (insert & ((1u << bits) - 1u)) << offset;
-        return clearedBase | maskedInsert;
-    }
-}
-
-__generic<let N:int>
-[__readNone]
-[ForceInline]
-[require(cpp_cuda_glsl_hlsl_spirv, GLSL_400)]
-public vector<uint,N> bitfieldInsert(vector<uint,N> base, vector<uint,N> insert, int offset, int bits)
-{
-    __target_switch
-    {
-    case glsl: __intrinsic_asm "bitfieldInsert";
-    case spirv: return spirv_asm {
-        result:$$vector<uint,N> = OpBitFieldInsert $base $insert $offset $bits
-    };
-    default:
-        vector<uint,N> result;
-        [ForceUnroll]
-        for (int i = 0; i < N; ++i)
-        {
-            result[i] = bitfieldInsert(base[i], insert[i], offset, bits);
-        }
-        return result;
-    }
-}
-
-[__readNone]
-[ForceInline]
-[require(cpp_cuda_glsl_hlsl_spirv, GLSL_400)]
-public int bitfieldInsert(int base, int insert, int offset, int bits)
-{
-    __target_switch
-    {
-    case glsl: __intrinsic_asm "bitfieldInsert";
-    case spirv: return spirv_asm {
-        result:$$int = OpBitFieldInsert $base $insert $offset $bits
-    };
-    default:
-        uint clearMask = ~(((1u << bits) - 1u) << offset);
-        uint clearedBase = base & clearMask;
-        uint maskedInsert = (insert & ((1u << bits) - 1u)) << offset;
-        return clearedBase | maskedInsert;
-    }
-}
-
-__generic<let N:int>
-[__readNone]
-[ForceInline]
-[require(cpp_cuda_glsl_hlsl_spirv, GLSL_400)]
-public vector<int,N> bitfieldInsert(vector<int,N> base, vector<int,N> insert, int offset, int bits)
-{
-    __target_switch
-    {
-    case glsl: __intrinsic_asm "bitfieldInsert";
-    case spirv: return spirv_asm {
-        result:$$vector<int,N> = OpBitFieldInsert $base $insert $offset $bits
-    };
-    default:
-        vector<int,N> result;
-        [ForceUnroll]
-        for (int i = 0; i < N; ++i)
-        {
-            result[i] = bitfieldInsert(base[i], insert[i], offset, bits);
-        }
-        return result;
-    }
-}
-
 [__readNone]
 [ForceInline]
 [require(cpp_cuda_glsl_hlsl_spirv, GLSL_400)]