Enable Link-Time Optimization (LTO) and codegen-units = 1

[zamazan4ik]

Hi!

I noticed that in the Cargo.toml file Link-Time Optimization (LTO) for the project is not enabled. I suggest switching it on since it will reduce the binary size (always a good thing to have) and will likely improve the application's performance a little bit. If you want to read more about LTO and its possible modes, I recommend starting from this Rustc documentation.

I think you can enable LTO only for the Release builds so as not to sacrifice the developers' experience while working on the project since LTO consumes an additional amount of time to finish the compilation routine. If you think that a regular Release build should not be affected by such a change as well, then I suggest adding an additional dist or release-lto profile where in addition to regular release optimizations LTO will also be added. Such a change simplifies life for maintainers and others interested in the project persons who want to build the most optimized version of the application. However, if we enable it on the Cargo profile level for the Release profile, users, who install the application with cargo install will get the LTO-optimized version of the game "automatically". E.g., check cargo-outdated Release profile. You also could be interested in other optimization options like codegen-units = 1, etc.

Basically, it can be enabled with the following lines:

[profile.release]
codegen-units = 1
lto = true

I have made quick tests (Fedora 41, Rust 1.82, the latest version of the project, CC=clang CXX=clang++ cargo build --release --workspace command) - here are the results (first results - the original Release profile, second results - the original Release profile + Fat LTO + codegen-units = 1):

• zluda_with: from 405 Kib to 361 Kib
• zluda_bindgen: from 8.4 Mib to 5.8 Mib
• libcuda_base.so: from 4.6 Mib to 4.4 Mib
• libnvcuda.so: from 5 Mib to 4.8 Mib
• libnvml.so: almost no changes
• libptx_parser_macros.so: from 4 to 3.7 Mib
• libzluda_dump.so: from 3.2 Mib to 2.8 Mib
• libzluda_redirect.so: almost no changes

I haven't made performance tests since I don't know how to do it for the project.

Thank you.

[vosen]

Makes sense, can you create a PR for an additional release-lto profile with those changes?

[zamazan4ik]

Sure. One question: do I need to change anything in a CI pipeline (if any exists for the project) to enable using the release-lto profile for GitHub releases? Since they are published at least for Windows, I suppose we need to enable using this heavily-optimized profile there too. However, if the releases are prepared manually by you - heh, just don't forget to use release-lto next time :D

[vosen]

There is no CI pipeline yet 😞. I'll add it after I get around to #287

[Sfinx]

Git this at fresh fee20e5 build:

...
Downloaded 26 crates (3.1 MB) in 1.33s (largest was `logos` at 1.5 MB)
error: profile `release-lto` is missing an `inherits` directive (`inherits` is required for all profiles except `dev` or `release`)

[zamazan4ik]

@Sfinx thank you for the report - it's my bad since I forgot to add "inherits". It's fixed in #319 . That's what happens when you make changes directly on GitHub without checking them locally/on CI :D