diff options
author | wm4 <wm4@nowhere> | 2016-07-03 16:31:38 +0200 |
---|---|---|
committer | wm4 <wm4@nowhere> | 2016-07-03 16:34:32 +0200 |
commit | b57debe5b3c8c85f14c1d7f895e61d4f4f26dcc4 (patch) | |
tree | f4a250de72339e0581aeca4c90de4dc19a3ac693 /video/out/opengl/utils.h | |
parent | 8c7f9dc1a89e90e44b2b7dfb13fa899ad52352ee (diff) | |
download | mpv-b57debe5b3c8c85f14c1d7f895e61d4f4f26dcc4.tar.bz2 mpv-b57debe5b3c8c85f14c1d7f895e61d4f4f26dcc4.tar.xz |
vo_opengl: use ringbuffer of PBOs
This is how PBOs are normally supposed to be used.
Unfortunately I can't see an any absolute improvement on nVidia binary
drivers and playing 4K material. Compared to the "old" PBO path with 1
buffer, the measured GL time decreases significantly, though.
Diffstat (limited to 'video/out/opengl/utils.h')
-rw-r--r-- | video/out/opengl/utils.h | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/video/out/opengl/utils.h b/video/out/opengl/utils.h index 35211f6485..ec54d19b8a 100644 --- a/video/out/opengl/utils.h +++ b/video/out/opengl/utils.h @@ -184,7 +184,8 @@ uint64_t gl_timer_peak_us(struct gl_timer *timer); struct gl_pbo_upload { GL *gl; - GLuint buffer; + int index; + GLuint buffers[2]; size_t buffer_size; }; |