SMAUG
Simulating Machine Learning Applications on gem5-Aladdin
|
Typedefs and macros for working with vector data. More...
Macros | |
#define | VECTOR_SIZE 8 |
Vector size used in SMV backends. | |
#define | VEC128_MASK(input, mask) ((v4fp_t)((v4sfx_t)input & mask)) |
Apply a elementwise mask to a 128-bit packed single precision FP vector. More... | |
#define | VEC256_MASK(input, mask) ((v8fp_t)((v8sfx_t)input & mask)) |
Same as VEC128_MASK, but for 256-bit vectors. More... | |
Typedefs | |
typedef fp16_t | v16fp_t |
16 packed 32-bit floating point values. | |
typedef fp_t | v8fp_t |
8 packed 32-bit floating point values. | |
typedef fp_t | v4fp_t |
4 packed 32-bit floating point values. | |
typedef fp16_t | v16ph_t |
16 packed 16-bit floating point values. | |
typedef fp16_t | v8ph_t |
8 packed 16-bit floating point values. | |
typedef fp16_t | v4ph_t |
4 packed 16-bit floating point values. | |
typedef sfx_t | v8sfx_t |
8 packed 32-bit integer values. | |
typedef sfx_t | v4sfx_t |
4 packed 32-bit integer values. | |
typedef uint8_t | v8bl_t |
8 packed 8-bit bool values. | |
Typedefs and macros for working with vector data.
Apply a elementwise mask to a 128-bit packed single precision FP vector.
The mask is a vector of either 0s or -1s (all 1s). Entries that are have a mask of 0 are zeroed out.
LLVM is smart enough to turn this into a SELECT instruction, rather than a bitwise mask!
input | A v4fp_t vector. |
mask | A v4sfx_t vector of either 0s or -1s. |