Attachment 348452 Details for Bug 189101 – Patch

[patch] Patch

bug-189101-20180829165837.patch (text/plain), 7.21 KB, created by Myles C. Maxfield on 2018-08-29 16:58:38 PDT

(hide)

Description:

Filename:

MIME Type:

Creator: Myles C. Maxfield

Created: 2018-08-29 16:58:38 PDT

Size: 7.21 KB

patch

obsolete

>Subversion Revision: 235487
>diff --git a/Tools/ChangeLog b/Tools/ChangeLog
>index 3ad6772a8a07b5ea6819e7bb50bf6e7ba66efd99..be5c5927fb808c91e1bb87e3ac9fcdfb9b3a5400 100644
>--- a/Tools/ChangeLog
>+++ b/Tools/ChangeLog
>@@ -1,3 +1,19 @@
>+2018-08-29  Myles C. Maxfield  <mmaxfield@apple.com>
>+
>+        [WHLSL] Test row-majorness of matrices
>+        https://bugs.webkit.org/show_bug.cgi?id=189101
>+
>+        The matrix multiplication functions are temporarily commented out of the standard library,
>+        so I've temporarily copy/pasted them into the test. Matrix multiplication is not
>+        commutative, so it requires the right indexing order.
>+
>+        Reviewed by NOBODY (OOPS!).
>+
>+        * WebGPUShadingLanguageRI/Intrinsics.js:
>+        * WebGPUShadingLanguageRI/StandardLibrary.js:
>+        (let.standardLibrary):
>+        * WebGPUShadingLanguageRI/Test.js:
>+
> 2018-08-29  Jer Noble  <jer.noble@apple.com>
> 
>         Unreviewed test gardening; NowPlayingTest API tests require High Sierra.
>diff --git a/Tools/WebGPUShadingLanguageRI/Intrinsics.js b/Tools/WebGPUShadingLanguageRI/Intrinsics.js
>index 79f59070daffb55a3f3a70a3b1badb4ce68cd4ef..b004067f5ebfb3a8b4bd98e437a8c0768be26f1e 100644
>--- a/Tools/WebGPUShadingLanguageRI/Intrinsics.js
>+++ b/Tools/WebGPUShadingLanguageRI/Intrinsics.js
>@@ -316,11 +316,11 @@ class Intrinsics {
>             }
>         }
> 
>-        for (let type of ["half", "float"]) {
>+        for (let typeName of ["half", "float"]) {
>             for (let height = 2; height <= 4; ++height) {
>                 for (let width = 2; width <= 4; ++width) {
>-                    this._map.set(`native typedef matrix<${type}, ${height}, ${width}>`, type => {
>-                        this[`matrix<${type}, ${height}, ${width}>`] = type;
>+                    this._map.set(`native typedef matrix<${typeName}, ${height}, ${width}>`, type => {
>+                        this[`matrix<${typeName}, ${height}, ${width}>`] = type;
>                     });
>                 }
>             }
>diff --git a/Tools/WebGPUShadingLanguageRI/StandardLibrary.js b/Tools/WebGPUShadingLanguageRI/StandardLibrary.js
>index 6a33236617a431bd8239baa469d3ed80c1a9a02f..3dd4fee18600d5cd2a895ad5a479721da7788cc1 100644
>--- a/Tools/WebGPUShadingLanguageRI/StandardLibrary.js
>+++ b/Tools/WebGPUShadingLanguageRI/StandardLibrary.js
>@@ -1876,9 +1876,9 @@ let standardLibrary = (function() {
>                         print(`    ${type}${i}x${k} result;`);
>                         for (var p = 0; p < i; ++p) {
>                             for (var r = 0; r < k; ++r) {
>-                                print(`    result[${p}][${k}] = 0;`);
>+                                print(`    result[${p}][${r}] = 0;`);
>                                 for (var q = 0; q < j; ++q) {
>-                                    print(`    result[${p}][${k}] += x[${p}][${q}] * y[${q}][${r}];`);
>+                                    print(`    result[${p}][${r}] += x[${p}][${q}] * y[${q}][${r}];`);
>                                 }
>                             }
>                         }
>diff --git a/Tools/WebGPUShadingLanguageRI/Test.js b/Tools/WebGPUShadingLanguageRI/Test.js
>index ab96bc477f7f1e29dbdf62f5260ce49e5e7100d8..2b5f872d8970c9e9e039a8b1827e377380789ffb 100644
>--- a/Tools/WebGPUShadingLanguageRI/Test.js
>+++ b/Tools/WebGPUShadingLanguageRI/Test.js
>@@ -5451,6 +5451,106 @@ tests.halfSimpleMath = function() {
>     checkHalf(program, callFunction(program, "foo", [makeHalf(program, 7), makeHalf(program, -2)]), -3.5);
> }
> 
>+tests.matrixMultiplication = function() {
>+    let program = doPrep(`
>+        float2x4 multiply(float2x3 x, float3x4 y) {
