dawn-cmake/test/tint/buffer/uniform/dynamic_index/read.wgsl.expected.fxc.hlsl

138 lines
6.8 KiB
HLSL
Raw Normal View History

cbuffer cbuffer_ub : register(b0, space0) {
uint4 ub[272];
};
struct tint_symbol_1 {
uint idx : SV_GroupIndex;
};
float2x2 tint_symbol_14(uint4 buffer[272], uint offset) {
transform: Fixes for DecomposeMemoryAccess CloneContext::Replace(T* what, T* with) is bug-prone, as complex transforms may want to clone `what` multiple times, or not at all. In both cases, this will likely result in an ICE as either the replacement will be reachable multiple times, or not at all. The CTS test: webgpu:shader,execution,robust_access:linear_memory:storageClass="storage";storageMode="read_write";access="read";atomic=true;baseType="i32" Was triggering this brokenness with DecomposeMemoryAccess's use of CloneContext::Replace(T*, T*). Switch the usage of CloneContext::Replace(T*, T*) to the new function form. As std::function is copyable, it cannot hold a captured std::unique_ptr. This prevented the Replace() lambdas from capturing the necessary `BufferAccess` data, as this held a `std::unique_ptr<Offset>`. To fix this, use a `BlockAllocator` for Offsets, and use raw pointers instead. Because the function passed to Replace() is called just before the node is cloned, insertion of new functions will occur just before the currently evaluated module-scope entity. This allows us to remove the "insert_after" arguments to LoadFunc(), StoreFunc(), and AtomicFunc(). We can also kill the icky InsertGlobal() and TypeDeclOf() helpers. Bug: tint:993 Change-Id: I60972bc13a2fa819a163ee2671f61e82d0e68d2a Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58222 Kokoro: Kokoro <noreply+kokoro@google.com> Reviewed-by: James Price <jrprice@google.com> Commit-Queue: Ben Clayton <bclayton@google.com>
2021-07-15 20:29:09 +00:00
const uint scalar_offset = ((offset + 0u)) / 4;
uint4 ubo_load = buffer[scalar_offset / 4];
const uint scalar_offset_1 = ((offset + 8u)) / 4;
uint4 ubo_load_1 = buffer[scalar_offset_1 / 4];
return float2x2(asfloat(((scalar_offset & 2) ? ubo_load.zw : ubo_load.xy)), asfloat(((scalar_offset_1 & 2) ? ubo_load_1.zw : ubo_load_1.xy)));
transform: Fixes for DecomposeMemoryAccess CloneContext::Replace(T* what, T* with) is bug-prone, as complex transforms may want to clone `what` multiple times, or not at all. In both cases, this will likely result in an ICE as either the replacement will be reachable multiple times, or not at all. The CTS test: webgpu:shader,execution,robust_access:linear_memory:storageClass="storage";storageMode="read_write";access="read";atomic=true;baseType="i32" Was triggering this brokenness with DecomposeMemoryAccess's use of CloneContext::Replace(T*, T*). Switch the usage of CloneContext::Replace(T*, T*) to the new function form. As std::function is copyable, it cannot hold a captured std::unique_ptr. This prevented the Replace() lambdas from capturing the necessary `BufferAccess` data, as this held a `std::unique_ptr<Offset>`. To fix this, use a `BlockAllocator` for Offsets, and use raw pointers instead. Because the function passed to Replace() is called just before the node is cloned, insertion of new functions will occur just before the currently evaluated module-scope entity. This allows us to remove the "insert_after" arguments to LoadFunc(), StoreFunc(), and AtomicFunc(). We can also kill the icky InsertGlobal() and TypeDeclOf() helpers. Bug: tint:993 Change-Id: I60972bc13a2fa819a163ee2671f61e82d0e68d2a Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58222 Kokoro: Kokoro <noreply+kokoro@google.com> Reviewed-by: James Price <jrprice@google.com> Commit-Queue: Ben Clayton <bclayton@google.com>
