|
@@ -295,7 +295,7 @@
|
|
* float x, y, z, w;
|
|
* float x, y, z, w;
|
|
* } SDL_ALIGNED(16) MySIMDAlignedData;
|
|
* } SDL_ALIGNED(16) MySIMDAlignedData;
|
|
*
|
|
*
|
|
-
|
|
|
|
|
|
+ *
|
|
* // make sure this one field in a struct is aligned to 16 bytes for SIMD access.
|
|
* // make sure this one field in a struct is aligned to 16 bytes for SIMD access.
|
|
* typedef struct {
|
|
* typedef struct {
|
|
* SomeStuff stuff;
|
|
* SomeStuff stuff;
|
|
@@ -312,9 +312,9 @@
|
|
* something that respects this alignment, use SDL_aligned_alloc() or some
|
|
* something that respects this alignment, use SDL_aligned_alloc() or some
|
|
* other mechanism.
|
|
* other mechanism.
|
|
*
|
|
*
|
|
- * On compilers without alignment support, this macro is defined to an
|
|
|
|
- * invalid symbol, to make it clear that the current compiler is likely to
|
|
|
|
- * generate incorrect code when it sees this macro.
|
|
|
|
|
|
+ * On compilers without alignment support, this macro is defined to an invalid
|
|
|
|
+ * symbol, to make it clear that the current compiler is likely to generate
|
|
|
|
+ * incorrect code when it sees this macro.
|
|
*
|
|
*
|
|
* \param x the byte count to align to, so the data's address will be a
|
|
* \param x the byte count to align to, so the data's address will be a
|
|
* multiple of this value.
|
|
* multiple of this value.
|