Support graphics cards without `ARB_half_float_vertex` #1179

illwieckz · 2024-06-04T03:14:21Z

This is to address those two issues:

Some graphics devices don't supports ARB_half_float_vertex like the Intel GMA Gen3 architecture with the Mesa driver.
Some graphics devices don't supports ARB_half_float_vertex but drivers may provide an unplayable slow emulation, like the ATi r300 architecture with the Mesa driver.
- Huge performance drop on r300 with default Thunder and Vega scene #1172

Those changes purposes to provide an alternative code path when ARB_half_float_vertex is not available or when the r_arb_half_float_vertex cvar is disabled on purpose to avoid a slow emulation provided by the driver.

It is the next step over preparatory work that was already merged:

Some improvements of our type system, helping the compiler to report more type errors and then making the conversion easier with the less risk of introducing silent errors:
Implement f16_t, f16vec2_t, f16vec4_t half float types for safety #728
Some enablement of GL 2.1 on the said Intel graphics chip so we can know GL 2.1 is supported but catch the lack of the half float vertex extension, meaning that if we implement a workaround one day (that is the day today), we would already have everything else already in place for it:
renderer/glimp: new GL detection and selection code #478

This branch doesn't break compatibility with the game but I implemented it over the illwieckz/image-fistcreen/sync branch anyway:

renderer: introduce IF_FITSCREEN and RSF_2D #1145

I did it that way because this other branch also helps running the game on this kind of hardware (to not upload textures too big for such hardware) and we better test this hardware over for-0.55.0/sync because of all compatibility-breaking changes purposed for performance that were merged there. The illwieckz/image-fistcreen/sync branch itself is already implemented over for-0.55.0/sync because of compatibility breaking.

This branch also includes the GL_DEPTH_CLAMP disablement when GL_ARB_depth_clamp to avoid spamming the console with warnings when running the game on ATi r300 graphics card, to make debugging easier:

renderer: use GL_DEPTH_CLAMP only if GL_ARB_depth_clamp is available #1173

~~⚠️🚧️ This branch is in very early state! 🚧️⚠️~~
~~⚠️🚧️ It's even not ready to review! 🚧️⚠️~~
~~⚠️🚧️ This may be full of useless halffloat-to-float-to-halffloat conversions! 🚧️⚠️~~
~~⚠️🚧️ I even haven't tested if it doesn't break the usual code path! 🚧️⚠️~~

~~I push this branch now because it starts to work and this is so precious I don't want the team to lose that work is something happens to me!~~

It's now ready to be reviewed, it works for me with both code paths, and is rebased on master.

illwieckz · 2024-06-11T01:30:30Z

For unknown reasons, this is broken:

mouse pointer
minimap
console text

illwieckz · 2024-06-11T06:17:14Z

For unknown reasons, this is broken:

mouse pointer

minimap

console text

It was just some stupid copy-paste mistake.

Now it looks like everything is working.

Also fixes the version printing.

illwieckz · 2024-06-11T07:29:20Z

Now that I verified the branch works, the branch is now targeting master.

It is ready to review.

…tToHalf4 to avoid confusion

VReaperV

Same question w. r. t. vector2Copy() vs floatToHalf() usage where the original code only has the former and the branches in this pr seem switched up.

src/engine/client/hunk_allocator.cpp

src/engine/renderer/tr_backend.cpp

src/engine/renderer/tr_surface.cpp

src/engine/renderer/tr_vbo.cpp

src/engine/renderer/tr_local.h

Generate the layout for interleaved vertex attribute data at runtime. The motivation for this is to support OpenGL implementations that don't provide half float support (#1179). The vertex "struct" may contain a 16-bit or 32-bit float, depending on the graphics card. Now, instead of defining a struct for the data to be uploaded into a VBO, one must separately specify inputs for each attribute. The input is defined by a type, base address, stride, etc.; very similarly to the arguments of glVertexAttribPointer itself. The new version of R_CreateStaticVBO takes these inputs and writes them to an interleaved format, performing any neede type conversions along the way. In this commit just skeletal models (IQM and MD5) are migrated to the new method.

slipher · 2024-08-09T06:38:36Z

You can see my current progress on the branch slipher/runtime-vbo-layout. The dynamically generated vertex attribute layout, which may be configured to use half float or not, is used for IQM, MD3, and MD5 models. It's not ready to test as other VBOs are not ported yet. Remaining steps:

