| //------------------------------------------------------------------------------------------------------- |
| // Copyright (C) Microsoft Corporation and contributors. All rights reserved. |
| // Licensed under the MIT license. See LICENSE.txt file in the project root for full license information. |
| //------------------------------------------------------------------------------------------------------- |
| this.WScript.LoadScriptFile("..\\UnitTestFramework\\SimdJsHelpers.js"); |
| function asmModule(stdlib, imports,buffer) { |
| "use asm"; |
| |
| |
| var i4 = stdlib.SIMD.Int32x4; |
| var i4check = i4.check; |
| var i4splat = i4.splat; |
| |
| var i4fromFloat32x4 = i4.fromFloat32x4; |
| var i4fromFloat32x4Bits = i4.fromFloat32x4Bits; |
| //var i4abs = i4.abs; |
| var i4neg = i4.neg; |
| var i4add = i4.add; |
| var i4sub = i4.sub; |
| var i4mul = i4.mul; |
| //var i4swizzle = i4.swizzle; |
| //var i4shuffle = i4.shuffle; |
| var i4lessThan = i4.lessThan; |
| var i4equal = i4.equal; |
| var i4greaterThan = i4.greaterThan; |
| var i4select = i4.select; |
| var i4and = i4.and; |
| var i4or = i4.or; |
| var i4xor = i4.xor; |
| var i4not = i4.not; |
| var i4shiftLeftByScalar = i4.shiftLeftByScalar; |
| var i4shiftRightByScalar = i4.shiftRightByScalar; |
| |
| var i8 = stdlib.SIMD.Int16x8; |
| var i8check = i8.check; |
| |
| |
| |
| var u4 = stdlib.SIMD.Uint32x4; |
| var u4check = u4.check; |
| |
| var u8 = stdlib.SIMD.Uint16x8; |
| var u8check = u8.check; |
| |
| |
| var i16 = stdlib.SIMD.Int8x16; |
| var i16check = i16.check; |
| var i16extractLane = i16.extractLane ; |
| var i16swizzle = i16.swizzle ; |
| var i16shuffle = i16.shuffle ; |
| |
| var i16splat = i16.splat ; |
| var i16replaceLane = i16.replaceLane ; |
| var i16and = i16.and ; |
| var i16or = i16.or ; |
| var i16xor = i16.xor ; |
| var i16not = i16.not ; |
| var i16add = i16.add ; |
| var i16sub = i16.sub ; |
| var i16mul = i16.mul ; |
| var i16shiftLeftByScalar = i16.shiftLeftByScalar ; |
| var i16shiftRightByScalar= i16.shiftRightByScalar ; |
| var i16addSaturate = i16.addSaturate ; |
| var i16subSaturate = i16.subSaturate ; |
| var i16load = i16.load ; |
| var i16store = i16.store ; |
| var i16fromFloat32x4Bits = i16.fromFloat32x4Bits ; |
| var i16fromInt32x4Bits = i16.fromInt32x4Bits ; |
| var i16fromInt16x8Bits = i16.fromInt16x8Bits ; |
| var i16fromUint32x4Bits = i16.fromUint32x4Bits ; |
| var i16fromUint16x8Bits = i16.fromUint16x8Bits ; |
| |
| |
| |
| |
| |
| |
| var f4 = stdlib.SIMD.Float32x4; |
| var f4check = f4.check; |
| var f4splat = f4.splat; |
| |
| var f4fromInt32x4 = f4.fromInt32x4; |
| var f4fromInt32x4Bits = f4.fromInt32x4Bits; |
| var f4abs = f4.abs; |
| var f4neg = f4.neg; |
| var f4add = f4.add; |
| var f4sub = f4.sub; |
| var f4mul = f4.mul; |
| var f4div = f4.div; |
| var f4min = f4.min; |
| var f4max = f4.max; |
| |
| |
| var f4sqrt = f4.sqrt; |
| //var f4swizzle = f4.swizzle; |
| //var f4shuffle = f4.shuffle; |
| var f4lessThan = f4.lessThan; |
