64bbad360f
Reason for revert: Unit tests fail on Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-Fast Original issue's description: > Expand _01 half<->float limitation to _finite. Simplify. > > It's become clear we need to sometimes deal with values <0 or >1. > I'm not yet convinced we care about NaN or +-inf. > > We had some fairly clever tricks and optimizations here for NEON > and SSE. I've thrown them out in favor of a single implementation. > If we find the specializations mattered, we can certainly figure out > how to extend them to this new range/domain. > > This happens to add a vectorized float -> half for ARMv7, which was > missing from the _01 version. (The SSE strategy was not portable to > platforms that flush denorm floats to zero.) > > I've tested the full float range for FloatToHalf on my desktop and a 5x. > > BUG=skia: > GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2145663003 > CQ_INCLUDE_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot;master.client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot,Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-Fast-Trybot > > Committed: https://skia.googlesource.com/skia/+/3296bee70d074bb8094b3229dbe12fa016657e90 TBR=msarett@google.com,mtklein@chromium.org # Skipping CQ checks because original CL landed less than 1 days ago. NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true BUG=skia: Review-Url: https://codereview.chromium.org/2151023003
104 lines
2.8 KiB
C++
104 lines
2.8 KiB
C++
/*
|
|
* Copyright 2016 Google Inc.
|
|
*
|
|
* Use of this source code is governed by a BSD-style license that can be
|
|
* found in the LICENSE file.
|
|
*/
|
|
|
|
#include "Test.h"
|
|
#include "SkAutoPixmapStorage.h"
|
|
#include "SkColor.h"
|
|
#include "SkHalf.h"
|
|
#include "SkOpts.h"
|
|
#include "SkPixmap.h"
|
|
#include "SkPM4f.h"
|
|
#include "SkRandom.h"
|
|
|
|
static bool eq_within_half_float(float a, float b) {
|
|
const float kTolerance = 1.0f / (1 << (8 + 10));
|
|
|
|
SkHalf ha = SkFloatToHalf(a);
|
|
SkHalf hb = SkFloatToHalf(b);
|
|
float a2 = SkHalfToFloat(ha);
|
|
float b2 = SkHalfToFloat(hb);
|
|
return fabsf(a2 - b2) <= kTolerance;
|
|
}
|
|
|
|
static bool eq_within_half_float(const SkPM4f& a, const SkPM4f& b) {
|
|
for (int i = 0; i < 4; ++i) {
|
|
if (!eq_within_half_float(a.fVec[i], b.fVec[i])) {
|
|
return false;
|
|
}
|
|
}
|
|
return true;
|
|
}
|
|
|
|
DEF_TEST(color_half_float, reporter) {
|
|
const int w = 100;
|
|
const int h = 100;
|
|
|
|
SkImageInfo info = SkImageInfo::Make(w, h, kRGBA_F16_SkColorType, kPremul_SkAlphaType);
|
|
|
|
SkAutoPixmapStorage pm;
|
|
pm.alloc(info);
|
|
REPORTER_ASSERT(reporter, pm.getSafeSize() == SkToSizeT(w * h * sizeof(uint64_t)));
|
|
|
|
SkColor4f c4 { 1, 0.5f, 0.25f, 0.5f };
|
|
pm.erase(c4);
|
|
|
|
SkPM4f origpm4 = c4.premul();
|
|
for (int y = 0; y < pm.height(); ++y) {
|
|
for (int x = 0; x < pm.width(); ++x) {
|
|
SkPM4f pm4 = SkPM4f::FromF16(pm.addrF16(x, y));
|
|
REPORTER_ASSERT(reporter, eq_within_half_float(origpm4, pm4));
|
|
}
|
|
}
|
|
}
|
|
|
|
static uint32_t u(float f) {
|
|
uint32_t x;
|
|
memcpy(&x, &f, 4);
|
|
return x;
|
|
}
|
|
|
|
DEF_TEST(HalfToFloat_01, r) {
|
|
for (uint16_t h = 0; h < 0x8000; h++) {
|
|
float f = SkHalfToFloat(h);
|
|
if (f >= 0 && f <= 1) {
|
|
float got = SkHalfToFloat_01(h)[0];
|
|
if (got != f) {
|
|
SkDebugf("0x%04x -> 0x%08x (%g), want 0x%08x (%g)\n",
|
|
h,
|
|
u(got), got,
|
|
u(f), f);
|
|
}
|
|
REPORTER_ASSERT(r, SkHalfToFloat_01(h)[0] == f);
|
|
REPORTER_ASSERT(r, SkFloatToHalf_01(SkHalfToFloat_01(h)) == h);
|
|
}
|
|
}
|
|
}
|
|
|
|
DEF_TEST(FloatToHalf_01, r) {
|
|
#if 0
|
|
for (uint32_t bits = 0; bits < 0x80000000; bits++) {
|
|
#else
|
|
SkRandom rand;
|
|
for (int i = 0; i < 1000000; i++) {
|
|
uint32_t bits = rand.nextU();
|
|
#endif
|
|
float f;
|
|
memcpy(&f, &bits, 4);
|
|
if (f >= 0 && f <= 1) {
|
|
uint16_t h1 = (uint16_t)SkFloatToHalf_01(Sk4f(f,0,0,0)),
|
|
h2 = SkFloatToHalf(f);
|
|
bool ok = (h1 == h2 || h1 == h2-1);
|
|
REPORTER_ASSERT(r, ok);
|
|
if (!ok) {
|
|
SkDebugf("%08x (%d) -> %04x (%d), want %04x (%d)\n",
|
|
bits, bits>>23, h1, h1>>10, h2, h2>>10);
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
}
|