x86: paddq

README.adoc

@@ -11493,7 +11493,7 @@ One "downside" of glibc is that it exercises much more kernel functionality on i
 
 == C
 
-Programs under link:userland/c/[] are examples of link:https://en.wikipedia.org/wiki/ANSI_C[ANSI C] programming.
+Programs under link:userland/c/[] are examples of link:https://en.wikipedia.org/wiki/ANSI_C[ANSI C] programming:
 
 * Standard library
 ** assert.h
@@ -11541,19 +11541,19 @@ What is POSIX:
 
 == Userland assembly
 
-Programs under `userland/arch/<arch>/` are examples of userland assembly programming:
+Programs under `userland/arch/<arch>/` are examples of userland assembly programming.
 
-* link:userland/arch/x86_64/[] moved from: https://github.com/cirosantilli/x86-assembly-cheat
-* link:userland/arch/arm/[] moved from: https://github.com/cirosantilli/arm-assembly-cheat
-* link:userland/arch/aarch64/[] moved from: https://github.com/cirosantilli/arm-assembly-cheat
+This section will document ISA agnostic concepts.
+
+ISA specifics are covered at:
+
+* <<x86-userland-assembly>> under link:userland/arch/x86_64/[], originally migrated from: https://github.com/cirosantilli/x86-assembly-cheat
+* <<arm-userland-assembly>> under originally migrated from https://github.com/cirosantilli/arm-assembly-cheat under:
+** link:userland/arch/arm/[]
+** link:userland/arch/aarch64/[]
 
 Like other userland programs, these programs can be run as explained at: <<userland-setup>>.
 
-This section will document ISA generic ideas. ISA specifics are documented on the following sections:
-
-* <<x86-userland-assembly>>
-* <<arm-userland-assembly>>
-
 The first example that you want to run for each arch is:
 
 ....
@@ -11836,7 +11836,10 @@ Once those are done, everything else left on userland is just to learn a huge li
 
 === x86 userland assembly instructions
 
-TODO
+==== x86 SIMD
+
+* SVE2
+** link:userland/arch/x86_64/paddq.S[]
 
 === rdtsc
 

userland/arch/aarch64/simd.S

@@ -13,7 +13,7 @@ ENTRY
 .data
     u32_0:          .word 0xF111F111, 0xF222F222, 0xF333F333, 0xF444F444
     u32_1:          .word 0x15551555, 0x16661666, 0x17771777, 0x18881888
-    u32_sum_expect: .word 0x06670666, 0x08890888, 0x0AAB0AAA, 0x0CCD0CCC
+    u32_sum_expect: .word 0x06670666, 0x08890889, 0x0AAB0AAA, 0x0CCD0CCC
 .bss
     u32_sum:        .skip 16
 .text

userland/arch/main.c

@@ -39,7 +39,7 @@ int assert_memcmp(const void *s1, const void *s2, size_t n) {
             printf(
                 "%s failed: "
                 "byte1, byte2, index: "
-                "0x%02" PRIX8 " 0x%02" PRIX8 " 0x%zx\n",
+                "0x%02" PRIX8 " 0x%02" PRIX8 " 0x%zX\n",
                 __func__,
                 b1,
                 b2,

userland/arch/x86_64/paddq.S

@@ -0,0 +1,21 @@
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#x86-simd
+ *
+ * Add 4 32-bit integeres in one go.
+ */
+
+#include "common.h"
+
+ENTRY
+.data
+    u32_0:      .long 0xF111F111, 0xF222F222, 0xF333F333, 0xF444F444
+    u32_1:      .long 0x15551555, 0x16661666, 0x17771777, 0x18881888
+    u32_expect: .long 0x06670666, 0x08890889, 0x0AAB0AAA, 0x0CCD0CCD
+.bss
+    u32_result: .skip 16
+.text
+    movups u32_0, %xmm0
+    movups u32_1, %xmm1
+    paddq %xmm1, %xmm0
+    movups %xmm0, u32_result
+    ASSERT_MEMCMP(u32_result, u32_expect, $0x10)
+EXIT