diff options
author | 2015-06-24 13:08:51 -0700 | |
---|---|---|
committer | 2015-06-24 13:08:52 -0700 | |
commit | 4760f32f2a68acdcfee53eae98bbdabb19238e8d (patch) | |
tree | 41d44095f742edc83534cb3cb4ed82341bcb7253 /src/core/SkPoint.cpp | |
parent | 0c752f4760f9b501eea5fb4b5fc629d5221473f9 (diff) |
Clarify iOS flush-to-zero status.
Modern processors support flush-to-zero and denormalized-are-zero
for floating point operations (with ARM NEON unable to disable them).
However, iOS on ARM is the only current system which defaults
processes to using both all the time. However, this is only iOS on
ARM, iOS on x86 (the simulator) does not. Correctly defining this
allows the math tests to run error free in the simulator.
Review URL: https://codereview.chromium.org/1203533003
Diffstat (limited to 'src/core/SkPoint.cpp')
-rw-r--r-- | src/core/SkPoint.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/src/core/SkPoint.cpp b/src/core/SkPoint.cpp index aabde2786d..345ef37448 100644 --- a/src/core/SkPoint.cpp +++ b/src/core/SkPoint.cpp @@ -161,7 +161,7 @@ bool SkPoint::setLength(float x, float y, float length) { // divide by inf. and return (0,0) vector. double xx = x; double yy = y; - #ifdef SK_DISCARD_DENORMALIZED_FOR_SPEED + #ifdef SK_CPU_FLUSH_TO_ZERO // The iOS ARM processor discards small denormalized numbers to go faster. // Casting this to a float would cause the scale to go to zero. Keeping it // as a double for the multiply keeps the scale non-zero. |