Add vbitsel

README.md

@@ -1,51 +1,17 @@
-# LoongArch SIMD 指令非官网文档
+# Unofficial LoongArch Intrinsics Guide
 
-从 QEMU 实现中整理。
+Visit <http://jia.je/unofficial-loongarch-intrinsics-guide/>
 
-GCC Intrinsic: <https://gcc.gnu.org/onlinedocs/gcc/LoongArch-SX-Vector-Intrinsics.html>
+Arranged from QEMU implementation and [GCC Intrinsics](https://gcc.gnu.org/onlinedocs/gcc/LoongArch-SX-Vector-Intrinsics.html)
 
-## LSX
-
-### vfmadd.s/d
-
-Vector Fused Multiply Add (`a * b + c`) Single/Double Precision
-
-- Single Precision: `__m128 __lsx_vfmadd_s (__m128 a, __m128 b, __m128 c)`
-- Double Precision: `__m128d __lsx_vfmadd_d (__m128d a, __m128d b, __m128d c)`
-
-### vfmsub.s/d
-
-Vector Fused Multiply Subtract (`a * b - c`) Single/Double Precision
-
-- Single Precision: `__m128 __lsx_vfmsub_s (__m128 a, __m128 b, __m128 c)`
-- Double Precision: `__m128d __lsx_vfmsub_d (__m128d a, __m128d b, __m128d c)`
-
-### vfnmadd.s/d
-
-Vector Fused Negative Multiply Add (`- a * b - c`) Single/Double Precision
-
-- Single Precision: `__m128 __lsx_vfnmadd_s (__m128 a, __m128 b, __m128 c)`
-- Double Precision: `__m128d __lsx_vfnmadd_d (__m128d a, __m128d b, __m128d c)`
-
-### vfnmsub.s/d
-
-Vector Fused Negative Multiply Subtract (`- a * b + c`) Single/Double Precision
-
-- Single Precision: `__m128 __lsx_vfnmsub_s (__m128 a, __m128 b, __m128 c)`
-- Double Precision: `__m128d __lsx_vfnmsub_d (__m128d a, __m128d b, __m128d c)`
+TODO List:
 
 ### vfcmp.cond.s/d
 
 Vector Float Compare
 
 <cmp>: See fcmp.cond.s/d instruction
 
-### vbitsel.v
-
-Vector Bit Selection: For each bit, `c ? b : a`
-
-`__m128i __lsx_vbitsel_v (__m128i a, __m128i b, __m128i c)`
-
 ### vshuf.b
 
 Vector Shuffle

docs/lsx_bitops/vbitsel.md

@@ -0,0 +1,24 @@
+# Bitwise Selection
+
+## __m128d __lsx_vfnmsub_d (__m128d a, __m128d b, __m128d c)
+
+### Synopsis
+
+```c++
+__m128i __lsx_vbitsel_v (__m128i a, __m128i b, __m128i c)
+#include <lsxintrin.h>
+Instruction: vbitsel.v vr, vr, vr
+CPU Flags: LSX
+```
+
+### Description
+
+Compute bitwise selection: for each bit position, if the bit in `c` equals to one, copy the bit from `b` to `dst`, otherwise copy from `a`.
+
+### Operation
+
+```c++
+for (int i = 0;i < 128;i++) {
+    dst.bit[i] = a.bit[i] ? b.bit[i] : c.bit[i];
+}
+```