Investigate the vertex attribute-related warning spam on station15. IIRC this is caused by some code which creates some verts with the qtangent attribute, then later overwrites that data with the incompatible orientation attribute. I should be careful to avoid breaking this.
Use the new vertex attribute layout method for the other static VBOs.
Change the dynamic VBO's vertex struct to (always) use float instead of f16_t. This is probably better anyway. Data in the dynamic VBO is only used once and (I believe; will double check) is usually small. Trying to compress it seems pointless. Also tess.verts is sometimes used for data which is not actually sent to the GPU, so avoiding useless conversions should be a win there.

illwieckz · 2024-08-09T18:12:58Z

Great! I like the way it looks!

As a side note, I prefer the cvar being named r_arb_half_float_vertex (not r_ext_half_float_vertex) as the extension is ARB_half_float_vertex, and ordered alphabetically like the others.

slipher · 2024-08-10T05:32:36Z

I spelled out my thoughts on the extension cvar naming. To me it looks like a case of cargo culting gone out of control. Maybe some of the first extensions were EXT ones, which makes some sense as part of the cvar name. But ARB really doesn't.

…t vertex

illwieckz · 2024-10-01T07:27:52Z

I rebased on latest master, which reduced the diff size as some code was just NUKED since.

I added a detection of early RV300 hardware were we should prefer non-half float, and added a workaround cvar that can be disabled to test the r300 driver half-float performance when not using such workaround.

This looks ready to me.

Regarding to the slipher/runtime-vbo-layout branch, I do believe it is the future, but I prefer to see that branch rebased on this one. This one also brings some optimizations and some renaming that makes easier to follow the code and I would like to merge that in all cases anyway.

We can already merge that as it already works, without waiting for slipher/runtime-vbo-layout. While I do believe such deeper redesign is the future, it can wait a bit.

Basically a rebase of slipher/runtime-vbo-layout would have to add an extra parent commit to unsplit the float/half-float if/else code paths. I can help at rebasing it.

slipher · 2024-10-07T00:03:16Z

I just looked at my vertex layout branch again and it's actually 95% done. The big thing it was waiting on was the autosprite rework (so that we no longer have overlapping attributes). I can have the PR ready in the next day or two.

illwieckz · 2024-10-07T11:19:31Z

Great!

Generate the layout for interleaved vertex attribute data at runtime. The motivation for this is to support OpenGL implementations that don't provide half float support (#1179). The vertex "struct" may contain a 16-bit or 32-bit float, depending on the graphics card. Now, instead of defining a struct for the data to be uploaded into a VBO, one must separately specify inputs for each attribute. The input is defined by a type, base address, stride, etc.; very similarly to the arguments of glVertexAttribPointer itself. The new version of R_CreateStaticVBO takes these inputs and writes them to an interleaved format, performing any neede type conversions along the way. In this commit just skeletal models (IQM and MD5) are migrated to the new method.

illwieckz · 2024-10-10T23:44:46Z

Obsoleted by #1344:

Runtime VBO layout: make half float usage optional #1344

slipher · 2024-10-11T00:07:56Z

We still need the commit to add a workaround flipping off half float vertex for the driver with poor quality of implementation.

illwieckz · 2024-10-11T00:08:20Z

Yes, I'm on it. :-)

illwieckz · 2024-10-11T00:14:23Z

See:

Detect RV300 and VC4 hardware and disable slow half float vertex on them #1347

renderer: add alternate code path without ARB_half_float_vertex

472dc72

illwieckz added A-Renderer T-Feature-Request Proposed new feature labels Jun 4, 2024

illwieckz marked this pull request as draft June 4, 2024 03:15

illwieckz changed the title ~~WIP: support graphis cards without ARB_half_float_vertex~~ WIP: support graphics cards without ARB_half_float_vertex Jun 4, 2024

illwieckz mentioned this pull request Jun 4, 2024

Huge performance drop on r300 with default Thunder and Vega scene #1172

Closed

illwieckz force-pushed the illwieckz/no-half-float-vertex branch 2 times, most recently from 08bd535 to e6ec633 Compare June 9, 2024 02:18

tr_shade_calc: optimize half-float round trip

f2c3caa

illwieckz force-pushed the illwieckz/no-half-float-vertex branch 2 times, most recently from 5cd8db6 to c1d888b Compare June 11, 2024 01:29

illwieckz force-pushed the illwieckz/no-half-float-vertex branch 3 times, most recently from 1898c46 to 38c78e7 Compare June 11, 2024 06:16

sdl_glimp: update the error message when OpenGL is too old

4da7953

Also fixes the version printing.