2021-07-15 20:29:09 +00:00
}
float2x3 tint_symbol_15(uint4 buffer[272], uint offset) {
const uint scalar_offset_2 = ((offset + 0u)) / 4;
const uint scalar_offset_3 = ((offset + 16u)) / 4;
return float2x3(asfloat(buffer[scalar_offset_2 / 4].xyz), asfloat(buffer[scalar_offset_3 / 4].xyz));
}
float2x4 tint_symbol_16(uint4 buffer[272], uint offset) {
const uint scalar_offset_4 = ((offset + 0u)) / 4;
const uint scalar_offset_5 = ((offset + 16u)) / 4;
return float2x4(asfloat(buffer[scalar_offset_4 / 4]), asfloat(buffer[scalar_offset_5 / 4]));
}
float3x2 tint_symbol_17(uint4 buffer[272], uint offset) {
const uint scalar_offset_6 = ((offset + 0u)) / 4;
uint4 ubo_load_2 = buffer[scalar_offset_6 / 4];
const uint scalar_offset_7 = ((offset + 8u)) / 4;
uint4 ubo_load_3 = buffer[scalar_offset_7 / 4];
const uint scalar_offset_8 = ((offset + 16u)) / 4;
uint4 ubo_load_4 = buffer[scalar_offset_8 / 4];
return float3x2(asfloat(((scalar_offset_6 & 2) ? ubo_load_2.zw : ubo_load_2.xy)), asfloat(((scalar_offset_7 & 2) ? ubo_load_3.zw : ubo_load_3.xy)), asfloat(((scalar_offset_8 & 2) ? ubo_load_4.zw : ubo_load_4.xy)));
}
float3x3 tint_symbol_18(uint4 buffer[272], uint offset) {
const uint scalar_offset_9 = ((offset + 0u)) / 4;
const uint scalar_offset_10 = ((offset + 16u)) / 4;
const uint scalar_offset_11 = ((offset + 32u)) / 4;
return float3x3(asfloat(buffer[scalar_offset_9 / 4].xyz), asfloat(buffer[scalar_offset_10 / 4].xyz), asfloat(buffer[scalar_offset_11 / 4].xyz));
}
float3x4 tint_symbol_19(uint4 buffer[272], uint offset) {
const uint scalar_offset_12 = ((offset + 0u)) / 4;
const uint scalar_offset_13 = ((offset + 16u)) / 4;
const uint scalar_offset_14 = ((offset + 32u)) / 4;
return float3x4(asfloat(buffer[scalar_offset_12 / 4]), asfloat(buffer[scalar_offset_13 / 4]), asfloat(buffer[scalar_offset_14 / 4]));
}
float4x2 tint_symbol_20(uint4 buffer[272], uint offset) {
const uint scalar_offset_15 = ((offset + 0u)) / 4;
uint4 ubo_load_5 = buffer[scalar_offset_15 / 4];
const uint scalar_offset_16 = ((offset + 8u)) / 4;
uint4 ubo_load_6 = buffer[scalar_offset_16 / 4];
const uint scalar_offset_17 = ((offset + 16u)) / 4;
uint4 ubo_load_7 = buffer[scalar_offset_17 / 4];
const uint scalar_offset_18 = ((offset + 24u)) / 4;
uint4 ubo_load_8 = buffer[scalar_offset_18 / 4];
return float4x2(asfloat(((scalar_offset_15 & 2) ? ubo_load_5.zw : ubo_load_5.xy)), asfloat(((scalar_offset_16 & 2) ? ubo_load_6.zw : ubo_load_6.xy)), asfloat(((scalar_offset_17 & 2) ? ubo_load_7.zw : ubo_load_7.xy)), asfloat(((scalar_offset_18 & 2) ? ubo_load_8.zw : ubo_load_8.xy)));
}
float4x3 tint_symbol_21(uint4 buffer[272], uint offset) {
const uint scalar_offset_19 = ((offset + 0u)) / 4;
const uint scalar_offset_20 = ((offset + 16u)) / 4;
const uint scalar_offset_21 = ((offset + 32u)) / 4;
const uint scalar_offset_22 = ((offset + 48u)) / 4;
return float4x3(asfloat(buffer[scalar_offset_19 / 4].xyz), asfloat(buffer[scalar_offset_20 / 4].xyz), asfloat(buffer[scalar_offset_21 / 4].xyz), asfloat(buffer[scalar_offset_22 / 4].xyz));