| var f4lessThanOrEqual = f4.lessThanOrEqual; |
| var f4equal = f4.equal; |
| var f4notEqual = f4.notEqual; |
| var f4greaterThan = f4.greaterThan; |
| var f4greaterThanOrEqual = f4.greaterThanOrEqual; |
| |
| var f4select = f4.select; |
| |
| |
| |
| var fround = stdlib.Math.fround; |
| |
| var globImportF4 = f4check(imports.g1); // global var import |
| var globImportI4 = i4check(imports.g2); // global var import |
| |
| |
| var f4g1 = f4(-5033.2,-3401.0,665.34,32234.1); // global var initialized |
| var f4g2 = f4(1194580.33,-11201.5,63236.93,334.8); // global var initialized |
| |
| var i4g1 = i4(1065353216, -1073741824, -1077936128, 1082130432); // global var initialized |
| var i4g2 = i4(353216, -492529, -1128, 1085); // global var initialized |
| |
| |
| |
| var i16g1 = i16(1065353216, -1073741824, -1077936128, 1082130432, 1065353216, -1073741824, -1077936128, 1082130432, 0, -124, 6128, 10432, 216, -824, -3128, 10); |
| var i16g2 = i16(0, -124, 6128, 10432, 216, -824, -3128, 10, -1077936128, 1082130432, 1065353216, -1073741824, -1077936128, 1082130432, 0, -124); |
| var gval = 1234; |
| var gval2 = 1234.0; |
| |
| |
| var Int8Heap = new stdlib.Int8Array (buffer); |
| var Uint8Heap = new stdlib.Uint8Array (buffer); |
| |
| var Int16Heap = new stdlib.Int16Array(buffer); |
| var Uint16Heap = new stdlib.Uint16Array(buffer); |
| var Int32Heap = new stdlib.Int32Array(buffer); |
| var Uint32Heap = new stdlib.Uint32Array(buffer); |
| var Float32Heap = new stdlib.Float32Array(buffer); |
| |
| var loopCOUNT = 3; |
| |
| function func1() |
| { |
| var x = i16(-1, -2, -3, -4, -5, -6, -7, -8, -1, -2, -3, -4, -5, -6, -7, -8);; |
| var y = 0; |
| var s = 0; |
| var loopIndex = 0; |
| var loopCOUNT = 3; |
| |
| x = i16g2; |
| while ( (loopIndex|0) < (loopCOUNT|0)) { |
| y = i16extractLane(x, 0); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 1); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 2); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 3); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 4); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 5); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 6); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 7); |
| s = ( s + y ) | 0; |
| |
| y = i16extractLane(x, 8); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 9); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 10); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 11); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 12); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 13); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 14); |
| s = ( s + y ) | 0; |
| y = i16extractLane(x, 15); |
| s = ( s + y ) | 0; |
| |
| loopIndex = (loopIndex + 1) | 0; |
| |
| } |
| |
| return s | 0; |
| } |
| |
| function func2(a) |
| { |
| a = i16check(a); |
| var x = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| var loopIndex = 0; |
| var loopCOUNT = 3; |
| |
| while ( (loopIndex|0) < (loopCOUNT|0)) { |
| |
