Try switching from rustacuda to cust

.github/workflows/rustdoc.yml

@@ -37,8 +37,8 @@ jobs:
             --enable-index-page \
             --extern-html-root-url const_type_layout=https://docs.rs/const-type-layout/0.3.2/ \
             --extern-html-root-url final=https://docs.rs/final/0.1.1/ \
-            --extern-html-root-url rustacuda=https://docs.rs/rustacuda/0.1.3/ \
-            --extern-html-root-url rustacuda_core=https://docs.rs/rustacuda_core/0.1.2/ \
+            --extern-html-root-url cust=https://docs.rs/cust/0.3.2/ \
+            --extern-html-root-url cust_core=https://docs.rs/cust_core/0.1/ \
             -Zunstable-options \
           " cargo doc \
             --all-features \

Cargo.toml

@@ -15,7 +15,7 @@ edition = "2021"
 authors = ["Juniper Tyree <[email protected]>"]
 repository = "https://github.com/juntyr/rust-cuda"
 license = "MIT OR Apache-2.0"
-rust-version = "1.81" # nightly
+rust-version = "1.84" # nightly
 
 [workspace.dependencies]
 # workspace-internal crates
@@ -24,12 +24,12 @@ rust-cuda-derive = { version = "0.1", path = "rust-cuda-derive", default-feature
 rust-cuda-kernel = { version = "0.1", path = "rust-cuda-kernel", default-features = false }
 
 # third-party dependencies with unpublished patches
-rustacuda = { git = "https://github.com/juntyr/RustaCUDA", rev = "c6ea7cc", default-features = false }
-rustacuda_core = { git = "https://github.com/juntyr/RustaCUDA", rev = "c6ea7cc", default-features = false }
+cust = { git = "https://github.com/juntyr/Rust-GPU-CUDA.git", rev = "5365c14", version = "0.3.2", default-features = false }
+cust_core = { git = "https://github.com/juntyr/Rust-GPU-CUDA.git", rev = "5365c14", version = "0.1", default-features = false }
 
 # crates.io third-party dependencies
 cargo_metadata = { version = "0.19", default-features = false }
-cargo-util = { version = "=0.2.16", default-features = false }  # TODO: keep in sync with toolchain
+cargo-util = { version = "=0.2.17", default-features = false }  # TODO: keep in sync with toolchain
 colored = { version = "3.0", default-features = false }
 const-type-layout = { version = "0.3.2", default-features = false }
 final = { version = "0.1.1", default-features = false }
@@ -92,16 +92,16 @@ default = []
 derive = ["dep:rust-cuda-derive"]
 device = []
 final = ["dep:final"]
-host = ["dep:rustacuda", "dep:regex", "dep:oneshot", "dep:safer_owning_ref"]
+host = ["dep:cust", "dep:regex", "dep:oneshot", "dep:safer_owning_ref"]
 kernel = ["dep:rust-cuda-kernel"]
 
 [dependencies]
 const-type-layout = { workspace = true, features = ["derive"] }
+cust = { workspace = true, optional = true }
+cust_core = { workspace = true }
 final = { workspace = true, optional = true }
 oneshot = { workspace = true, features = ["std", "async"], optional = true }
 regex = { workspace = true, optional = true }
-rustacuda = { workspace = true, optional = true }
-rustacuda_core = { workspace = true }
 rust-cuda-derive = { workspace = true, optional = true }
 rust-cuda-kernel = { workspace = true, optional = true }
 safer_owning_ref = { workspace = true, optional = true }

README.md

@@ -3,7 +3,7 @@
 [CI Status]: https://img.shields.io/github/actions/workflow/status/juntyr/rust-cuda/ci.yml?branch=main
 [workflow]: https://github.com/juntyr/rust-cuda/actions/workflows/ci.yml?query=branch%3Amain
 
-[MSRV]: https://img.shields.io/badge/MSRV-1.81.0--nightly-orange
+[MSRV]: https://img.shields.io/badge/MSRV-1.84.0--nightly-orange
 [repo]: https://github.com/juntyr/rust-cuda
 
