diff --git a/OpenCL/m08900-pure.cl b/OpenCL/m08900-pure.cl
index 5f96f8ec4..3064263a2 100644
--- a/OpenCL/m08900-pure.cl
+++ b/OpenCL/m08900-pure.cl
@@ -327,6 +327,9 @@ KERNEL_FQ void m08900_init (KERN_ATTR_TMPS (scrypt_tmp_t))
     #if defined IS_CUDA || defined IS_HIP
     const uint4 tmp0 = make_uint4 (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = make_uint4 (digest[4], digest[5], digest[6], digest[7]);
+    #elif defined IS_METAL
+    const uint4 tmp0 = uint4 (digest[0], digest[1], digest[2], digest[3]);
+    const uint4 tmp1 = uint4 (digest[4], digest[5], digest[6], digest[7]);
     #else
     const uint4 tmp0 = (uint4) (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = (uint4) (digest[4], digest[5], digest[6], digest[7]);
@@ -357,6 +360,11 @@ KERNEL_FQ void m08900_init (KERN_ATTR_TMPS (scrypt_tmp_t))
     X[1] = make_uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
     X[2] = make_uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
     X[3] = make_uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
+    #elif defined IS_METAL
+    X[0] = uint4 (T[0].x, T[1].y, T[2].z, T[3].w);
+    X[1] = uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
+    X[2] = uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
+    X[3] = uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
     #else
     X[0] = (uint4) (T[0].x, T[1].y, T[2].z, T[3].w);
     X[1] = (uint4) (T[1].x, T[2].y, T[3].z, T[0].w);
@@ -467,6 +475,11 @@ KERNEL_FQ void m08900_comp (KERN_ATTR_TMPS (scrypt_tmp_t))
     T[1] = make_uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
     T[2] = make_uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
     T[3] = make_uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
+    #elif defined IS_METAL
+    T[0] = uint4 (X[0].x, X[3].y, X[2].z, X[1].w);
+    T[1] = uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
+    T[2] = uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
+    T[3] = uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
     #else
     T[0] = (uint4) (X[0].x, X[3].y, X[2].z, X[1].w);
     T[1] = (uint4) (X[1].x, X[0].y, X[3].z, X[2].w);
diff --git a/OpenCL/m15700-pure.cl b/OpenCL/m15700-pure.cl
index 20cab27d1..e500b4f70 100644
--- a/OpenCL/m15700-pure.cl
+++ b/OpenCL/m15700-pure.cl
@@ -463,6 +463,9 @@ KERNEL_FQ void m15700_init (KERN_ATTR_TMPS_ESALT (scrypt_tmp_t, ethereum_scrypt_
     #if defined IS_CUDA || defined IS_HIP
     const uint4 tmp0 = make_uint4 (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = make_uint4 (digest[4], digest[5], digest[6], digest[7]);
+    #elif defined IS_METAL
+    const uint4 tmp0 = uint4 (digest[0], digest[1], digest[2], digest[3]);
+    const uint4 tmp1 = uint4 (digest[4], digest[5], digest[6], digest[7]);
     #else
     const uint4 tmp0 = (uint4) (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = (uint4) (digest[4], digest[5], digest[6], digest[7]);
@@ -493,6 +496,11 @@ KERNEL_FQ void m15700_init (KERN_ATTR_TMPS_ESALT (scrypt_tmp_t, ethereum_scrypt_
     X[1] = make_uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
     X[2] = make_uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
     X[3] = make_uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
+    #elif defined IS_METAL
+    X[0] = uint4 (T[0].x, T[1].y, T[2].z, T[3].w);
+    X[1] = uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
+    X[2] = uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
+    X[3] = uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
     #else
     X[0] = (uint4) (T[0].x, T[1].y, T[2].z, T[3].w);
     X[1] = (uint4) (T[1].x, T[2].y, T[3].z, T[0].w);
@@ -603,6 +611,11 @@ KERNEL_FQ void m15700_comp (KERN_ATTR_TMPS_ESALT (scrypt_tmp_t, ethereum_scrypt_
     T[1] = make_uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
     T[2] = make_uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
     T[3] = make_uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
+    #elif defined IS_METAL
+    T[0] = uint4 (X[0].x, X[3].y, X[2].z, X[1].w);
+    T[1] = uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
+    T[2] = uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
+    T[3] = uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
     #else
     T[0] = (uint4) (X[0].x, X[3].y, X[2].z, X[1].w);
     T[1] = (uint4) (X[1].x, X[0].y, X[3].z, X[2].w);
diff --git a/OpenCL/m22700-pure.cl b/OpenCL/m22700-pure.cl
index 7d65fd52e..a29df1c03 100644
--- a/OpenCL/m22700-pure.cl
+++ b/OpenCL/m22700-pure.cl
