SIMD/Types/uint8x16: Difference between revisions
< SIMD
Jump to navigation
Jump to search
Maikmerten (talk | contribs) (Preliminary definition for uint8x16) |
Maikmerten (talk | contribs) (Uses for uint8x16) |
||
| Line 1: | Line 1: | ||
The data type uint8x16 represents 16 unsigned 8-bit integers, packed into one 128 bit SIMD register. | The data type uint8x16 represents 16 unsigned 8-bit integers, packed into one 128 bit SIMD register. | ||
== Uses for uint8x16 == | |||
* In computer graphics, pixel values for each color channel are often represented as uint8. Thus uint8x16 can be very useful when operating on pixel values (for instance, for [[SIMD/Uses/SAD |SAD]]). | |||
* In digital audio, some [http://en.wikipedia.org/wiki/Pcm PCM] formats store samples as uint8 values. Manipulating audio content stored in such a fashion may benefit from working with uint8x16. | |||
Latest revision as of 16:24, 10 April 2014
The data type uint8x16 represents 16 unsigned 8-bit integers, packed into one 128 bit SIMD register.
Uses for uint8x16
- In computer graphics, pixel values for each color channel are often represented as uint8. Thus uint8x16 can be very useful when operating on pixel values (for instance, for SAD).
- In digital audio, some PCM formats store samples as uint8 values. Manipulating audio content stored in such a fashion may benefit from working with uint8x16.