 [Rust Doc]: https://img.shields.io/badge/docs-main-blue

examples/lifetime/src/main.rs

@@ -2,30 +2,30 @@
 
 use lifetime::{kernel, link};
 
-fn main() -> rust_cuda::deps::rustacuda::error::CudaResult<()> {
+fn main() -> rust_cuda::deps::cust::error::CudaResult<()> {
     // Link the lifetime-only-generic CUDA kernel
     struct KernelPtx<'a, 'b>(core::marker::PhantomData<(&'a (), &'b ())>);
     link! { impl kernel<'a, 'b> for KernelPtx }
 
     // Initialize the CUDA API
-    rust_cuda::deps::rustacuda::init(rust_cuda::deps::rustacuda::CudaFlags::empty())?;
+    rust_cuda::deps::cust::init(rust_cuda::deps::cust::CudaFlags::empty())?;
 
     // Get the first CUDA GPU device
-    let device = rust_cuda::deps::rustacuda::device::Device::get_device(0)?;
+    let device = rust_cuda::deps::cust::device::Device::get_device(0)?;
 
     // Create a CUDA context associated to this device
     let _context = rust_cuda::host::CudaDropWrapper::from(
-        rust_cuda::deps::rustacuda::context::Context::create_and_push(
-            rust_cuda::deps::rustacuda::context::ContextFlags::MAP_HOST
-                | rust_cuda::deps::rustacuda::context::ContextFlags::SCHED_AUTO,
+        rust_cuda::deps::cust::context::legacy::Context::create_and_push(
+            rust_cuda::deps::cust::context::legacy::ContextFlags::MAP_HOST
+                | rust_cuda::deps::cust::context::legacy::ContextFlags::SCHED_AUTO,
             device,
         )?,
     );
 
     // Create a new CUDA stream to submit kernels to
     let mut stream =
-        rust_cuda::host::CudaDropWrapper::from(rust_cuda::deps::rustacuda::stream::Stream::new(
-            rust_cuda::deps::rustacuda::stream::StreamFlags::NON_BLOCKING,
+        rust_cuda::host::CudaDropWrapper::from(rust_cuda::deps::cust::stream::Stream::new(
+            rust_cuda::deps::cust::stream::StreamFlags::NON_BLOCKING,
             None,
         )?);
 
@@ -34,8 +34,8 @@ fn main() -> rust_cuda::deps::rustacuda::error::CudaResult<()> {
     // Create a new instance of the CUDA kernel and prepare the launch config
     let mut kernel = rust_cuda::kernel::TypedPtxKernel::<kernel>::new::<KernelPtx>(None);
     let config = rust_cuda::kernel::LaunchConfig {
-        grid: rust_cuda::deps::rustacuda::function::GridSize::x(1),
-        block: rust_cuda::deps::rustacuda::function::BlockSize::x(4),
+        grid: rust_cuda::deps::cust::function::GridSize::x(1),
+        block: rust_cuda::deps::cust::function::BlockSize::x(4),
         ptx_jit: false,
     };
 

examples/print/src/main.rs

@@ -2,38 +2,38 @@
 
 use print::{kernel, link, Action};
 
-fn main() -> rust_cuda::deps::rustacuda::error::CudaResult<()> {
+fn main() -> rust_cuda::deps::cust::error::CudaResult<()> {
     // Link the non-generic CUDA kernel
     struct KernelPtx;
     link! { impl kernel for KernelPtx }
 
     // Initialize the CUDA API
-    rust_cuda::deps::rustacuda::init(rust_cuda::deps::rustacuda::CudaFlags::empty())?;
+    rust_cuda::deps::cust::init(rust_cuda::deps::cust::CudaFlags::empty())?;
 
     // Get the first CUDA GPU device
-    let device = rust_cuda::deps::rustacuda::device::Device::get_device(0)?;
+    let device = rust_cuda::deps::cust::device::Device::get_device(0)?;
 