@@ -400,6 +400,9 @@ KERNEL_FQ void m22700_init (KERN_ATTR_TMPS (scrypt_tmp_t))
     #if defined IS_CUDA || defined IS_HIP
     const uint4 tmp0 = make_uint4 (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = make_uint4 (digest[4], digest[5], digest[6], digest[7]);
+    #elif defined IS_METAL
+    const uint4 tmp0 = uint4 (digest[0], digest[1], digest[2], digest[3]);
+    const uint4 tmp1 = uint4 (digest[4], digest[5], digest[6], digest[7]);
     #else
     const uint4 tmp0 = (uint4) (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = (uint4) (digest[4], digest[5], digest[6], digest[7]);
@@ -430,6 +433,11 @@ KERNEL_FQ void m22700_init (KERN_ATTR_TMPS (scrypt_tmp_t))
     X[1] = make_uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
     X[2] = make_uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
     X[3] = make_uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
+    #elif defined IS_METAL
+    X[0] = uint4 (T[0].x, T[1].y, T[2].z, T[3].w);
+    X[1] = uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
+    X[2] = uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
+    X[3] = uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
     #else
     X[0] = (uint4) (T[0].x, T[1].y, T[2].z, T[3].w);
     X[1] = (uint4) (T[1].x, T[2].y, T[3].z, T[0].w);
@@ -605,6 +613,11 @@ KERNEL_FQ void m22700_comp (KERN_ATTR_TMPS (scrypt_tmp_t))
     T[1] = make_uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
     T[2] = make_uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
     T[3] = make_uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
+    #elif defined IS_METAL
+    T[0] = uint4 (X[0].x, X[3].y, X[2].z, X[1].w);
+    T[1] = uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
+    T[2] = uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
+    T[3] = uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
     #else
     T[0] = (uint4) (X[0].x, X[3].y, X[2].z, X[1].w);
     T[1] = (uint4) (X[1].x, X[0].y, X[3].z, X[2].w);
diff --git a/OpenCL/m27700-pure.cl b/OpenCL/m27700-pure.cl
index feaca11de..c62dc90d6 100644
--- a/OpenCL/m27700-pure.cl
+++ b/OpenCL/m27700-pure.cl
@@ -351,6 +351,9 @@ KERNEL_FQ void m27700_init (KERN_ATTR_TMPS (scrypt_tmp_t))
     #if defined IS_CUDA || defined IS_HIP
     const uint4 tmp0 = make_uint4 (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = make_uint4 (digest[4], digest[5], digest[6], digest[7]);
+    #elif defined IS_METAL
+    const uint4 tmp0 = uint4 (digest[0], digest[1], digest[2], digest[3]);
+    const uint4 tmp1 = uint4 (digest[4], digest[5], digest[6], digest[7]);
     #else
     const uint4 tmp0 = (uint4) (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = (uint4) (digest[4], digest[5], digest[6], digest[7]);
@@ -381,6 +384,11 @@ KERNEL_FQ void m27700_init (KERN_ATTR_TMPS (scrypt_tmp_t))
     X[1] = make_uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
     X[2] = make_uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
     X[3] = make_uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
+    #elif defined IS_METAL
+    X[0] = uint4 (T[0].x, T[1].y, T[2].z, T[3].w);
+    X[1] = uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
+    X[2] = uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
+    X[3] = uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
     #else
     X[0] = (uint4) (T[0].x, T[1].y, T[2].z, T[3].w);
     X[1] = (uint4) (T[1].x, T[2].y, T[3].z, T[0].w);
@@ -557,6 +565,11 @@ KERNEL_FQ void m27700_comp (KERN_ATTR_TMPS (scrypt_tmp_t))
     T[1] = make_uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
     T[2] = make_uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
     T[3] = make_uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
+    #elif defined IS_METAL
+    T[0] = uint4 (X[0].x, X[3].y, X[2].z, X[1].w);
+    T[1] = uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
+    T[2] = uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
+    T[3] = uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
     #else
     T[0] = (uint4) (X[0].x, X[3].y, X[2].z, X[1].w);
     T[1] = (uint4) (X[1].x, X[0].y, X[3].z, X[2].w);
diff --git a/OpenCL/m28200-pure.cl b/OpenCL/m28200-pure.cl
index 2ead5b4bc..2260e931b 100644
--- a/OpenCL/m28200-pure.cl
+++ b/OpenCL/m28200-pure.cl
@@ -337,6 +337,9 @@ KERNEL_FQ void m28200_init (KERN_ATTR_TMPS_ESALT (exodus_tmp_t, exodus_t))
     #if defined IS_CUDA || defined IS_HIP
