Remove stdsimd requirement

README.md

@@ -32,7 +32,7 @@ hashset.insert("hello world");
 ## Compatibility
 - ARM 64-bit using `NEON` intrinsics.
 - x86-64 bit using `SSE2` + `AES` intrinsics.
-- (optional) with `avx2` feature enabled, gxhash will use `AVX2` intrinsics, for up to twice as much performance for large inputs. Only compatible on `AVX2` enabled x86-64 platforms.
+- (optional and only on unstable toolchain) with `avx2` feature enabled, gxhash will use `AVX2` intrinsics, for up to twice as much performance for large inputs. Only compatible on `AVX2` enabled x86-64 platforms.
 
 > **Warning**
 > Other platforms are currently not supported (there is no fallback)

src/gxhash/platform/x86_128.rs

@@ -36,7 +36,7 @@ pub unsafe fn get_partial_safe(data: *const State, len: usize) -> State {
     // Copy data into the buffer
     std::ptr::copy(data as *const i8, buffer.as_mut_ptr(), len);
     // Load the buffer into a __m256i vector
-    let partial_vector = _mm_loadu_epi8(buffer.as_ptr());
+    let partial_vector = _mm_loadu_si128(buffer.as_ptr() as *const State);
     _mm_add_epi8(partial_vector, _mm_set1_epi8(len as i8))
 }
 

src/gxhash/platform/x86_256.rs

@@ -36,7 +36,7 @@ pub unsafe fn get_partial_safe(data: *const State, len: usize) -> State {
     // Copy data into the buffer
     std::ptr::copy(data as *const i8, buffer.as_mut_ptr(), len);
     // Load the buffer into a __m256i vector
-    let partial_vector = _mm256_loadu_epi8(buffer.as_ptr());
+    let partial_vector = _mm256_loadu_si256(buffer.as_ptr() as *const State);
     _mm256_add_epi8(partial_vector, _mm256_set1_epi8(len as i8))
 }
 

src/lib.rs

@@ -1,7 +1,5 @@
-//#![feature(core_intrinsics)]
-//#![feature(pointer_byte_offsets)]
-#![feature(stdsimd)]
-//#![feature(stmt_expr_attributes)]
+// Feature 'avx2' currently requires unstable 'stdsimd'
+#![cfg_attr(all(feature = "avx2", target_arch = "x86_64"), feature(stdsimd))]
 
 mod gxhash;
 mod hasher;