| x = i16swizzle(x, 1, 2, 3, 4, 0, 7, 7, 7, 11, 13, 12, 14, 10, 15, 15, 15); |
| x = i16shuffle(a, x, 0, 1, 10, 11, 2, 3, 12, 13, 20, 21, 31, 3, 22, 0, 18, 31); |
| loopIndex = (loopIndex + 1) | 0; |
| } |
| |
| return i16check(x); |
| } |
| |
| function func3(a) |
| { |
| a = i16check(a); |
| var x = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| var y = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| var loopIndex = 0; |
| var loopCOUNT = 3; |
| |
| |
| while ( (loopIndex|0) < (loopCOUNT|0)) { |
| x = i16splat(loopIndex); |
| x = i16replaceLane(x, 0, 200); |
| x = i16replaceLane(y, 9, loopIndex); |
| y = x; |
| loopIndex = (loopIndex + 1) | 0; |
| } |
| y = i16add(y, i16(0,1,1,1,loopIndex, loopIndex, loopIndex, loopIndex,1,1,1,1,0,0,0,0)); |
| return i16check(y); |
| } |
| |
| |
| function func4(a, b, c, d, e) |
| { |
| a = i16check(a); |
| b = i16check(b); |
| c = i16check(c); |
| d = i16check(d); |
| e = i16check(e); |
| var x = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| var y = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| var loopIndex = 0; |
| var loopCOUNT = 3; |
| |
| |
| while ( (loopIndex|0) < (loopCOUNT|0)) { |
| x = i16and(a, b); |
| x = i16or(x, d); |
| x = i16xor(x, e); |
| x = i16not(x); |
| |
| loopIndex = (loopIndex + 1) | 0; |
| } |
| |
| return i16check(x); |
| } |
| |
| function func5(a, b, c, d, e) |
| { |
| a = i16check(a); |
| b = i16check(b); |
| c = i16check(c); |
| d = i16check(d); |
| e = i16check(e); |
| var x = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| var y = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| var loopIndex = 0; |
| var loopCOUNT = 3; |
| |
| |
| while ( (loopIndex|0) < (loopCOUNT|0)) { |
| x = i16add(a, b); |
| x = i16sub(x, d); |
| x = i16mul(x, e); |
| |
| loopIndex = (loopIndex + 1) | 0; |
| } |
| |
| return i16check(x); |
| } |
| |
| |
| function func7() |
| { |
| var loopIndex = 0; |
| var loopCOUNT = 3; |
| var index = 100; |
| var size = 10; |
| var x = i16(1, 2, 3,0xff, -1, -2, -3, -4, 16, 24, 32, 41, -18, -27, -63, 0xff); |
| var y = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| var t = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| |
| for (loopIndex = 0; (loopIndex | 0) < (size | 0) ; loopIndex = (loopIndex + 1) | 0) |
| { |
| i16store(Uint32Heap, index >> 2, x); |
| index = (index + 16 ) | 0; |
| } |
| |
| index = 100; |
| for (loopIndex = 0; (loopIndex | 0) < (size | 0) ; loopIndex = (loopIndex + 1) | 0) |
| { |
| t = i16load(Float32Heap, index >> 2); |
| y = i16add(y, t); |
| index = (index + 16 ) | 0; |
| } |
| return i16check(y); |
| } |
| |
| function func8() |
| { |
| var loopIndex = 0; |
| var loopCOUNT = 3; |
| var index = 100; |
| var size = 10; |
| var x = i16(1, 2, 3,0xff, -1, -2, -3, -4, 16, 24, 32, 41, -18, -27, -63, 0xff); |
| var y = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| var t = i16(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); |
| var v_i4 = i4(-29,34,10,4); |
| var v_f4 = f4(1.0, -200.0, 123.443, 3000.1); |