     const uint4 tmp0 = make_uint4 (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = make_uint4 (digest[4], digest[5], digest[6], digest[7]);
+    #elif defined IS_METAL
+    const uint4 tmp0 = uint4 (digest[0], digest[1], digest[2], digest[3]);
+    const uint4 tmp1 = uint4 (digest[4], digest[5], digest[6], digest[7]);
     #else
     const uint4 tmp0 = (uint4) (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = (uint4) (digest[4], digest[5], digest[6], digest[7]);
@@ -367,6 +370,11 @@ KERNEL_FQ void m28200_init (KERN_ATTR_TMPS_ESALT (exodus_tmp_t, exodus_t))
     X[1] = make_uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
     X[2] = make_uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
     X[3] = make_uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
+    #elif defined IS_METAL
+    X[0] = uint4 (T[0].x, T[1].y, T[2].z, T[3].w);
+    X[1] = uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
+    X[2] = uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
+    X[3] = uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
     #else
     X[0] = (uint4) (T[0].x, T[1].y, T[2].z, T[3].w);
     X[1] = (uint4) (T[1].x, T[2].y, T[3].z, T[0].w);
@@ -525,6 +533,11 @@ KERNEL_FQ void m28200_comp (KERN_ATTR_TMPS_ESALT (exodus_tmp_t, exodus_t))
     T[1] = make_uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
     T[2] = make_uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
     T[3] = make_uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
+    #elif defined IS_METAL
+    T[0] = uint4 (X[0].x, X[3].y, X[2].z, X[1].w);
+    T[1] = uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
+    T[2] = uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
+    T[3] = uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
     #else
     T[0] = (uint4) (X[0].x, X[3].y, X[2].z, X[1].w);
     T[1] = (uint4) (X[1].x, X[0].y, X[3].z, X[2].w);
diff --git a/OpenCL/m29800-pure.cl b/OpenCL/m29800-pure.cl
index de4260279..c72f5c267 100644
--- a/OpenCL/m29800-pure.cl
+++ b/OpenCL/m29800-pure.cl
@@ -351,6 +351,9 @@ KERNEL_FQ void m29800_init (KERN_ATTR_TMPS (scrypt_tmp_t))
     #if defined IS_CUDA || defined IS_HIP
     const uint4 tmp0 = make_uint4 (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = make_uint4 (digest[4], digest[5], digest[6], digest[7]);
+    #elif defined IS_METAL
+    const uint4 tmp0 = uint4 (digest[0], digest[1], digest[2], digest[3]);
+    const uint4 tmp1 = uint4 (digest[4], digest[5], digest[6], digest[7]);
     #else
     const uint4 tmp0 = (uint4) (digest[0], digest[1], digest[2], digest[3]);
     const uint4 tmp1 = (uint4) (digest[4], digest[5], digest[6], digest[7]);
@@ -381,6 +384,11 @@ KERNEL_FQ void m29800_init (KERN_ATTR_TMPS (scrypt_tmp_t))
     X[1] = make_uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
     X[2] = make_uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
     X[3] = make_uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
+    #elif defined IS_METAL
+    X[0] = uint4 (T[0].x, T[1].y, T[2].z, T[3].w);
+    X[1] = uint4 (T[1].x, T[2].y, T[3].z, T[0].w);
+    X[2] = uint4 (T[2].x, T[3].y, T[0].z, T[1].w);
+    X[3] = uint4 (T[3].x, T[0].y, T[1].z, T[2].w);
     #else
     X[0] = (uint4) (T[0].x, T[1].y, T[2].z, T[3].w);
     X[1] = (uint4) (T[1].x, T[2].y, T[3].z, T[0].w);
@@ -557,6 +565,11 @@ KERNEL_FQ void m29800_comp (KERN_ATTR_TMPS (scrypt_tmp_t))
     T[1] = make_uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
     T[2] = make_uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
     T[3] = make_uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
+    #elif defined IS_METAL
+    T[0] = uint4 (X[0].x, X[3].y, X[2].z, X[1].w);
+    T[1] = uint4 (X[1].x, X[0].y, X[3].z, X[2].w);
+    T[2] = uint4 (X[2].x, X[1].y, X[0].z, X[3].w);
+    T[3] = uint4 (X[3].x, X[2].y, X[1].z, X[0].w);
     #else
     T[0] = (uint4) (X[0].x, X[3].y, X[2].z, X[1].w);
     T[1] = (uint4) (X[1].x, X[0].y, X[3].z, X[2].w);
diff --git a/docs/changes.txt b/docs/changes.txt
index 90e8a6f0e..c977aaf87 100644
--- a/docs/changes.txt
+++ b/docs/changes.txt
@@ -70,6 +70,7 @@
 - Fixed bug in 29600 module OPTS_TYPE setting
 - Fixed bug in grep out-of-memory workaround on Unit Test
 - Fixed bug in input_tokenizer when TOKEN_ATTR_FIXED_LENGTH is used and refactor modules
+- Fixed a bug in all SCRYPT-based hash modes with Apple Metal
 - Added verification of token buffer length when using TOKEN_ATTR_FIXED_LENGTH
 - Fixed build failed for 4410 with vector width > 1
 - Fixed build failed for 10700 optimized with Apple Metal