MyouProject/pixie
3
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

wider neon

Browse source
This commit is contained in:
Ryan Oldenburg 2022-07-31 00:22:17 -05:00
parent 6fb4b02eb6
commit db4c1bea73
1 changed files with 43 additions and 25 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

68
src/pixie/simd/neon.nim
View file

@ -416,34 +416,43 @@ proc blitLineNormalNeon*(
a[i] = blendNormal(a[i], b[i]) a[i] = blendNormal(a[i], b[i])
inc i inc i
let vec255 = vmov_n_u8(255) let vec255 = vmovq_n_u8(255)
while i < len - 8: while i < len - 16:
let let
source = vld4_u8(b[i].addr) source = vld4q_u8(b[i].addr)
eq255 = vceq_u8(source.val[3], vec255) eq255 = vceqq_u8(source.val[3], vec255)
if vget_lane_u64(cast[uint64x1](eq255), 0) == uint64.high: mask = vget_lane_u64(cast[uint64x1](
vst4_u8(a[i].addr, source) vand_u8(vget_low_u8(eq255), vget_high_u8(eq255)
)), 0)
if mask == uint64.high:
vst4q_u8(a[i].addr, source)
else: else:
template multiply(c, a: uint8x8): uint8x8 = template multiply(c, a: uint8x8): uint8x8 =
let ca = vmull_u8(c, a) let ca = vmull_u8(c, a)
vraddhn_u16(ca, vrshrq_n_u16(ca, 8)) vraddhn_u16(ca, vrshrq_n_u16(ca, 8))
let template multiply(c, a: uint8x16): uint8x16 =
backdrop = vld4_u8(a[i].addr) vcombine_u8(
multiplier = vsub_u8(vec255, source.val[3]) multiply(vget_low_u8(c), vget_low_u8(a)),
multiply(vget_high_u8(c), vget_high_u8(a))
)
var blended: uint8x8x4 let
backdrop = vld4q_u8(a[i].addr)
multiplier = vsubq_u8(vec255, source.val[3])
var blended: uint8x16x4
blended.val[0] = multiply(backdrop.val[0], multiplier) blended.val[0] = multiply(backdrop.val[0], multiplier)
blended.val[1] = multiply(backdrop.val[1], multiplier) blended.val[1] = multiply(backdrop.val[1], multiplier)
blended.val[2] = multiply(backdrop.val[2], multiplier) blended.val[2] = multiply(backdrop.val[2], multiplier)
blended.val[3] = multiply(backdrop.val[3], multiplier) blended.val[3] = multiply(backdrop.val[3], multiplier)
blended.val[0] = vadd_u8(blended.val[0], source.val[0]) blended.val[0] = vaddq_u8(blended.val[0], source.val[0])
blended.val[1] = vadd_u8(blended.val[1], source.val[1]) blended.val[1] = vaddq_u8(blended.val[1], source.val[1])
blended.val[2] = vadd_u8(blended.val[2], source.val[2]) blended.val[2] = vaddq_u8(blended.val[2], source.val[2])
blended.val[3] = vadd_u8(blended.val[3], source.val[3]) blended.val[3] = vaddq_u8(blended.val[3], source.val[3])
vst4_u8(a[i].addr, blended) vst4q_u8(a[i].addr, blended)
i += 8 i += 16
for i in i ..< len: for i in i ..< len:
a[i] = blendNormal(a[i], b[i]) a[i] = blendNormal(a[i], b[i])
@ -456,27 +465,36 @@ proc blitLineMaskNeon*(
a[i] = blendMask(a[i], b[i]) a[i] = blendMask(a[i], b[i])
inc i inc i
let vec255 = vmov_n_u8(255) let vec255 = vmovq_n_u8(255)
while i < len - 8: while i < len - 16:
let let
source = vld4_u8(b[i].addr) source = vld4q_u8(b[i].addr)
eq255 = vceq_u8(source.val[3], vec255) eq255 = vceqq_u8(source.val[3], vec255)
if vget_lane_u64(cast[uint64x1](eq255), 0) == uint64.high: mask = vget_lane_u64(cast[uint64x1](
vand_u8(vget_low_u8(eq255), vget_high_u8(eq255)
)), 0)
if mask == uint64.high:
discard discard
else: else:
template multiply(c, a: uint8x8): uint8x8 = template multiply(c, a: uint8x8): uint8x8 =
let ca = vmull_u8(c, a) let ca = vmull_u8(c, a)
vraddhn_u16(ca, vrshrq_n_u16(ca, 8)) vraddhn_u16(ca, vrshrq_n_u16(ca, 8))
let backdrop = vld4_u8(a[i].addr) template multiply(c, a: uint8x16): uint8x16 =
var blended: uint8x8x4 vcombine_u8(
multiply(vget_low_u8(c), vget_low_u8(a)),
multiply(vget_high_u8(c), vget_high_u8(a))
)
let backdrop = vld4q_u8(a[i].addr)
var blended: uint8x16x4
blended.val[0] = multiply(backdrop.val[0], source.val[3]) blended.val[0] = multiply(backdrop.val[0], source.val[3])
blended.val[1] = multiply(backdrop.val[1], source.val[3]) blended.val[1] = multiply(backdrop.val[1], source.val[3])
blended.val[2] = multiply(backdrop.val[2], source.val[3]) blended.val[2] = multiply(backdrop.val[2], source.val[3])
blended.val[3] = multiply(backdrop.val[3], source.val[3]) blended.val[3] = multiply(backdrop.val[3], source.val[3])
vst4_u8(a[i].addr, blended) vst4q_u8(a[i].addr, blended)
i += 8 i += 16
for i in i ..< len: for i in i ..< len:
a[i] = blendMask(a[i], b[i]) a[i] = blendMask(a[i], b[i])