Let simd/x86 pass new transmute/cast vet.

core/simd/x86/sse2.odin

@@ -35,7 +35,7 @@ _mm_add_epi32 :: #force_inline proc "c" (a, b: __m128i)  -> __m128i {
 }
 @(require_results, enable_target_feature="sse2")
 _mm_add_epi64 :: #force_inline proc "c" (a, b: __m128i)  -> __m128i {
-	return transmute(__m128i)simd.add(transmute(i64x2)a, transmute(i64x2)b)
+	return simd.add(a, b)
 }
 @(require_results, enable_target_feature="sse2")
 _mm_adds_epi8 :: #force_inline proc "c" (a, b: __m128i)  -> __m128i {
@@ -118,7 +118,7 @@ _mm_sub_epi32 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
 }
 @(require_results, enable_target_feature="sse2")
 _mm_sub_epi64 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
-	return transmute(__m128i)simd.sub(transmute(i64x2)a, transmute(i64x2)b)
+	return simd.sub(a, b)
 }
 @(require_results, enable_target_feature="sse2")
 _mm_subs_epi8 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
@@ -229,7 +229,7 @@ _mm_slli_epi64 :: #force_inline proc "c" (a: __m128i, $IMM8: u32) -> __m128i {
 }
 @(require_results, enable_target_feature="sse2")
 _mm_sll_epi64 :: #force_inline proc "c" (a, count: __m128i) -> __m128i {
-	return transmute(__m128i)psllq(transmute(i64x2)a, transmute(i64x2)count)
+	return psllq(a, count)
 }
 @(require_results, enable_target_feature="sse2")
 _mm_srai_epi16 :: #force_inline proc "c" (a: __m128i, $IMM8: u32) -> __m128i {
@@ -275,7 +275,7 @@ _mm_srli_epi64 :: #force_inline proc "c" (a: __m128i, $IMM8: u32) -> __m128i {
 }
 @(require_results, enable_target_feature="sse2")
 _mm_srl_epi64 :: #force_inline proc "c" (a, count: __m128i) -> __m128i {
-	return transmute(__m128i)psrlq(transmute(i64x2)a, transmute(i64x2)count)
+	return psrlq(a, count)
 }
 
 
@@ -363,7 +363,7 @@ _mm_cvtsi128_si32 :: #force_inline proc "c" (a: __m128i) -> i32 {
 
 @(require_results, enable_target_feature="sse2")
 _mm_set_epi64x :: #force_inline proc "c" (e1, e0: i64) -> __m128i {
-	return transmute(__m128i)i64x2{e0, e1}
+	return i64x2{e0, e1}
 }
 @(require_results, enable_target_feature="sse2")
 _mm_set_epi32 :: #force_inline proc "c" (e3, e2, e1, e0: i32) -> __m128i {
@@ -453,7 +453,7 @@ _mm_stream_si32 :: #force_inline proc "c" (mem_addr: ^i32, a: i32) {
 @(require_results, enable_target_feature="sse2")
 _mm_move_epi64 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 	zero := _mm_setzero_si128()
-	return transmute(__m128i)simd.shuffle(transmute(i64x2)a, transmute(i64x2)zero, 0, 2)
+	return simd.shuffle(a, zero, 0, 2)
 }
 
 
@@ -545,7 +545,7 @@ _mm_unpackhi_epi32 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
 }
 @(require_results, enable_target_feature="sse2")
 _mm_unpackhi_epi64 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
-	return transmute(__m128i)simd.shuffle(transmute(i64x2)a, transmute(i64x2)b, 1, 3)
+	return simd.shuffle(a, b, 1, 3)
 }
 @(require_results, enable_target_feature="sse2")
 _mm_unpacklo_epi8 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
@@ -565,7 +565,7 @@ _mm_unpacklo_epi32 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
 }
 @(require_results, enable_target_feature="sse2")
 _mm_unpacklo_epi64 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
-	return transmute(__m128i)simd.shuffle(transmute(i64x2)a, transmute(i64x2)b, 0, 2)
+	return simd.shuffle(a, b, 0, 2)
 }
 
 
@@ -1023,7 +1023,7 @@ when ODIN_ARCH == .amd64 {
 	}
 	@(require_results, enable_target_feature="sse2")
 	_mm_cvtsi128_si64 :: #force_inline proc "c" (a: __m128i) -> i64 {
-		return simd.extract(transmute(i64x2)a, 0)
+		return simd.extract(a, 0)
 	}
 	@(require_results, enable_target_feature="sse2")
 	_mm_cvtsi128_si64x :: #force_inline proc "c" (a: __m128i) -> i64 {

