diff --git a/HISTORY b/HISTORY
index cd480d76..11e069f6 100644
--- a/HISTORY
+++ b/HISTORY
@@ -18,6 +18,12 @@ SnapRAID HISTORY
    intelligently mixes live access for the disks being repaired with
    snapshot access for reference disks to maintain mathematical
    integrity across the array.
+ * Implemented AVX-512BW (Byte and Word) support for parity generation
+   and data recovery, optimizing performance on modern architectures
+   like Intel Sapphire Rapids and AMD Zen 4/5.
+   By scaling to 512-bit ZMM registers, the implementation now processes
+   64 bytes per iteration, effectively doubling the data width of the
+   previous AVX2 routines.
 
 14.10 2026/08
 =============
diff --git a/cmdline/speed.c b/cmdline/speed.c
index 10ae8c8b..fe993bc5 100644
--- a/cmdline/speed.c
+++ b/cmdline/speed.c
@@ -284,6 +284,7 @@ void speed(int period, int nd, int size)
 	printf("%8s", "avx2");
 #ifdef CONFIG_X86_64
 	printf("%8s", "avx2e");
+	printf("%8s", "avx512");
 #endif
 #endif
 	printf("\n");
@@ -338,6 +339,19 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_X86_64
+	printf("%8s", "");
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			raid_gen1_avx512bw(nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
+#endif
 #endif
 	printf("\n");
 
@@ -397,6 +411,19 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_X86_64
+	printf("%8s", "");
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			raid_gen2_avx512bw(nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
+#endif
 #endif
 	printf("\n");
 
@@ -522,6 +549,16 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			raid_gen3_avx512bw(nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
 #endif
 #endif
 	printf("\n");
@@ -587,6 +624,16 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			raid_gen4_avx512bw(nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
 #endif
 #endif
 	printf("\n");
@@ -652,6 +699,16 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			raid_gen5_avx512bw(nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
 #endif
 #endif
 	printf("\n");
@@ -717,6 +774,16 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			raid_gen6_avx512bw(nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
 #endif
 #endif
 	printf("\n");
@@ -730,6 +797,9 @@ void speed(int period, int nd, int size)
 #ifdef CONFIG_X86
 	printf("%8s", "ssse3");
 	printf("%8s", "avx2");
+#ifdef CONFIG_X86_64
+	printf("%8s", "avx512");
+#endif
 #endif
 	printf("\n");
 
@@ -774,6 +844,21 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_X86_64
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			/* ensure to use same hardware in the delta step */
+			raid_gen_force(1, raid_gen1_avx512bw);
+			/* +1 to avoid GEN1 optimized case */
+			raid_rec1_avx512bw(1, id, ip + 1, nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
+#endif
 #endif
 	printf("\n");
 
@@ -814,7 +899,7 @@ void speed(int period, int nd, int size)
 		SPEED_START {
 			/* ensure to use same hardware in the delta step */
 			raid_gen_force(2, raid_gen2_avx2);
-			/* +1 to avoid GEN1 optimized case */
+			/* +1 to avoid GEN2 optimized case */
 			raid_rec2_avx2(2, id, ip + 1, nd, size, v);
 		} SPEED_STOP
 
@@ -822,6 +907,21 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_X86_64
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			/* ensure to use same hardware in the delta step */
+			raid_gen_force(2, raid_gen2_avx512bw);
+			/* +1 to avoid GEN2 optimized case */
+			raid_rec2_avx512bw(2, id, ip + 1, nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
+#endif
 #endif
 	printf("\n");
 
@@ -871,6 +971,21 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_X86_64
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			/* ensure to use same hardware in the delta step */
+			raid_gen_force(3, raid_gen3_avx512bw);
+			/* +1 to avoid GEN1 optimized case */
+			raid_recX_avx512bw(3, id, ip, nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
+#endif
 #endif
 	printf("\n");
 
@@ -921,6 +1036,21 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_X86_64
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			/* ensure to use same hardware in the delta step */
+			raid_gen_force(4, raid_gen4_avx512bw);
+			/* +1 to avoid GEN1 optimized case */
+			raid_recX_avx512bw(4, id, ip, nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
+#endif
 #endif
 	printf("\n");
 
@@ -970,6 +1100,21 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_X86_64
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			/* ensure to use same hardware in the delta step */
+			raid_gen_force(5, raid_gen5_avx512bw);
+			/* +1 to avoid GEN1 optimized case */
+			raid_recX_avx512bw(5, id, ip, nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
+#endif
 #endif
 	printf("\n");
 
@@ -1019,6 +1164,21 @@ void speed(int period, int nd, int size)
 		fflush(stdout);
 	}
 #endif
+#ifdef CONFIG_X86_64
+#ifdef CONFIG_AVX512BW
+	if (raid_cpu_has_avx512bw()) {
+		SPEED_START {
+			/* ensure to use same hardware in the delta step */
+			raid_gen_force(6, raid_gen6_avx512bw);
+			/* +1 to avoid GEN1 optimized case */
+			raid_recX_avx512bw(6, id, ip, nd, size, v);
+		} SPEED_STOP
+
+		printf("%8" PRIu64, ds / dt);
+		fflush(stdout);
+	}
+#endif
+#endif
 #endif
 	printf("\n");
 	printf("\n");