| var v_u4 = u4(-29,0xffff,10,4); |
| var v_i8 = i8(-29,0xffff,10,4, -104, -9999, 1, 22); |
| var v_u8 = u8(-29,0xffff,10,4, -104, -9999, 1, 22); |
| |
| |
| for (loopIndex = 0; (loopIndex | 0) < (size | 0) ; loopIndex = (loopIndex + 1) | 0) |
| { |
| x = i16add(x, i16fromFloat32x4Bits(v_f4)); |
| x = i16add(x, i16fromInt32x4Bits(v_i4)); |
| x = i16add(x, i16fromInt16x8Bits(v_i8)); |
| x = i16add(x, i16fromUint32x4Bits(v_u4)); |
| x = i16add(x, i16fromUint16x8Bits(v_u8)); |
| } |
| |
| return i16check(x); |
| } |
| |
| function bug1() //simd tmp reg reuse. |
| { |
| var a = i4(1,2,3,4); |
| var x = i16(-1, -2, -3, -4,-1, -2, -3, -4, -1, -2, -3, -4,-1, -2, -3, -4); |
| i16add(x,x); |
| return i4check(a); |
| } |
| |
| return {func1:func1, func2: func2, func3:func3, func4:func4, |
| func5: func5, func7:func7, func8:func8, bug1:bug1}; |
| } |
| |
| var buffer = new ArrayBuffer(0x10000); |
| var m = asmModule(this, {g1:SIMD.Float32x4(90934.2,123.9,419.39,449.0), g2:SIMD.Int32x4(-1065353216, -1073741824,-1077936128, -1082130432)}, buffer); |
| |
| |
| |
| var v1 = SIMD.Int8x16(1, 2, 3, 4, 5, 6, 7, 8, 1, 2, 3, 4, 5, 6, 7, 8); |
| var v2 = SIMD.Int8x16(134, 211, -333, 422, -165, 999996, 0xffff, 0xf0f0, 134, 211, -333, 422, -165, 999996, 0xffff, 0xf0f0); |
| var v3 = SIMD.Int8x16(0xcccc, -211, 189, 422, -165, -999996, 0xffff, 0xf0f0, 0xcccc, -211, 189, 422, -165, -999996, 0xffff, 0xf0f0); |
| var ret1 = m.func1(); |
| // print (ret1); |
| var ret2 = m.func2(v1); |
| // printSimdBaseline(ret2, "SIMD.Int8x16", "ret2", "func2"); |
| |
| |
| var ret3 = m.func3(v1); |
| // printSimdBaseline(ret3, "SIMD.Int8x16", "ret3", "func3"); |
| |
| |
| var ret4 = m.func4(v1, v2, v3, v1, v2); |
| // printSimdBaseline(ret4, "SIMD.Int8x16", "ret4", "func4"); |
| |
| |
| var ret5 = m.func5(v1, v2, v3, v1, v2); |
| // printSimdBaseline(ret5, "SIMD.Int8x16", "ret5", "func5"); |
| |
| |
| var ret7 = m.func7(v1, v2, v3, v1, v2); |
| // printSimdBaseline(ret7, "SIMD.Int8x16", "ret7", "func7"); |
| |
| |
| var ret8 = m.func8(v1, v2, v3, v1, v2); |
| // printSimdBaseline(ret8, "SIMD.Int8x16", "ret8", "func8"); |
| |
| var ret9 = m.bug1(); |
| // printSimdBaseline(ret9, "SIMD.Int32x4", "ret9", "bug1"); |
| |
| equal(-1410, ret1); |
| equalSimd([1, 2, 3, 4, 3, 4, 5, 6, 1, 6, 0, 4, 6, 1, 4, 0], ret2, SIMD.Int8x16, "func2") |
| equalSimd([0, 1, 1, 1, 3, 3, 3, 3, 1, 3, 1, 1, 0, 0, 0, 0], ret3, SIMD.Int8x16, "func3") |
| equalSimd([120, 46, 79, 93, -95, -59, 7, 7, 120, 46, 79, 93, -95, -59, 7, 7], ret4, SIMD.Int8x16, "func4") |
| equalSimd([36, -23, 41, -92, 89, 16, 1, 0, 36, -23, 41, -92, 89, 16, 1, 0], ret5, SIMD.Int8x16, "func5") |
| equalSimd([10, 20, 30, -10, -10, -20, -30, -40, -96, -16, 64, -102, 76, -14, -118, -10], ret7, SIMD.Int8x16, "func7") |
| equalSimd([121, -38, -37, 77, 17, -12, 29, -102, -30, -40, -112, -99, 86, -17, -57, -79], ret8, SIMD.Int8x16, "func8") |
| equalSimd([1, 2, 3, 4], ret9, SIMD.Int32x4, "bug1") |
| |
| print("PASS"); |
| |
| |
