diff --git a/nimsimd.nimble b/nimsimd.nimble
index 0bfdbd4..04093b0 100644
--- a/nimsimd.nimble
+++ b/nimsimd.nimble
@@ -1,4 +1,4 @@
-version     = "1.3.0"
+version     = "1.3.1"
 author      = "Ryan Oldenburg"
 description = "Pleasant Nim bindings for SIMD instruction sets."
 license     = "MIT"
diff --git a/src/nimsimd/avx.nim b/src/nimsimd/avx.nim
index cd8c226..eb53ec3 100644
--- a/src/nimsimd/avx.nim
+++ b/src/nimsimd/avx.nim
@@ -3,9 +3,9 @@ import sse42
 export sse42
 
 type
-  M256* {.importc: "__m256", header: "immintrin.h".} = object
-  M256i* {.importc: "__m256i", header: "immintrin.h".} = object
-  M256d* {.importc: "__m256d", header: "immintrin.h".} = object
+  M256* {.importc: "__m256", header: "immintrin.h", bycopy.} = object
+  M256i* {.importc: "__m256i", header: "immintrin.h", bycopy.} = object
+  M256d* {.importc: "__m256d", header: "immintrin.h", bycopy.} = object
 
 {.push header: "immintrin.h".}
 
@@ -157,9 +157,15 @@ func mm256_load_si256*(p: pointer): M256i {.importc: "_mm256_load_si256".}
 
 func mm256_loadu_pd*(p: pointer): M256d {.importc: "_mm256_loadu_pd".}
 
-func mm256_loadu_ps*(p: pointer): M256 {.importc: "_mm256_loadu_ps".}
+func mm256_loadu_ps(p: ptr float32): M256 {.importc: "_mm256_loadu_ps".}
 
-func mm256_loadu_si256*(p: pointer): M256i {.importc: "_mm256_loadu_si256".}
+template mm256_loadu_ps*(p: pointer): M256 =
+  mm256_loadu_ps(cast[ptr float32](p))
+
+func mm256_loadu_si256(p: ptr M256i): M256i {.importc: "_mm256_loadu_si256".}
+
+template mm256_loadu_si256*(p: pointer): M256i =
+  mm256_loadu_si256(cast[ptr M256i](p))
 
 func mm256_loadu2_m128*(hi, lo: pointer): M256 {.importc: "_mm256_loadu2_m128".}
 
@@ -309,9 +315,15 @@ func mm256_store_si256*(p: pointer, a: M256i) {.importc: "_mm256_store_si256".}
 
 func mm256_storeu_pd*(p: pointer, a: M256d) {.importc: "_mm256_storeu_pd".}
 
-func mm256_storeu_ps*(p: pointer, a: M256) {.importc: "_mm256_storeu_ps".}
+func mm256_storeu_ps(p: ptr float32, a: M256) {.importc: "_mm256_storeu_ps".}
+
+template mm256_storeu_ps*(p: pointer, a: M256) =
+  mm256_storeu_ps(cast[ptr float32](p), a)
+
+func mm256_storeu_si256(p: ptr M256i, a: M256i) {.importc: "_mm256_storeu_si256".}
 
-func mm256_storeu_si256*(p: pointer, a: M256i) {.importc: "_mm256_storeu_si256".}
+template mm256_storeu_si256*(p: pointer, a: M256i) =
+  mm256_storeu_si256(cast[ptr M256i](p), a)
 
 func mm256_storeu2_m128*(hi, lo: pointer) {.importc: "_mm256_storeu2_m128".}
 
diff --git a/src/nimsimd/avx2.nim b/src/nimsimd/avx2.nim
index 885d49f..1aff039 100644
--- a/src/nimsimd/avx2.nim
+++ b/src/nimsimd/avx2.nim
@@ -164,11 +164,17 @@ func mm256_madd_epi16*(a, b: M256i): M256i {.importc: "_mm256_madd_epi16".}
 
 func mm256_maddubs_epi16*(a, b: M256i): M256i {.importc: "_mm256_maddubs_epi16".}
 
-func mm256_maskload_epi32*(p: pointer, mask: M256i): M256i {.importc: "_mm256_maskload_epi32".}
+func mm256_maskload_epi32(p: ptr int32, mask: M256i): M256i {.importc: "_mm256_maskload_epi32".}
+
+template mm256_maskload_epi32*(p: pointer, mask: M256i): M256i =
+  mm256_maskload_epi32(cast[ptr int32](p), mask)
 
 func mm256_maskload_epi64*(p: pointer, mask: M256i): M256i {.importc: "_mm256_maskload_epi64".}
 