>+            // Copied and pasted from the standard library
>+            float2x4 result;
>+            result[0][0] = 0;
>+            result[0][0] += x[0][0] * y[0][0];
>+            result[0][0] += x[0][1] * y[1][0];
>+            result[0][0] += x[0][2] * y[2][0];
>+            result[0][1] = 0;
>+            result[0][1] += x[0][0] * y[0][1];
>+            result[0][1] += x[0][1] * y[1][1];
>+            result[0][1] += x[0][2] * y[2][1];
>+            result[0][2] = 0;
>+            result[0][2] += x[0][0] * y[0][2];
>+            result[0][2] += x[0][1] * y[1][2];
>+            result[0][2] += x[0][2] * y[2][2];
>+            result[0][3] = 0;
>+            result[0][3] += x[0][0] * y[0][3];
>+            result[0][3] += x[0][1] * y[1][3];
>+            result[0][3] += x[0][2] * y[2][3];
>+            result[1][0] = 0;
>+            result[1][0] += x[1][0] * y[0][0];
>+            result[1][0] += x[1][1] * y[1][0];
>+            result[1][0] += x[1][2] * y[2][0];
>+            result[1][1] = 0;
>+            result[1][1] += x[1][0] * y[0][1];
>+            result[1][1] += x[1][1] * y[1][1];
>+            result[1][1] += x[1][2] * y[2][1];
>+            result[1][2] = 0;
>+            result[1][2] += x[1][0] * y[0][2];
>+            result[1][2] += x[1][1] * y[1][2];
>+            result[1][2] += x[1][2] * y[2][2];
>+            result[1][3] = 0;
>+            result[1][3] += x[1][0] * y[0][3];
>+            result[1][3] += x[1][1] * y[1][3];
>+            result[1][3] += x[1][2] * y[2][3];
>+            return result;
>+        }
>+        float2x3 matrix1() {
>+            float2x3 x;
>+            x[0][0] = 2;
>+            x[0][1] = 3;
>+            x[0][2] = 5;
>+            x[1][0] = 7;
>+            x[1][1] = 11;
>+            x[1][2] = 13;
>+            return x;
>+        }
>+        float3x4 matrix2() {
>+            float3x4 y;
>+            y[0][0] = 17;
>+            y[0][1] = 19;
>+            y[0][2] = 23;
>+            y[0][3] = 29;
>+            y[1][0] = 31;
>+            y[1][1] = 37;
>+            y[1][2] = 41;
>+            y[1][3] = 43;
>+            y[2][0] = 47;
>+            y[2][1] = 53;
>+            y[2][2] = 59;
>+            y[2][3] = 61;
>+            return y;
>+        }
>+        float foo00() {
>+            return multiply(matrix1(), matrix2())[0][0];
>+        }
>+        float foo01() {
>+            return multiply(matrix1(), matrix2())[0][1];
>+        }
>+        float foo02() {
>+            return multiply(matrix1(), matrix2())[0][2];
>+        }
>+        float foo03() {
>+            return multiply(matrix1(), matrix2())[0][3];
>+        }
>+        float foo10() {
>+            return multiply(matrix1(), matrix2())[1][0];
>+        }
>+        float foo11() {
>+            return multiply(matrix1(), matrix2())[1][1];
>+        }
>+        float foo12() {
>+            return multiply(matrix1(), matrix2())[1][2];
>+        }
>+        float foo13() {
>+            return multiply(matrix1(), matrix2())[1][3];
>+        }
>+    `);
>+    checkFloat(program, callFunction(program, "foo00", []), 17 * 2 + 31 * 3 + 47 * 5);
>+    checkFloat(program, callFunction(program, "foo01", []), 19 * 2 + 37 * 3 + 53 * 5);
>+    checkFloat(program, callFunction(program, "foo02", []), 23 * 2 + 41 * 3 + 59 * 5);
>+    checkFloat(program, callFunction(program, "foo03", []), 29 * 2 + 43 * 3 + 61 * 5);
>+    checkFloat(program, callFunction(program, "foo10", []), 17 * 7 + 31 * 11 + 47 * 13);
>+    checkFloat(program, callFunction(program, "foo11", []), 19 * 7 + 37 * 11 + 53 * 13);
>+    checkFloat(program, callFunction(program, "foo12", []), 23 * 7 + 41 * 11 + 59 * 13);
>+    checkFloat(program, callFunction(program, "foo13", []), 29 * 7 + 43 * 11 + 61 * 13);
>+}
>+
> okToTest = true;
> 
> let testFilter = /.*/; // run everything by default

Flags:

ews-feeder: review-
mmaxfield: commit-queue+

Actions: View | Formatted Diff | Diff

Attachments on bug 189101: 348440 | 348451 | 348452 | 348457 | 348458