From f47eba7d7d4c6a1d9501e027b63bbab04bb7d417 Mon Sep 17 00:00:00 2001
From: Takamichi Horikawa <takamichiho@gmail.com>
Date: Thu, 30 Mar 2017 23:31:39 +0000
Subject: add NEON optimization

---
 fmdsp/fmdsp-vramlookup-neon.s | 83 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 83 insertions(+)
 create mode 100644 fmdsp/fmdsp-vramlookup-neon.s

(limited to 'fmdsp/fmdsp-vramlookup-neon.s')

diff --git a/fmdsp/fmdsp-vramlookup-neon.s b/fmdsp/fmdsp-vramlookup-neon.s
new file mode 100644
index 0000000..3cfb957
--- /dev/null
+++ b/fmdsp/fmdsp-vramlookup-neon.s
@@ -0,0 +1,83 @@
+@ neon register map
+@  0,  3,  6,  9, 12, 15, 18, 21  b
+@  1,  4,  7, 10, 13, 16, 19, 22  g
+@  2,  5,  9, 11, 14, 17, 20, 23  r
+
+@ 16, 17, 18, 19, 20, 21, 22, 23  vram
+
+@ 26, 27 r palette
+@ 28, 29 g palette
+@ 30, 31 b palette
+
+.global fmdsp_vramlookup_neon
+
+@ r0: uint8_t *vram32
+@   4 bytes aligned
+@   b, g, r, 0,
+@ r1: const uint8_t *vram
+@ r2: const uint8_t *palette
+@   r0, g0, b0, r1, g1, b1, ...
+@ r3: int stride
+fmdsp_vramlookup_neon:
+  push {lr}
+@ load palette
+  vld3.8 {d26, d28, d30}, [r2]!
+  vld1.8 {d27}, [r2]!
+  vld1.8 {d29}, [r2]!
+  vld1.8 {d31}, [r2]!
+
+  mov r14, #400
+.loopcol:
+  mov r2, r0
+  mov r12, #10
+.looprow:
+@ row address
+
+@ load vram
+  vld1.8 {d16-d19}, [r1]!
+  vld1.8 {d20-d23}, [r1]!
+
+@ lookup
+  vtbl.8 d0,  {d30-d31}, d16
+  vtbl.8 d1,  {d28-d29}, d16
+  vtbl.8 d2,  {d26-d27}, d16
+  vtbl.8 d3,  {d30-d31}, d17
+  vtbl.8 d4,  {d28-d29}, d17
+  vtbl.8 d5,  {d26-d27}, d17
+  vtbl.8 d6,  {d30-d31}, d18
+  vtbl.8 d7,  {d28-d29}, d18
+  vtbl.8 d8,  {d26-d27}, d18
+  vtbl.8 d9,  {d30-d31}, d19
+  vtbl.8 d10, {d28-d29}, d19
+  vtbl.8 d11, {d26-d27}, d19
+  vtbl.8 d12, {d30-d31}, d20
+  vtbl.8 d13, {d28-d29}, d20
+  vtbl.8 d14, {d26-d27}, d20
+  vtbl.8 d15, {d30-d31}, d21
+  vtbl.8 d16, {d28-d29}, d21
+  vtbl.8 d17, {d26-d27}, d21
+  vtbl.8 d18, {d30-d31}, d22
+  vtbl.8 d19, {d28-d29}, d22
+  vtbl.8 d20, {d26-d27}, d22
+  vtbl.8 d21, {d30-d31}, d23
+  vtbl.8 d22, {d28-d29}, d23
+  vtbl.8 d23, {d26-d27}, d23
+
+@ store vram32
+  vst4.8 {d0-d3},   [r2]!
+  vst4.8 {d3-d6},   [r2]!
+  vst4.8 {d6-d9},   [r2]!
+  vst4.8 {d9-d12},  [r2]!
+  vst4.8 {d12-d15}, [r2]!
+  vst4.8 {d15-d18}, [r2]!
+  vst4.8 {d18-d21}, [r2]!
+  vst4.8 {d21-d24}, [r2]!
+
+  subs r12, #1
+  bne .looprow
+
+  add r0, r3
+  subs r14, #1
+  bne .loopcol
+
+  pop {pc}
-- 
cgit v1.2.3