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

Review-Url: https://codereview.chromium.org/2145663003

1 files changed, 19 insertions, 0 deletions

diff --git a/tests/SkNxTest.cpp b/tests/SkNxTest.cpp
index 5509814c07..51d937dd4d 100644
--- a/tests/SkNxTest.cpp
+++ b/tests/SkNxTest.cpp
@@ -288,3 +288,22 @@ DEF_TEST(SkNx_u16_float, r) {
         REPORTER_ASSERT(r, !memcmp(s16, d16, sizeof(s16)));
     }
 }
+
+// The SSE2 implementation of SkNx_cast<uint16_t>(Sk4i) is non-trivial, so worth a test.
+DEF_TEST(SkNx_int_u16, r) {
+    // These are pretty hard to get wrong.
+    for (int i = 0; i <= 0x7fff; i++) {
+        uint16_t expected = (uint16_t)i;
+        uint16_t actual = SkNx_cast<uint16_t>(Sk4i(i))[0];
+
+        REPORTER_ASSERT(r, expected == actual);
+    }
+
+    // A naive implementation with _mm_packs_epi32 would succeed up to 0x7fff but fail here:
+    for (int i = 0x8000; (1) && i <= 0xffff; i++) {
+        uint16_t expected = (uint16_t)i;
+        uint16_t actual = SkNx_cast<uint16_t>(Sk4i(i))[0];
+
+        REPORTER_ASSERT(r, expected == actual);
+    }
+}