core/simd/x86/sse41.odin

@@ -106,7 +106,7 @@ _mm_packus_epi32 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_cmpeq_epi64 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
-	return transmute(__m128i)simd.lanes_eq(transmute(i64x2)a, transmute(i64x2)b)
+	return transmute(__m128i)simd.lanes_eq(a, b)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_cvtepi8_epi16 :: #force_inline proc "c" (a: __m128i) -> __m128i {
@@ -124,7 +124,7 @@ _mm_cvtepi8_epi32 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 _mm_cvtepi8_epi64 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 	x := transmute(i8x16)a
 	y := simd.shuffle(x, x, 0, 1)
-	return transmute(__m128i)i64x2(y)
+	return i64x2(y)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_cvtepi16_epi32 :: #force_inline proc "c" (a: __m128i) -> __m128i {
@@ -136,13 +136,13 @@ _mm_cvtepi16_epi32 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 _mm_cvtepi16_epi64 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 	x := transmute(i16x8)a
 	y := simd.shuffle(x, x, 0, 1)
-	return transmute(__m128i)i64x2(y)
+	return i64x2(y)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_cvtepi32_epi64 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 	x := transmute(i32x4)a
 	y := simd.shuffle(x, x, 0, 1)
-	return transmute(__m128i)i64x2(y)
+	return i64x2(y)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_cvtepu8_epi16 :: #force_inline proc "c" (a: __m128i) -> __m128i {
@@ -160,7 +160,7 @@ _mm_cvtepu8_epi32 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 _mm_cvtepu8_epi64 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 	x := transmute(u8x16)a
 	y := simd.shuffle(x, x, 0, 1)
-	return transmute(__m128i)i64x2(y)
+	return i64x2(y)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_cvtepu16_epi32 :: #force_inline proc "c" (a: __m128i) -> __m128i {
@@ -172,13 +172,13 @@ _mm_cvtepu16_epi32 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 _mm_cvtepu16_epi64 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 	x := transmute(u16x8)a
 	y := simd.shuffle(x, x, 0, 1)
-	return transmute(__m128i)i64x2(y)
+	return i64x2(y)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_cvtepu32_epi64 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 	x := transmute(u32x4)a
 	y := simd.shuffle(x, x, 0, 1)
-	return transmute(__m128i)i64x2(y)
+	return i64x2(y)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_dp_pd :: #force_inline proc "c" (a, b: __m128d, $IMM8: u8) -> __m128d {
@@ -242,7 +242,7 @@ _mm_minpos_epu16 :: #force_inline proc "c" (a: __m128i) -> __m128i {
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_mul_epi32 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
-	return transmute(__m128i)pmuldq(transmute(i32x4)a, transmute(i32x4)b)
+	return pmuldq(transmute(i32x4)a, transmute(i32x4)b)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_mullo_epi32 :: #force_inline proc "c" (a, b: __m128i) -> __m128i {
@@ -254,15 +254,15 @@ _mm_mpsadbw_epu8 :: #force_inline proc "c" (a, b: __m128i, $IMM8: u8) -> __m128i
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_testz_si128 :: #force_inline proc "c" (a: __m128i, mask: __m128i) -> i32 {
-	return ptestz(transmute(i64x2)a, transmute(i64x2)mask)
+	return ptestz(a, mask)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_testc_si128 :: #force_inline proc "c" (a: __m128i, mask: __m128i) -> i32 {
-	return ptestc(transmute(i64x2)a, transmute(i64x2)mask)
+	return ptestc(a, mask)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_testnzc_si128 :: #force_inline proc "c" (a: __m128i, mask: __m128i) -> i32 {
-	return ptestnzc(transmute(i64x2)a, transmute(i64x2)mask)
+	return ptestnzc(a, mask)
 }
 @(require_results, enable_target_feature="sse4.1")
 _mm_test_all_zeros :: #force_inline proc "c" (a: __m128i, mask: __m128i) -> i32 {

core/simd/x86/sse42.odin

@@ -94,7 +94,7 @@ _mm_crc32_u32 :: #force_inline proc "c" (crc: u32, v: u32) -> u32 {
 }
 @(require_results, enable_target_feature="sse4.2")
 _mm_cmpgt_epi64 :: #force_inline proc "c" (a: __m128i, b: __m128i) -> __m128i {
-	return transmute(__m128i)simd.lanes_gt(transmute(i64x2)a, transmute(i64x2)b)
+	return transmute(__m128i)simd.lanes_gt(a, b)
 }
 
 when ODIN_ARCH == .amd64 {

examples/all/all_main.odin

@@ -114,6 +114,7 @@ import relative         "core:relative"
 import reflect          "core:reflect"
 import runtime          "base:runtime"
 import simd             "core:simd"
+import x86              "core:simd/x86"
 import slice            "core:slice"
 import slice_heap       "core:slice/heap"
 import sort             "core:sort"
@@ -234,6 +235,7 @@ _ :: relative
 _ :: reflect
 _ :: runtime
 _ :: simd
+_ :: x86
 _ :: slice
 _ :: slice_heap
 _ :: sort