octo
/
yuzu-mainline
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #9777 from vonchenplus/speed_up_video_copy 

video_core: Speed up video frame data copy
This commit is contained in:
liamwhite 2023-02-11 12:35:16 -05:00 committed by GitHub
parent 709879cfc1 4adf39edf2
commit c97b5c9ff9
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 5 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
src/video_core/host1x/vic.cpp
View File

@ -189,9 +189,7 @@ void Vic::WriteYUVFrame(const AVFrame* frame, const VicConfig& config) {
for (std::size_t y = 0; y < frame_height; ++y) {
const std::size_t src = y * stride;
const std::size_t dst = y * aligned_width;
for (std::size_t x = 0; x < frame_width; ++x) {
luma_buffer[dst + x] = luma_src[src + x];
}
std::memcpy(luma_buffer.data() + dst, luma_src + src, frame_width);
}
host1x.MemoryManager().WriteBlock(output_surface_luma_address, luma_buffer.data(),
luma_buffer.size());
@ -205,15 +203,15 @@ void Vic::WriteYUVFrame(const AVFrame* frame, const VicConfig& config) {
// Frame from FFmpeg software
// Populate chroma buffer from both channels with interleaving.
const std::size_t half_width = frame_width / 2;
u8* chroma_buffer_data = chroma_buffer.data();
const u8* chroma_b_src = frame->data[1];
const u8* chroma_r_src = frame->data[2];
for (std::size_t y = 0; y < half_height; ++y) {
const std::size_t src = y * half_stride;
const std::size_t dst = y * aligned_width;
for (std::size_t x = 0; x < half_width; ++x) {
chroma_buffer[dst + x * 2] = chroma_b_src[src + x];
chroma_buffer[dst + x * 2 + 1] = chroma_r_src[src + x];
chroma_buffer_data[dst + x * 2] = chroma_b_src[src + x];
chroma_buffer_data[dst + x * 2 + 1] = chroma_r_src[src + x];
}
}
break;
@ -225,9 +223,7 @@ void Vic::WriteYUVFrame(const AVFrame* frame, const VicConfig& config) {
for (std::size_t y = 0; y < half_height; ++y) {
const std::size_t src = y * stride;
const std::size_t dst = y * aligned_width;
for (std::size_t x = 0; x < frame_width; ++x) {
chroma_buffer[dst + x] = chroma_src[src + x];
}
std::memcpy(chroma_buffer.data() + dst, chroma_src + src, frame_width);
}
break;
}