Commit Graph

45 Commits

Author SHA1 Message Date
Henrik Rydgård
ab1072224f Cleanup, add another GPU stat 2024-10-29 08:18:41 +01:00
Henrik Rydgård
533fa14e27 Add a stat for number of verts decoded. It's quite high but not crazy for UFC 2010...
See #18806
2024-02-01 19:19:53 +01:00
Henrik Rydgård
0905b6a5ad Frustum-cull small draws
Some games do a poor job of culling stuff, and some transparent
sprites can be very expensive if they cause a copy.
Skipping them if outside the viewport makes sense in that case.

One example are the flame sprites in #17797 .

Additionally, we should be able to cull through-mode draws easily, this
one doesn't even try.
2023-12-09 15:55:51 +01:00
Henrik Rydgård
632fa1c9d6 Cache and hash data for DrawPixels.
We already had a cache to reuse texture objects so just
opportunistically reuse them when easy to do so.
2023-11-11 19:58:12 +01:00
Henrik Rydgård
e4ea4831e9 Delete the vertex cache option from the code. 2023-10-10 15:43:43 +02:00
Henrik Rydgård
226d25721a Add a block transfer GPU stat, remove a redundant one 2023-10-03 13:15:55 +02:00
Henrik Rydgård
e63bb0459c Add a new stat, so we can see per game if the optimization has an effect 2023-10-03 11:01:37 +02:00
Henrik Rydgård
061131ec8a Cache planes used for BBOX culling
This isn't a huge performance boost for the games that use BBOX (like
Tekken), but it'll be more valuable if we start using soft culling more
widely automatically, see #17808
2023-07-30 14:42:22 +02:00
Henrik Rydgård
ea552bc573 Add a new GPU stat (DrawPixels), kinda heavy since creates textures. GoW does 20-ish per frame. 2023-05-30 10:15:34 +02:00
Henrik Rydgård
06b83115a5 Replacer: Avoid tracking video textures 2023-03-10 12:20:55 +01:00
Henrik Rydgård
2d0252feb8
Merge pull request #16916 from hrydgard/implement-delayed-readback
Implement delayed depth readbacks, Vulkan only
2023-02-08 21:34:29 +01:00
Henrik Rydgård
9827dd536c Count listsyncs and drawsyncs in gpu stats. Eat some cycles in listSync
Chose a conservative number (same as sceGeContinue)
2023-02-08 00:41:26 +01:00
Henrik Rydgård
b3e6b81e43 Implement delayed depth readbacks for Vulkan only 2023-02-07 23:13:36 +01:00
Henrik Rydgård
62484f01ed Make ReadbackFramebufferSync able to use the stretch ability of ReadbackDepthbufferSync 2023-02-04 12:05:50 +01:00
Henrik Rydgård
d49b39e422 Remove gpuCommandsAtCallLevel stats, compute another stat more efficiently
These stats have proven rather useless since we don't plan to cache and
optimize display lists.
2023-01-04 17:07:45 +01:00
Unknown W. Brackets
be84bc9934 softgpu: Detect binner alloc fail and bail. 2023-01-01 06:59:14 -08:00
Henrik Rydgård
694dbe1271 Add ShaderDepalMode enum, use in shader ID. Replaces smoothed bit. 2022-09-11 14:40:01 +02:00
Henrik Rydgård
e69bb4f616 Buildfix, warn fixes while at it 2022-09-05 23:40:43 +02:00
Henrik Rydgård
16212cae73 NFS speed effect: Improve framebuffer choice logging, tighten up stride checks. 2022-09-05 23:31:30 +02:00
Henrik Rydgård
847e05140b Add another stat, for self-tex 2022-08-29 17:41:29 +02:00
Henrik Rydgård
f228de76ef Fix OpenGL bug when framebuffer fetch is available. Add two new stats. 2022-08-29 15:39:29 +02:00
Henrik Rydgård
51686f4936 Copy color from overlapping framebuffers on bind, under certain conditions.
Leads to much faster performance in Juiced 2.

This will later be expanded to handle more things in a more elegant way,
like the framebuffer overlap in God of War for the shadows and
color reinterpretation in a generic way.

