Merge pull request #12 from haskell-works/fix-simd-vector-type

Use correct type __m256i instead of __m256

Author: newhoggy

cbits/simd_avx2.c

@@ -39,15 +39,15 @@ void avx2_cmpeq8(
 #if defined(AVX2_ENABLED)
   uint32_t *target32 = (uint32_t *)target;
 
-  __m256 v_comparand = _mm256_set1_epi8(byte);
+  __m256i v_comparand = _mm256_set1_epi8(byte);
 
   uint32_t *out_mask = (uint32_t*)target;
 
   size_t i;
 
   for (i = 0; i < target_length * 2; ++i) {
-    __m256 v_data_a = *(__m256*)(source + (i * 32));
-    __m256 v_results_a = _mm256_cmpeq_epi8(v_data_a, v_comparand);
+    __m256i v_data_a = *(__m256i *)(source + (i * 32));
+    __m256i v_results_a = _mm256_cmpeq_epi8(v_data_a, v_comparand);
     uint32_t mask = (uint32_t)_mm256_movemask_epi8(v_results_a);
     target32[i] = mask;
   }
@@ -63,10 +63,10 @@ void avx2_and_bits(
   size_t i;
 
   for (i = 0; i < target_length; i += 32) {
-    __m256 v_data_a   = *(__m256*)(source_a + i);
-    __m256 v_data_b   = *(__m256*)(source_b + i);
-    __m256 v_results  = _mm256_and_si256(v_data_a, v_data_b);
-    *(__m256*)(target + i) = v_results;
+    __m256i v_data_a   = *(__m256i *)(source_a + i);
+    __m256i v_data_b   = *(__m256i *)(source_b + i);
+    __m256i v_results  = _mm256_and_si256(v_data_a, v_data_b);
+    *(__m256i *)(target + i) = v_results;
   }
 #endif
 }
@@ -80,10 +80,10 @@ void avx2_and_not_bits(
   size_t i;
 
   for (i = 0; i < target_length; i += 32) {
-    __m256 v_data_a   = *(__m256*)(source_a + i);
-    __m256 v_data_b   = *(__m256*)(source_b + i);
-    __m256 v_results  = _mm256_andnot_si256(v_data_a, v_data_b);
-    *(__m256*)(target + i) = v_results;
+    __m256i v_data_a   = *(__m256i *)(source_a + i);
+    __m256i v_data_b   = *(__m256i *)(source_b + i);
+    __m256i v_results  = _mm256_andnot_si256(v_data_a, v_data_b);
+    *(__m256i *)(target + i) = v_results;
   }
 #endif
 }
@@ -93,14 +93,14 @@ void avx2_not_bits(
     size_t target_length,
     uint8_t *source) {
 #if defined(AVX2_ENABLED)
-  __m256 ones = _mm256_set1_epi8(0xff);
+  __m256i ones = _mm256_set1_epi8(0xff);
 
   size_t i;
 
   for (i = 0; i < target_length; i += 32) {
-    __m256 v_data     = *(__m256*)(source + i);
-    __m256 v_results  = _mm256_xor_si256(v_data, ones);
-    *(__m256*)(target + i) = v_results;
+    __m256i v_data     = *(__m256i *)(source + i);
+    __m256i v_results  = _mm256_xor_si256(v_data, ones);
+    *(__m256i *)(target + i) = v_results;
   }
 #endif
 }
@@ -114,10 +114,10 @@ void avx2_or_bits(
   size_t i;
 
   for (i = 0; i < target_length; i += 32) {
-    __m256 v_data_a   = *(__m256*)(source_a + i);
-    __m256 v_data_b   = *(__m256*)(source_b + i);
-    __m256 v_results  = _mm256_or_si256(v_data_a, v_data_b);
-    *(__m256*)(target + i) = v_results;
+    __m256i v_data_a   = *(__m256i *)(source_a + i);
+    __m256i v_data_b   = *(__m256i *)(source_b + i);
+    __m256i v_results  = _mm256_or_si256(v_data_a, v_data_b);
+    *(__m256i *)(target + i) = v_results;
   }
 #endif
 }
@@ -131,10 +131,10 @@ void avx2_xor_bits(
   size_t i;
 
   for (i = 0; i < target_length; i += 32) {
-    __m256 v_data_a   = *(__m256*)(source_a + i);
-    __m256 v_data_b   = *(__m256*)(source_b + i);
-    __m256 v_results  = _mm256_xor_si256(v_data_a, v_data_b);
-    *(__m256*)(target + i) = v_results;
+    __m256i v_data_a   = *(__m256i *)(source_a + i);
+    __m256i v_data_b   = *(__m256i *)(source_b + i);
+    __m256i v_results  = _mm256_xor_si256(v_data_a, v_data_b);
+    *(__m256i *)(target + i) = v_results;
   }
 #endif
 }

stack.yaml

@@ -5,7 +5,7 @@ packages:
 
 extra-deps:
 - hw-hedgehog-0.1.0.2
-- hw-prim-0.6.2.12
+- hw-prim-0.6.2.14
 - hw-rankselect-0.12.0.3
 - lazy-csv-0.5.1
 - sv-0.1