     // Create a CUDA context associated to this device
     let _context = rust_cuda::host::CudaDropWrapper::from(
-        rust_cuda::deps::rustacuda::context::Context::create_and_push(
-            rust_cuda::deps::rustacuda::context::ContextFlags::MAP_HOST
-                | rust_cuda::deps::rustacuda::context::ContextFlags::SCHED_AUTO,
+        rust_cuda::deps::cust::context::legacy::Context::create_and_push(
+            rust_cuda::deps::cust::context::legacy::ContextFlags::MAP_HOST
+                | rust_cuda::deps::cust::context::legacy::ContextFlags::SCHED_AUTO,
             device,
         )?,
     );
 
     // Create a new CUDA stream to submit kernels to
     let mut stream =
-        rust_cuda::host::CudaDropWrapper::from(rust_cuda::deps::rustacuda::stream::Stream::new(
-            rust_cuda::deps::rustacuda::stream::StreamFlags::NON_BLOCKING,
+        rust_cuda::host::CudaDropWrapper::from(rust_cuda::deps::cust::stream::Stream::new(
+            rust_cuda::deps::cust::stream::StreamFlags::NON_BLOCKING,
             None,
         )?);
 
     // Create a new instance of the CUDA kernel and prepare the launch config
     let mut kernel = rust_cuda::kernel::TypedPtxKernel::<kernel>::new::<KernelPtx>(None);
     let config = rust_cuda::kernel::LaunchConfig {
-        grid: rust_cuda::deps::rustacuda::function::GridSize::x(1),
-        block: rust_cuda::deps::rustacuda::function::BlockSize::x(4),
+        grid: rust_cuda::deps::cust::function::GridSize::x(1),
+        block: rust_cuda::deps::cust::function::BlockSize::x(4),
         ptx_jit: false,
     };
 

rust-cuda-derive/src/lib.rs

@@ -5,7 +5,7 @@
 //! [CI Status]: https://img.shields.io/github/actions/workflow/status/juntyr/rust-cuda/ci.yml?branch=main
 //! [workflow]: https://github.com/juntyr/rust-cuda/actions/workflows/ci.yml?query=branch%3Amain
 //!
-//! [MSRV]: https://img.shields.io/badge/MSRV-1.81.0--nightly-orange
+//! [MSRV]: https://img.shields.io/badge/MSRV-1.84.0--nightly-orange
 //! [repo]: https://github.com/juntyr/rust-cuda
 //!
 //! [Rust Doc]: https://img.shields.io/badge/docs-main-blue

rust-cuda-derive/src/rust_to_cuda/field_ty.rs

@@ -1,6 +1,5 @@
 use syn::{parse_quote, spanned::Spanned};
 
-#[expect(clippy::module_name_repetitions)]
 pub enum CudaReprFieldTy {
     SafeDeviceCopy,
     RustToCuda {

rust-cuda-derive/src/rust_to_cuda/impl.rs

@@ -84,7 +84,7 @@ pub fn rust_to_cuda_trait(
             unsafe fn borrow<CudaAllocType: #crate_path::alloc::CudaAlloc>(
                 &self,
                 alloc: CudaAllocType,
-            ) -> #crate_path::deps::rustacuda::error::CudaResult<(
+            ) -> #crate_path::deps::cust::error::CudaResult<(
                 #crate_path::utils::ffi::DeviceAccessible<Self::CudaRepresentation>,
                 #crate_path::alloc::CombinedCudaAlloc<Self::CudaAllocation, CudaAllocType>
             )> {
@@ -107,7 +107,7 @@ pub fn rust_to_cuda_trait(
                 alloc: #crate_path::alloc::CombinedCudaAlloc<
                     Self::CudaAllocation, CudaAllocType
                 >,
-            ) -> #crate_path::deps::rustacuda::error::CudaResult<CudaAllocType> {
+            ) -> #crate_path::deps::cust::error::CudaResult<CudaAllocType> {
                 let (alloc_front, alloc_tail) = alloc.split();
 
