encounter/dawn-cmake
1
0
Fork 0
mirror of https://github.com/encounter/dawn-cmake.git synced 2025-12-14 15:46:28 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix FXC compile errors on divide by zero

Browse Source 
FXC fails to compile when it determines that the rhs of an integral
division is zero with "error X4010: Unsigned integer divide by zero".

bclayton's fix (https://dawn-review.googlesource.com/c/tint/+/60500)
addressed cases for division by an integer constant 0. This CL adds the
missing support for division by integral vectors with 0 components.

FXC also fails on division by integral expressions that it can fold to
0. To handle these cases, we now emit a runtime check for 0 and replace
by 1. In the cases I've tested, FXC seems able to optimize these checks
away.

Bug: tint:1083
Change-Id: I02f08e9077882f03c1e42b62dacb742a48fa48ba
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/73580
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
This commit is contained in:
Antonio Maiorano
2021-12-21 14:48:26 +00:00
committed by Tint LUCI CQ
parent 5965c6ed1f
commit 821f9bb525
187 changed files with 2691 additions and 40 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/f32.wgsl Normal file
View File

@@ -0,0 +1,6 @@
[[stage(compute), workgroup_size(1)]]
fn f() {
var a = 4.;
var b = vec3<f32>(0., 2., 0.);
let r : vec3<f32> = a / (b + b);
}

7
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/f32.wgsl.expected.hlsl Normal file
View File

@@ -0,0 +1,7 @@
[numthreads(1, 1, 1)]
void f() {
float a = 4.0f;
float3 b = float3(0.0f, 2.0f, 0.0f);
const float3 r = (a / (b + b));
return;
}

10
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/f32.wgsl.expected.msl Normal file
View File

@@ -0,0 +1,10 @@
#include <metal_stdlib>
using namespace metal;
kernel void f() {
float a = 4.0f;
float3 b = float3(0.0f, 2.0f, 0.0f);
float3 const r = (a / (b + b));
return;
}

39
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/f32.wgsl.expected.spvasm Normal file
View File

@@ -0,0 +1,39 @@
; SPIR-V
; Version: 1.3
; Generator: Google Tint Compiler; 0
; Bound: 24
; Schema: 0
OpCapability Shader
OpMemoryModel Logical GLSL450
OpEntryPoint GLCompute %f "f"
OpExecutionMode %f LocalSize 1 1 1
OpName %f "f"
OpName %a "a"
OpName %b "b"
%void = OpTypeVoid
%1 = OpTypeFunction %void
%float = OpTypeFloat 32
%float_4 = OpConstant %float 4
%_ptr_Function_float = OpTypePointer Function %float
%9 = OpConstantNull %float
%v3float = OpTypeVector %float 3
%float_0 = OpConstant %float 0
%float_2 = OpConstant %float 2
%13 = OpConstantComposite %v3float %float_0 %float_2 %float_0
%_ptr_Function_v3float = OpTypePointer Function %v3float
%16 = OpConstantNull %v3float
%f = OpFunction %void None %1
%4 = OpLabel
%a = OpVariable %_ptr_Function_float Function %9
%b = OpVariable %_ptr_Function_v3float Function %16
%22 = OpVariable %_ptr_Function_v3float Function %16
OpStore %a %float_4
OpStore %b %13
%17 = OpLoad %float %a
%18 = OpLoad %v3float %b
%19 = OpLoad %v3float %b
%20 = OpFAdd %v3float %18 %19
%23 = OpCompositeConstruct %v3float %17 %17 %17
%21 = OpFDiv %v3float %23 %20
OpReturn
OpFunctionEnd

6
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/f32.wgsl.expected.wgsl Normal file
View File

@@ -0,0 +1,6 @@
[[stage(compute), workgroup_size(1)]]
fn f() {
var a = 4.0;
var b = vec3<f32>(0.0, 2.0, 0.0);
let r : vec3<f32> = (a / (b + b));
}

6
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/i32.wgsl Normal file
View File

@@ -0,0 +1,6 @@
[[stage(compute), workgroup_size(1)]]
fn f() {
var a = 4;
var b = vec3<i32>(0, 2, 0);
let r : vec3<i32> = a / (b + b);
}

11
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/i32.wgsl.expected.hlsl Normal file
View File

@@ -0,0 +1,11 @@
int3 value_or_one_if_zero_int3(int3 value) {
return value == int3(0, 0, 0) ? int3(1, 1, 1) : value;
}
[numthreads(1, 1, 1)]
void f() {
int a = 4;
int3 b = int3(0, 2, 0);
const int3 r = (a / value_or_one_if_zero_int3((b + b)));
return;
}

10
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/i32.wgsl.expected.msl Normal file
View File

@@ -0,0 +1,10 @@
#include <metal_stdlib>
using namespace metal;
kernel void f() {
int a = 4;
int3 b = int3(0, 2, 0);
int3 const r = (a / as_type<int3>((as_type<uint3>(b) + as_type<uint3>(b))));
return;
}

39
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/i32.wgsl.expected.spvasm Normal file
View File

