SIMD: implement _mm256_i8gather, _mm256_i16gather
I want to ask a question about SIMD. Intel doesn't offer gather functions like: _mm256_i8gather_epi8, _mm256_i16gather_epi16.
They should simply do the samiliar things like _mm256_i32gather_epi32 and _mm256_i64gather_epi64 , but just load 32 int8_t or 16 int16_t.
How can we implement these functions with AVX2?
They should simply do the samiliar things like _mm256_i32gather_epi32 and _mm256_i64gather_epi64 , but just load 32 int8_t or 16 int16_t.
How can we implement these functions with AVX2?
Topic archived. No new replies allowed.