                 #(#r2c_field_destructors)*
@@ -192,7 +192,7 @@ pub fn rust_to_cuda_async_trait(
                 &self,
                 alloc: CudaAllocType,
                 stream: #crate_path::host::Stream<'stream>,
-            ) -> #crate_path::deps::rustacuda::error::CudaResult<(
+            ) -> #crate_path::deps::cust::error::CudaResult<(
                 #crate_path::utils::r#async::Async<
                     '_, 'stream,
                     #crate_path::utils::ffi::DeviceAccessible<Self::CudaRepresentation>,
@@ -220,7 +220,7 @@ pub fn rust_to_cuda_async_trait(
                     Self::CudaAllocationAsync, CudaAllocType
                 >,
                 stream: #crate_path::host::Stream<'stream>,
-            ) -> #crate_path::deps::rustacuda::error::CudaResult<(
+            ) -> #crate_path::deps::cust::error::CudaResult<(
                 #crate_path::utils::r#async::Async<
                     'a, 'stream,
                     #crate_path::deps::owning_ref::BoxRefMut<'a, CudaRestoreOwner, Self>,

rust-cuda-derive/src/rust_to_cuda/mod.rs

@@ -10,7 +10,7 @@ fn get_cuda_repr_ident(rust_repr_ident: &proc_macro2::Ident) -> proc_macro2::Ide
     format_ident!("{}CudaRepresentation", rust_repr_ident)
 }
 
-#[expect(clippy::module_name_repetitions, clippy::too_many_lines)]
+#[expect(clippy::too_many_lines)]
 pub fn impl_rust_to_cuda(ast: &syn::DeriveInput) -> proc_macro::TokenStream {
     let (mut struct_fields_cuda, struct_semi_cuda) = if let syn::Data::Struct(s) = &ast.data {
         (s.fields.clone(), s.semi_token)

rust-cuda-kernel/build.rs

@@ -5,7 +5,7 @@
 //! [CI Status]: https://img.shields.io/github/actions/workflow/status/juntyr/rust-cuda/ci.yml?branch=main
 //! [workflow]: https://github.com/juntyr/rust-cuda/actions/workflows/ci.yml?query=branch%3Amain
 //!
-//! [MSRV]: https://img.shields.io/badge/MSRV-1.81.0--nightly-orange
+//! [MSRV]: https://img.shields.io/badge/MSRV-1.84.0--nightly-orange
 //! [repo]: https://github.com/juntyr/rust-cuda
 //!
 //! [Rust Doc]: https://img.shields.io/badge/docs-main-blue

rust-cuda-kernel/src/kernel/link/mod.rs

@@ -321,7 +321,7 @@ fn check_kernel_ptx_and_report(
         Ok(Some(binary)) => {
             if ptx_lint_levels
                 .get(&PtxLint::DumpAssembly)
-                .map_or(false, |level| *level > LintLevel::Allow)
+                .is_some_and(|level| *level > LintLevel::Allow)
             {
                 const HEX: [char; 16] = [
                     '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd', 'e', 'f',
@@ -336,7 +336,7 @@ fn check_kernel_ptx_and_report(
 
                 if ptx_lint_levels
                     .get(&PtxLint::DumpAssembly)
-                    .map_or(false, |level| *level > LintLevel::Warn)
+                    .is_some_and(|level| *level > LintLevel::Warn)
                 {
                     emit_call_site_error!(
                         "{} compiled binary:\n{}\n\n{}",
@@ -459,31 +459,31 @@ fn check_kernel_ptx(
 
             if ptx_lint_levels
                 .get(&PtxLint::Verbose)
-                .map_or(false, |level| *level > LintLevel::Warn)
+                .is_some_and(|level| *level > LintLevel::Warn)
             {
                 options.push(c"--verbose");
             }
             if ptx_lint_levels
                 .get(&PtxLint::DoublePrecisionUse)
-                .map_or(false, |level| *level > LintLevel::Warn)
+                .is_some_and(|level| *level > LintLevel::Warn)
             {
                 options.push(c"--warn-on-double-precision-use");
             }
             if ptx_lint_levels
                 .get(&PtxLint::LocalMemoryUse)
-                .map_or(false, |level| *level > LintLevel::Warn)
+                .is_some_and(|level| *level > LintLevel::Warn)
             {
                 options.push(c"--warn-on-local-memory-usage");
             }
             if ptx_lint_levels
                 .get(&PtxLint::RegisterSpills)
-                .map_or(false, |level| *level > LintLevel::Warn)
+                .is_some_and(|level| *level > LintLevel::Warn)
             {
                 options.push(c"--warn-on-spills");
             }
             if ptx_lint_levels
                 .get(&PtxLint::DynamicStackSize)
-                .map_or(true, |level| *level <= LintLevel::Warn)
+                .is_none_or(|level| *level <= LintLevel::Warn)
             {
                 options.push(c"--suppress-stack-size-warning");
             }
@@ -505,31 +505,31 @@ fn check_kernel_ptx(
 