@@ -0,0 +1,39 @@
; SPIR-V
; Version: 1.3
; Generator: Google Tint Compiler; 0
; Bound: 24
; Schema: 0
OpCapability Shader
OpMemoryModel Logical GLSL450
OpEntryPoint GLCompute %f "f"
OpExecutionMode %f LocalSize 1 1 1
OpName %f "f"
OpName %a "a"
OpName %b "b"
%void = OpTypeVoid
%1 = OpTypeFunction %void
%int = OpTypeInt 32 1
%int_4 = OpConstant %int 4
%_ptr_Function_int = OpTypePointer Function %int
%9 = OpConstantNull %int
%v3int = OpTypeVector %int 3
%int_0 = OpConstant %int 0
%int_2 = OpConstant %int 2
%13 = OpConstantComposite %v3int %int_0 %int_2 %int_0
%_ptr_Function_v3int = OpTypePointer Function %v3int
%16 = OpConstantNull %v3int
%f = OpFunction %void None %1
%4 = OpLabel
%a = OpVariable %_ptr_Function_int Function %9
%b = OpVariable %_ptr_Function_v3int Function %16
%22 = OpVariable %_ptr_Function_v3int Function %16
OpStore %a %int_4
OpStore %b %13
%17 = OpLoad %int %a
%18 = OpLoad %v3int %b
%19 = OpLoad %v3int %b
%20 = OpIAdd %v3int %18 %19
%23 = OpCompositeConstruct %v3int %17 %17 %17
%21 = OpSDiv %v3int %23 %20
OpReturn
OpFunctionEnd

6
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/i32.wgsl.expected.wgsl Normal file
View File

@@ -0,0 +1,6 @@
[[stage(compute), workgroup_size(1)]]
fn f() {
var a = 4;
var b = vec3<i32>(0, 2, 0);
let r : vec3<i32> = (a / (b + b));
}

6
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/u32.wgsl Normal file
View File

@@ -0,0 +1,6 @@
[[stage(compute), workgroup_size(1)]]
fn f() {
var a = 4u;
var b = vec3<u32>(0u, 2u, 0u);
let r : vec3<u32> = a / (b + b);
}

11
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/u32.wgsl.expected.hlsl Normal file
View File

@@ -0,0 +1,11 @@
uint3 value_or_one_if_zero_uint3(uint3 value) {
return value == uint3(0u, 0u, 0u) ? uint3(1u, 1u, 1u) : value;
}
[numthreads(1, 1, 1)]
void f() {
uint a = 4u;
uint3 b = uint3(0u, 2u, 0u);
const uint3 r = (a / value_or_one_if_zero_uint3((b + b)));
return;
}

10
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/u32.wgsl.expected.msl Normal file
View File

@@ -0,0 +1,10 @@
#include <metal_stdlib>
using namespace metal;
kernel void f() {
uint a = 4u;
uint3 b = uint3(0u, 2u, 0u);
uint3 const r = (a / (b + b));
return;
}

39
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/u32.wgsl.expected.spvasm Normal file
View File

@@ -0,0 +1,39 @@
; SPIR-V
; Version: 1.3
; Generator: Google Tint Compiler; 0
; Bound: 24
; Schema: 0
OpCapability Shader
OpMemoryModel Logical GLSL450
OpEntryPoint GLCompute %f "f"
OpExecutionMode %f LocalSize 1 1 1
OpName %f "f"
OpName %a "a"
OpName %b "b"
%void = OpTypeVoid
%1 = OpTypeFunction %void
%uint = OpTypeInt 32 0
%uint_4 = OpConstant %uint 4
%_ptr_Function_uint = OpTypePointer Function %uint
%9 = OpConstantNull %uint
%v3uint = OpTypeVector %uint 3
%uint_0 = OpConstant %uint 0
%uint_2 = OpConstant %uint 2
%13 = OpConstantComposite %v3uint %uint_0 %uint_2 %uint_0
%_ptr_Function_v3uint = OpTypePointer Function %v3uint
%16 = OpConstantNull %v3uint
%f = OpFunction %void None %1
%4 = OpLabel
%a = OpVariable %_ptr_Function_uint Function %9
%b = OpVariable %_ptr_Function_v3uint Function %16
%22 = OpVariable %_ptr_Function_v3uint Function %16
OpStore %a %uint_4
OpStore %b %13
%17 = OpLoad %uint %a
%18 = OpLoad %v3uint %b
%19 = OpLoad %v3uint %b
%20 = OpIAdd %v3uint %18 %19
%23 = OpCompositeConstruct %v3uint %17 %17 %17
%21 = OpUDiv %v3uint %23 %20
OpReturn
OpFunctionEnd

6
test/expressions/binary/div_by_zero/by_expression/scalar-vec3/u32.wgsl.expected.wgsl Normal file
View File

@@ -0,0 +1,6 @@
[[stage(compute), workgroup_size(1)]]
fn f() {
var a = 4u;
var b = vec3<u32>(0u, 2u, 0u);
let r : vec3<u32> = (a / (b + b));
}