64461329e7
More minor fixes and cleanups
2026-01-02 14:20:20 +01:00
7cc4a0f3a3
Add more accurate sleep function (well, not more accurate on Windows unfortunately)
2025-10-21 12:11:23 +02:00
9218fe9b59
Rework the shutdown logic in the OpenGL backend. Eliminated race conditions.
2025-08-31 10:57:06 +02:00
e1527233d9
Add the EmuScreen pass assert on OpenGL as well
2024-10-14 14:35:35 +02:00
2ba4eaf3dd
First part of the const changes etc
2024-09-17 15:13:13 +02:00
3c66f149d3
[Common/Core/Windows] Removed excess check pointer before delete or free()
2024-09-17 11:34:42 +02:00
3065979528
iOS: Fix hang when task switching in GL mode
2024-05-29 18:21:58 +02:00
8e580045cd
Fixed condition always false using unsigned int and logical AND 0 & 4
2024-04-11 16:56:59 +03:00
cbb4236cd8
Vulkan: Attempt to fix some more shutdown race conditions, simplify.
2024-01-17 14:49:49 +01:00
d891aaf9cd
Remove code that pretended that we supported multiple vertex streams
...
Don't really see that we'll have much use for this feature, so simplify
it away. Only single vertex stream data is now supported by the thin3d
API.
2023-11-13 01:15:28 +01:00
4d465678cd
GL: Add resilience to delete-null bugs (but assert in debug builds)
2023-09-26 00:49:59 +02:00
691c8b8d8d
Write to the frame time history from the other backends too. Needed for the upcoming timing code.
2023-08-16 12:22:49 +02:00
e651d6e59f
Finish the separation of EndFrame and Present.
2023-08-13 18:40:35 +02:00
377704f195
OpenGL: Split submit and presentation into two tasks, but schedule them together
2023-08-13 18:40:35 +02:00
86fcb28d4d
Remove "bool retainControl_" for simplicity. Not needed.
2023-08-13 18:40:35 +02:00
0ccd29f2ba
Revert "OpenGL: Separate submit/present for this backend as well"
...
This reverts commit 0143d67f9b
2023-08-12 13:35:21 +02:00
0143d67f9b
OpenGL: Separate submit/present for this backend as well
2023-08-10 17:46:10 +02:00
2bee5b64e4
Add new icon cache, for caching small images.
...
It doesn't try to insert or fetch missing things itself, re-fetching is up to the caller.
This will be required for handling the many achievement icons.
Saving/loading the cache to a single file on disk is implemented but not
hooked up yet. It works without it, though of course will have to
re-fetch things on the next startup.
2023-06-18 14:11:12 +02:00
880379c15d
Extract some minor changes from #17497
2023-06-12 20:20:06 +02:00
ad8827ae70
Cleanup, address feedback
2023-05-26 10:28:10 +02:00
62b41c6640
OpenGL: Add a simple pass list to gpu profiler
2023-05-24 14:08:19 +02:00
edd208791e
Add tag field to GLRFramebuffer for debugging
2023-05-24 13:47:51 +02:00
84da0327d6
GLQueueRunner: Make DrawIndexed parameters more consistent.
2023-05-23 17:49:11 +02:00
d7ea2ebf8a
GL: Add tags to push buffers for debug purposes. Double the default vertex buffer size.
2023-05-23 15:53:41 +02:00
cb38c43d7e
GL render manager: Merge the two stencil commands, for more compact command lists
2023-05-23 09:14:23 +02:00
37906384ee
Use the faster pushes.
2023-05-23 08:54:41 +02:00
558e29a9bb
Switch to the copy-free method of initializing initsteps
2023-05-23 08:54:41 +02:00
956d784bde
Add FastVec, start using it for InitSteps
2023-05-23 08:54:40 +02:00
78eaa8c235
Make sure we never copy GLRRenderThreadTask objects
2023-05-23 08:53:48 +02:00
c30895b6b4
Remove unnecessary variable
2023-05-23 08:53:48 +02:00
e4a729d371
Remove unnecessary zero-initialization of GLRRenderCommand structs
2023-05-23 08:53:48 +02:00
d7a5edeb6f
GL: Break out GLPushBuffer from GLRenderManager.cpp/h
2023-05-23 08:41:09 +02:00
05b6bbdc56
Add a trivial profiling tool to the OpenGL backend - meaasure the time it takes to run a frame of commands.
...
Accessed from the in-game dev menu just like the Vulkan frame profiler.
With this we can easily see that actually submitting the GL commands is often the bottleneck on old
devices like a Galaxy S3.
2023-05-17 14:38:11 +02:00
c729519cf5
Merge pull request #17442 from hrydgard/glr-render-command-shrink
...
Shrink the GLRRenderCommand struct from 152 to 88 bytes
2023-05-16 16:38:41 +02:00
b27c427547
We always bind and draw together, so let's combine them to one command.
2023-05-10 10:56:25 +02:00
c7caefe6d8
Have Draw and DrawIndexed share command and struct. Will make the next change easier.
2023-05-10 10:40:33 +02:00
d5e0299b0b
GLR: Share union struct between Draw/DrawIndex
2023-05-10 10:32:13 +02:00
8f2069ef71
GLQueueRunner: Combine BindVertexBuffer with BindIndexBuffer
2023-05-10 09:23:10 +02:00
cc16eb8321
Shrink the GLRRenderCommand struct from 152 to 88 bytes
...
Turns out the VR work bloated it a bit, which can't be good.
Think it's fine to allocate these view matrices on the heap to get them
out of the way, there won't be that crazy many per frame usually.
2023-05-10 00:24:17 +02:00
5724bbd8e9
Correct size calculations in GLPushBuffer. I don't see how this failed as badly as in that crash report though...
2023-05-03 23:49:41 +02:00
d4249c1d73
OpenGL: Add an assert to catch a class of crash bugs early. Also assorted paranoia.
2023-05-01 11:56:26 +02:00
3d477003c4
Fix issue where if Break on Load was set, we'd crash once we started running in a few games
2023-04-30 10:06:43 +02:00
7cbb7a09e5
[what][style][gpu] some coding style refine
2023-04-18 19:18:38 +08:00
d502198aa3
Move GLFrameData out of GLRenderManager.
...
Small refactor to prepare for adding delayed readback support to the
OpenGL backend.
2023-02-07 12:12:00 +01:00
937975000b
Add a ReadbackMode parameter to more functions in the FramebufferManager
2023-02-05 13:57:45 +01:00
cb16f193ec
Fix the problem. Still need to leave the loop with SDL even if we supply a swap function.
2023-01-30 15:45:05 +01:00
436a3e0d61
Cleaner exit from emuthread, as used by OpenGL. Don't know how it worked before...
2023-01-30 11:49:31 +01:00
a0c893cc8f
Fix ThreadFrame. All OK now except shutdown, it seems
2023-01-30 11:49:31 +01:00
2659fd6f66
Android work. Make the Android OpenGL exit prodecure make more sense
2023-01-30 11:49:31 +01:00
b62572a78f
Refactor the framedata/GPU thread management.
2023-01-30 11:49:31 +01:00
Henrik Rydgård