         if ptx_lint_levels
             .get(&PtxLint::Verbose)
-            .map_or(false, |level| *level > LintLevel::Allow)
+            .is_some_and(|level| *level > LintLevel::Allow)
         {
             options.push(c"--verbose");
         }
         if ptx_lint_levels
             .get(&PtxLint::DoublePrecisionUse)
-            .map_or(false, |level| *level > LintLevel::Allow)
+            .is_some_and(|level| *level > LintLevel::Allow)
         {
             options.push(c"--warn-on-double-precision-use");
         }
         if ptx_lint_levels
             .get(&PtxLint::LocalMemoryUse)
-            .map_or(false, |level| *level > LintLevel::Allow)
+            .is_some_and(|level| *level > LintLevel::Allow)
         {
             options.push(c"--warn-on-local-memory-usage");
         }
         if ptx_lint_levels
             .get(&PtxLint::RegisterSpills)
-            .map_or(false, |level| *level > LintLevel::Allow)
+            .is_some_and(|level| *level > LintLevel::Allow)
         {
             options.push(c"--warn-on-spills");
         }
         if ptx_lint_levels
             .get(&PtxLint::DynamicStackSize)
-            .map_or(true, |level| *level < LintLevel::Warn)
+            .is_none_or(|level| *level < LintLevel::Warn)
         {
             options.push(c"--suppress-stack-size-warning");
         }

rust-cuda-kernel/src/kernel/lints.rs

@@ -180,7 +180,7 @@ pub trait NestedMetaParser {
     ) -> syn::Result<()>;
 }
 
-impl<'a> NestedMetaParser for syn::meta::ParseNestedMeta<'a> {
+impl NestedMetaParser for syn::meta::ParseNestedMeta<'_> {
     fn path(&self) -> &syn::Path {
         &self.path
     }

rust-cuda-kernel/src/lib.rs

@@ -5,7 +5,7 @@
 //! [CI Status]: https://img.shields.io/github/actions/workflow/status/juntyr/rust-cuda/ci.yml?branch=main
 //! [workflow]: https://github.com/juntyr/rust-cuda/actions/workflows/ci.yml?query=branch%3Amain
 //!
-//! [MSRV]: https://img.shields.io/badge/MSRV-1.81.0--nightly-orange
+//! [MSRV]: https://img.shields.io/badge/MSRV-1.84.0--nightly-orange
 //! [repo]: https://github.com/juntyr/rust-cuda
 //!
 //! [Rust Doc]: https://img.shields.io/badge/docs-main-blue

rust-toolchain

@@ -1,5 +1,5 @@
 [toolchain]
-# Pin to final 1.81.0 nightly
-channel = "nightly-2024-07-21"
+# Pin to final 1.84.0 nightly
+channel = "nightly-2024-11-22"
 components = [ "cargo", "rustfmt", "clippy", "llvm-bitcode-linker", "llvm-tools" ]
-targets = [ "x86_64-unknown-linux-gnu", "nvptx64-nvidia-cuda" ]
+targets = [ "nvptx64-nvidia-cuda" ]

src/deps.rs

@@ -7,6 +7,6 @@ pub extern crate const_type_layout;
 pub extern crate owning_ref;
 
 #[cfg(feature = "host")]
-pub extern crate rustacuda;
+pub extern crate cust;
 
-pub extern crate rustacuda_core;
+pub extern crate cust_core;