-func mm256_maskstore_epi32*(p: pointer, mask, a: M256i) {.importc: "_mm256_maskstore_epi32".}
+func mm256_maskstore_epi32(p: ptr int32, mask, a: M256i) {.importc: "_mm256_maskstore_epi32".}
+
+template mm256_maskstore_epi32*(p: pointer, mask, a: M256i) =
+  mm256_maskstore_epi32(cast[ptr int32](p), mask, a)
 
 func mm256_maskstore_epi64*(p: pointer, mask, a: M256i) {.importc: "_mm256_maskstore_epi64".}
 
diff --git a/src/nimsimd/runtimecheck.nim b/src/nimsimd/runtimecheck.nim
index 69cf484..8fe32df 100644
--- a/src/nimsimd/runtimecheck.nim
+++ b/src/nimsimd/runtimecheck.nim
@@ -12,10 +12,10 @@ when defined(amd64):
       PCLMULQDQ
       SHA
       AES
-      CMPXCHG16B # Atomic CompareExchange 16-byte, avail. since Haswell
+      CMPXCHG16B
       F16C
-      BM1
-      BM2
+      BMI1
+      BMI2
 
     InstructionSetCheckInfo = object
       leaf, register, bit: int
diff --git a/src/nimsimd/sse2.nim b/src/nimsimd/sse2.nim
index 1fa43be..1f00539 100644
--- a/src/nimsimd/sse2.nim
+++ b/src/nimsimd/sse2.nim
@@ -1,9 +1,9 @@
 ## SSE and SSE2 intrinsics
 
 type
-  M128* {.importc: "__m128", header: "xmmintrin.h".} = object
-  M128i* {.importc: "__m128i", header: "emmintrin.h".} = object
-  M128d* {.importc: "__m128d", header: "emmintrin.h".} = object
+  M128* {.importc: "__m128", header: "xmmintrin.h", bycopy.} = object
+  M128i* {.importc: "__m128i", header: "emmintrin.h", bycopy.} = object
+  M128d* {.importc: "__m128d", header: "emmintrin.h", bycopy.} = object
 
 template MM_SHUFFLE*(z, y, x, w: int | uint): int32 =
   ((z shl 6) or (y shl 4) or (x shl 2) or w).int32
@@ -191,7 +191,10 @@ func mm_store1_ps*(p: pointer, a: M128) {.importc: "_mm_store1_ps".}
 
 func mm_storer_ps*(p: pointer, a: M128) {.importc: "_mm_storer_ps".}
 
-func mm_storeu_ps*(p: pointer, a: M128) {.importc: "_mm_storeu_ps".}
+func mm_storeu_ps(p: ptr float32, a: M128) {.importc: "_mm_storeu_ps".}
+
+template mm_storeu_ps*(p: pointer, a: M128) =
+  mm_storeu_ps(cast[ptr float32](p), a)
 
 func mm_storeu_si16*(p: pointer, a: M128) {.importc: "_mm_storeu_si16".}
 
@@ -419,13 +422,19 @@ func mm_load_pd1*(p: pointer): M128d {.importc: "_mm_load_pd1".}
 
 func mm_load_sd*(p: pointer): M128d {.importc: "_mm_load_sd".}
 
-func mm_load_si128*(p: pointer): M128i {.importc: "_mm_load_si128".}
+func mm_load_si128(p: ptr M128i): M128i {.importc: "_mm_load_si128".}
+
+template mm_load_si128*(p: pointer): M128i =
+  mm_load_si128(cast[ptr M128i](p))
 
 func mm_load1_pd*(p: pointer): M128d {.importc: "_mm_load1_pd".}
 
 func mm_loadh_pd*(a: M128d, p: pointer): M128d {.importc: "_mm_loadh_pd".}
 
-func mm_loadl_epi64*(p: pointer): M128i {.importc: "_mm_loadl_epi64".}
+func mm_loadl_epi64(p: ptr M128i): M128i {.importc: "_mm_loadl_epi64".}
+
+template mm_loadl_epi64*(p: pointer): M128i =
+  mm_loadl_epi64(cast[ptr M128i](p))
 
 func mm_loadl_pd*(a: M128d, p: pointer): M128i {.importc: "_mm_loadl_pd".}
 
@@ -435,7 +444,10 @@ func mm_loadr_pd*(p: pointer): M128d {.importc: "_mm_loadr_pd".}
 
 func mm_loadu_pd*(p: pointer): M128d {.importc: "_mm_loadu_pd".}
 
-func mm_loadu_si128*(p: pointer): M128i {.importc: "_mm_loadu_si128".}
+func mm_loadu_si128(p: ptr M128i): M128i {.importc: "_mm_loadu_si128".}
+
+template mm_loadu_si128*(p: pointer): M128i =
+  mm_loadu_si128(cast[ptr M128i](p))
 
 func mm_loadu_si32*(p: pointer): M128i {.importc: "_mm_loadu_si32".}
 
@@ -595,7 +607,10 @@ func mm_storer_pd*(p: pointer, a: M128d) {.importc: "_mm_storer_pd".}
 
 func mm_storeu_pd*(p: pointer, a: M128d) {.importc: "_mm_storeu_pd".}
 
-func mm_storeu_si128*(p: pointer, a: M128i) {.importc: "_mm_storeu_si128".}
+func mm_storeu_si128(p: ptr M128i, a: M128i) {.importc: "_mm_storeu_si128".}
+
+template mm_storeu_si128*(p: pointer, a: M128i) =
+  mm_storeu_si128(cast[ptr M128i](p), a)
 
 func mm_storeu_si32*(p: pointer, a: M128i) {.importc: "_mm_storeu_si32".}