illwieckz force-pushed the illwieckz/no-half-float-vertex branch from b5c0125 to 1b58b39 Compare June 11, 2024 07:04

illwieckz changed the title ~~WIP: support graphics cards without ARB_half_float_vertex~~ Support graphics cards without ARB_half_float_vertex Jun 11, 2024

illwieckz force-pushed the illwieckz/no-half-float-vertex branch from 1b58b39 to 2f8e40f Compare June 11, 2024 07:27

illwieckz marked this pull request as ready for review June 11, 2024 07:27

illwieckz changed the base branch from for-0.55.0/sync to master June 11, 2024 07:28

illwieckz force-pushed the illwieckz/no-half-float-vertex branch from 2f8e40f to 4a01370 Compare June 11, 2024 07:28

renderer: rename halfToFloat and floatToHalf as halfToFloat4 and floa…

8c53a2f

…tToHalf4 to avoid confusion

illwieckz force-pushed the illwieckz/no-half-float-vertex branch 4 times, most recently from 8d8d9f5 to 0d6b114 Compare June 11, 2024 08:05

VReaperV reviewed Jun 11, 2024

View reviewed changes

illwieckz force-pushed the illwieckz/no-half-float-vertex branch from e86ea98 to 7f896cb Compare August 7, 2024 14:49

sdl_glimp: detect ATI RV300 and use float vertex instead of half-floa…

3697d53

…t vertex

illwieckz force-pushed the illwieckz/no-half-float-vertex branch from 7f896cb to fe10c3a Compare September 30, 2024 19:43

illwieckz force-pushed the illwieckz/no-half-float-vertex branch 6 times, most recently from 7e1282e to 3697d53 Compare October 2, 2024 11:43

slipher mentioned this pull request Oct 9, 2024

Runtime VBO layout: make half float usage optional #1344

Merged

illwieckz closed this Oct 10, 2024

slipher reopened this Oct 11, 2024

illwieckz mentioned this pull request Oct 11, 2024

Detect RV300 and VC4 hardware and disable slow half float vertex on them #1347

Merged

slipher closed this Oct 11, 2024

illwieckz mentioned this pull request Oct 29, 2024

sdl_glimp: update the error message when OpenGL is too old #1401

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Support graphics cards without `ARB_half_float_vertex` #1179

Support graphics cards without `ARB_half_float_vertex` #1179

illwieckz commented Jun 4, 2024 •

edited

Loading

illwieckz commented Jun 11, 2024

illwieckz commented Jun 11, 2024 •

edited

Loading

illwieckz commented Jun 11, 2024

VReaperV left a comment

slipher commented Aug 9, 2024

illwieckz commented Aug 9, 2024

slipher commented Aug 10, 2024

illwieckz commented Oct 1, 2024 •

edited

Loading

slipher commented Oct 7, 2024

illwieckz commented Oct 7, 2024

illwieckz commented Oct 10, 2024

slipher commented Oct 11, 2024

illwieckz commented Oct 11, 2024

illwieckz commented Oct 11, 2024

Support graphics cards without ARB_half_float_vertex #1179

Support graphics cards without ARB_half_float_vertex #1179

Conversation

illwieckz commented Jun 4, 2024 • edited Loading

illwieckz commented Jun 11, 2024

illwieckz commented Jun 11, 2024 • edited Loading

illwieckz commented Jun 11, 2024

VReaperV left a comment

Choose a reason for hiding this comment

slipher commented Aug 9, 2024

illwieckz commented Aug 9, 2024

slipher commented Aug 10, 2024

illwieckz commented Oct 1, 2024 • edited Loading

slipher commented Oct 7, 2024

illwieckz commented Oct 7, 2024

illwieckz commented Oct 10, 2024

slipher commented Oct 11, 2024

illwieckz commented Oct 11, 2024

illwieckz commented Oct 11, 2024

Support graphics cards without `ARB_half_float_vertex` #1179

Support graphics cards without `ARB_half_float_vertex` #1179

illwieckz commented Jun 4, 2024 •

edited

Loading

illwieckz commented Jun 11, 2024 •

edited

Loading

illwieckz commented Oct 1, 2024 •

edited

Loading