}
float4x4 tint_symbol_22(uint4 buffer[272], uint offset) {
const uint scalar_offset_23 = ((offset + 0u)) / 4;
const uint scalar_offset_24 = ((offset + 16u)) / 4;
const uint scalar_offset_25 = ((offset + 32u)) / 4;
const uint scalar_offset_26 = ((offset + 48u)) / 4;
return float4x4(asfloat(buffer[scalar_offset_23 / 4]), asfloat(buffer[scalar_offset_24 / 4]), asfloat(buffer[scalar_offset_25 / 4]), asfloat(buffer[scalar_offset_26 / 4]));
}
typedef float3 tint_symbol_23_ret[2];
tint_symbol_23_ret tint_symbol_23(uint4 buffer[272], uint offset) {
float3 arr_1[2] = (float3[2])0;
{
for(uint i = 0u; (i < 2u); i = (i + 1u)) {
const uint scalar_offset_27 = ((offset + (i * 16u))) / 4;
arr_1[i] = asfloat(buffer[scalar_offset_27 / 4].xyz);
}
}
return arr_1;
}
void main_inner(uint idx) {
const uint scalar_offset_28 = ((544u * idx)) / 4;
const float scalar_f32 = asfloat(ub[scalar_offset_28 / 4][scalar_offset_28 % 4]);
const uint scalar_offset_29 = (((544u * idx) + 4u)) / 4;
const int scalar_i32 = asint(ub[scalar_offset_29 / 4][scalar_offset_29 % 4]);
const uint scalar_offset_30 = (((544u * idx) + 8u)) / 4;
const uint scalar_u32 = ub[scalar_offset_30 / 4][scalar_offset_30 % 4];
const uint scalar_offset_31 = (((544u * idx) + 16u)) / 4;
uint4 ubo_load_9 = ub[scalar_offset_31 / 4];
const float2 vec2_f32 = asfloat(((scalar_offset_31 & 2) ? ubo_load_9.zw : ubo_load_9.xy));
const uint scalar_offset_32 = (((544u * idx) + 24u)) / 4;
uint4 ubo_load_10 = ub[scalar_offset_32 / 4];
const int2 vec2_i32 = asint(((scalar_offset_32 & 2) ? ubo_load_10.zw : ubo_load_10.xy));
const uint scalar_offset_33 = (((544u * idx) + 32u)) / 4;
uint4 ubo_load_11 = ub[scalar_offset_33 / 4];
const uint2 vec2_u32 = ((scalar_offset_33 & 2) ? ubo_load_11.zw : ubo_load_11.xy);
const uint scalar_offset_34 = (((544u * idx) + 48u)) / 4;
const float3 vec3_f32 = asfloat(ub[scalar_offset_34 / 4].xyz);
const uint scalar_offset_35 = (((544u * idx) + 64u)) / 4;
const int3 vec3_i32 = asint(ub[scalar_offset_35 / 4].xyz);
const uint scalar_offset_36 = (((544u * idx) + 80u)) / 4;
const uint3 vec3_u32 = ub[scalar_offset_36 / 4].xyz;
const uint scalar_offset_37 = (((544u * idx) + 96u)) / 4;
const float4 vec4_f32 = asfloat(ub[scalar_offset_37 / 4]);
const uint scalar_offset_38 = (((544u * idx) + 112u)) / 4;
const int4 vec4_i32 = asint(ub[scalar_offset_38 / 4]);
const uint scalar_offset_39 = (((544u * idx) + 128u)) / 4;
const uint4 vec4_u32 = ub[scalar_offset_39 / 4];
const float2x2 mat2x2_f32 = tint_symbol_14(ub, ((544u * idx) + 144u));
const float2x3 mat2x3_f32 = tint_symbol_15(ub, ((544u * idx) + 160u));
const float2x4 mat2x4_f32 = tint_symbol_16(ub, ((544u * idx) + 192u));
const float3x2 mat3x2_f32 = tint_symbol_17(ub, ((544u * idx) + 224u));
const float3x3 mat3x3_f32 = tint_symbol_18(ub, ((544u * idx) + 256u));
const float3x4 mat3x4_f32 = tint_symbol_19(ub, ((544u * idx) + 304u));
const float4x2 mat4x2_f32 = tint_symbol_20(ub, ((544u * idx) + 352u));
const float4x3 mat4x3_f32 = tint_symbol_21(ub, ((544u * idx) + 384u));
const float4x4 mat4x4_f32 = tint_symbol_22(ub, ((544u * idx) + 448u));
const float3 arr2_vec3_f32[2] = tint_symbol_23(ub, ((544u * idx) + 512u));
}
[numthreads(1, 1, 1)]
void main(tint_symbol_1 tint_symbol) {
main_inner(tint_symbol.idx);
return;
}