memchr

Mojo module 🡭

memchr

Fast memchr implementations.

There are two here, memchr and memchr_wide. memchr_wide will do more comparisons at once, but needs to do more loading first. If you know you have some distance between needles, then it should be faster. memchr is just vanilla memchr.

Aliases

• SIMD_U8_WIDTH = simdwidthof[::DType,__mlir_type.!kgen.target]()
• LOOP_SIZE = (simdwidthof[::DType,__mlir_type.!kgen.target]() * 4)

Functions

• memchr
• memchr_wide