Skip to content

Commit

Permalink
pipeline optimize
Browse files Browse the repository at this point in the history
  • Loading branch information
@nihui
nihui committed Dec 24, 2017
1 parent 0564021 commit 1f5c646
Showing 1 changed file with 33 additions and 30 deletions.
63 changes: 33 additions & 30 deletions src/layer/arm/convolution_1x1.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -190,49 +190,49 @@ static void conv1x1s1_neon(const Mat& bottom_blob, Mat& top_blob, const Mat& _ke
asm volatile(
"pld [%5, #256] \n"
"vld1.f32 {d12-d15}, [%5 :128]! \n"
"0: \n"
"pld [%1, #256] \n"
"vld1.f32 {d16-d19}, [%1 :128] \n"
"0: \n"

"vmla.f32 q8, q6, %e18[0] \n"
"vmla.f32 q9, q7, %e18[0] \n"

"pld [%2, #256] \n"
"vld1.f32 {d20-d23}, [%2 :128] \n"
"vmla.f32 q9, q7, %e18[0] \n"

"vmla.f32 q10, q6, %e19[0] \n"
"vmla.f32 q11, q7, %e19[0] \n"

"pld [%3, #256] \n"
"vld1.f32 {d24-d27}, [%3 :128] \n"
"vmla.f32 q11, q7, %e19[0] \n"

"vmla.f32 q12, q6, %e20[0] \n"
"vmla.f32 q13, q7, %e20[0] \n"

"pld [%4, #256] \n"
"vld1.f32 {d28-d31}, [%4 :128] \n"

"vmla.f32 q14, q6, %e21[0] \n"
"vmla.f32 q15, q7, %e21[0] \n"
"vmla.f32 q13, q7, %e20[0] \n"

"pld [%6, #256] \n"
"vld1.f32 {d12-d15}, [%6 :128]! \n"
"vld1.f32 {d8-d11}, [%6 :128]! \n"

"vmla.f32 q8, q6, %e18[1] \n"
"vmla.f32 q9, q7, %e18[1] \n"
"vmla.f32 q14, q6, %e21[0] \n"
"vmla.f32 q15, q7, %e21[0] \n"

"vmla.f32 q10, q6, %e19[1] \n"
"vmla.f32 q11, q7, %e19[1] \n"
"vmla.f32 q8, q4, %e18[1] \n"
"vmla.f32 q9, q5, %e18[1] \n"

"vmla.f32 q12, q6, %e20[1] \n"
"vmla.f32 q13, q7, %e20[1] \n"
"vmla.f32 q10, q4, %e19[1] \n"
"vmla.f32 q11, q5, %e19[1] \n"

"vmla.f32 q14, q6, %e21[1] \n"
"vmla.f32 q15, q7, %e21[1] \n"
"vmla.f32 q12, q4, %e20[1] \n"
"vmla.f32 q13, q5, %e20[1] \n"

"pld [%7, #256] \n"
"vld1.f32 {d12-d15}, [%7 :128]! \n"

"vmla.f32 q14, q4, %e21[1] \n"
"vmla.f32 q15, q5, %e21[1] \n"

"vmla.f32 q8, q6, %f18[0] \n"
"vmla.f32 q9, q7, %f18[0] \n"

Expand All @@ -242,32 +242,35 @@ static void conv1x1s1_neon(const Mat& bottom_blob, Mat& top_blob, const Mat& _ke
"vmla.f32 q12, q6, %f20[0] \n"
"vmla.f32 q13, q7, %f20[0] \n"

"pld [%8, #256] \n"
"vld1.f32 {d8-d11}, [%8 :128]! \n"

"vmla.f32 q14, q6, %f21[0] \n"
"vmla.f32 q15, q7, %f21[0] \n"

"pld [%8, #256] \n"
"vld1.f32 {d12-d15}, [%8 :128]! \n"
"vmla.f32 q8, q4, %f18[1] \n"
"vmla.f32 q9, q5, %f18[1] \n"

"vmla.f32 q8, q6, %f18[1] \n"
"vmla.f32 q9, q7, %f18[1] \n"

"vmla.f32 q10, q6, %f19[1] \n"
"vmla.f32 q11, q7, %f19[1] \n"
"vmla.f32 q10, q4, %f19[1] \n"
"vmla.f32 q11, q5, %f19[1] \n"

"vst1.f32 {d16-d19}, [%1 :128]! \n"

"vmla.f32 q12, q6, %f20[1] \n"
"vmla.f32 q13, q7, %f20[1] \n"
"vmla.f32 q12, q4, %f20[1] \n"
"vmla.f32 q13, q5, %f20[1] \n"

"vst1.f32 {d20-d23}, [%2 :128]! \n"

"vmla.f32 q14, q6, %f21[1] \n"
"vmla.f32 q15, q7, %f21[1] \n"
"pld [%5, #256] \n"
"vld1.f32 {d12-d15}, [%5 :128]! \n"

"vmla.f32 q14, q4, %f21[1] \n"
"vmla.f32 q15, q5, %f21[1] \n"

"vst1.f32 {d24-d27}, [%3 :128]! \n"

"pld [%5, #256] \n"
"vld1.f32 {d12-d15}, [%5 :128]! \n"
"pld [%1, #256] \n"
"vld1.f32 {d16-d19}, [%1 :128] \n"

"subs %0, #1 \n"
"vst1.f32 {d28-d31}, [%4 :128]! \n"
Expand Down Expand Up @@ -296,7 +299,7 @@ static void conv1x1s1_neon(const Mat& bottom_blob, Mat& top_blob, const Mat& _ke
"w"(_k1), // %19
"w"(_k2), // %20
"w"(_k3) // %21
: "cc", "memory", "q6", "q7", "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15"
: "cc", "memory", "q4", "q5", "q6", "q7", "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15"
);
}
#endif // __aarch64__
Expand Down

0 comments on commit 1f5c646

Please sign in to comment.