Fixes #15728
2022-08-22 16:06:55 +02:00
Henrik Rydgård
886679c7ec Remove the color-to-depth mode 2022-08-20 08:29:33 +02:00
Henrik Rydgård
131098c4d4 Some enum renaming, move RasterChannel to GPU.h. 2022-08-09 19:58:48 +02:00
Henrik Rydgård
90a44579bf Implement color-to-depth for Vulkan, start implementing for D3D11 2022-08-01 11:30:36 +02:00
Henrik Rydgård
89845eae7a Add GPU stat for number of depth copies per frame
Will be useful for evaluating #15700
2022-07-24 17:12:43 +02:00
Henrik Rydgård
3093b9c7bf Share GPUS stats formatting code between the backends (easier to add new stats) 2020-09-20 22:09:25 +02:00
Henrik Rydgård
7630fae4b9 Fix bug in new framebuffer code, reducing unnecessary texture invalidations.
Improves the performance issue in #13431
2020-09-16 01:03:56 +02:00
Henrik Rydgård
f6db1c2b01 Show how many times we evaluate framebuffers per frame in debug statistics. 2020-09-12 15:37:03 +02:00
Unknown W. Brackets
5871ab0538 UI: Stop caching the draw context in coreParam.
This is possibly getting outdated in some paths of graphics reinit, and
then causing crashes.  Let's just always get it from the graphicsContext.
2019-09-28 21:58:15 -07:00
Unknown W. Brackets
0151c5e649 GPU: Track stats on uploads.
A lot can lead to slowdown, so it's useful when logging frame drops.
2017-12-28 19:27:45 -08:00
Unknown W. Brackets
f9750dd137 Display: Avoid limiting FPS without a clear.
A bit of a dirty heuristic to avoid the slowdown in #8538.
2017-12-26 19:37:40 -08:00
Unknown W. Brackets
60c4ac58f0 System: Allow GPU to have delayed loading. 2017-12-03 11:29:07 -08:00
Henrik Rydgård
d618b3673b Count readbacks in a frame and show in GPU debug stats. 2017-11-08 11:57:53 +01:00
Henrik Rydgård
f86b9f7900 Guess what? More renaming 2016-12-27 11:59:06 +01:00
Henrik Rydgård
10393f5d4e Move Thin3D into the Draw namespace. 2016-12-27 11:58:03 +01:00
Henrik Rydgard
727058fc27 Remove an entry from GPUStats that didn't really belong. 2016-04-02 11:06:05 +02:00
Henrik Rydgard
c1e9df0777 Adjust the displayed debug stats for each backend. Cleanup. 2016-03-31 10:23:40 +02:00
Henrik Rydgard
29bc07eb0a SoftGPU with Vulkan runs but displays black 2016-03-20 20:04:49 +01:00
Henrik Rydgard
f7113bb2e3 Use separate pushbuffers for UBO/Index/Vertex. This will make decoding directly into the pushbuffers easier, plus other benefits later.
(For example, we'll often be able to avoid rebinding the vertex and
index buffers at new offsets by just keep counting upwards between draws,
if the vertex format is the same but other state changed)
2016-03-20 19:31:02 +01:00
Henrik Rydgard
e45c24b75f Show how much pushbuffer space is used each frame.
Optimizing this number (by avoiding redundant UBO uploads etc) will probably help performance.
2016-03-20 19:31:02 +01:00
Henrik Rydgard
e11d0a7e1c Minor GPU interface cleanup 2016-01-06 23:49:02 +01:00
Henrik Rydgard
15de6e6b98 GraphicsContext: Abstract away things like swapbuffers etc before adding even more backends.
Needed to prevent clutter all over the codebase.

Does not go all the way yet, goal would be a common render loop between platforms but not there yet.
2016-01-01 14:40:16 +01:00
Unknown W. Brackets
f982e777e5 Stop logging alphatested draws for debug stats.
Measured a small perf hit here, and we're not really using these stats
much anymore.
2015-12-30 10:24:54 -08:00
Henrik Rydgard
2430c283a5 More GPU cleaning, removing uses of GPUState.h where not needed.
Want to get rid of direct accesses to GPUState in modules that may be reused in
my future next-gen backends, that will reformat display lists into command lists that will
then be optimized and executed, out of sync with the real GPUState.

Candidate modules that may be reused in full are Framebuffer and Depal, possibly TextureCache to some degree.
2015-07-29 12:37:49 +02:00