Continue replacing vek with glam

Author: jorge-ortega

crates/cuda_std/src/thread.rs

@@ -154,7 +154,7 @@ pub fn grid_dim_z() -> u32 {
 #[inline(always)]
 pub fn thread_idx() -> UVec3 {
     unsafe {
-        Vec3::new(
+        UVec3::new(
             __nvvm_thread_idx_x(),
             __nvvm_thread_idx_y(),
             __nvvm_thread_idx_z(),
@@ -167,7 +167,7 @@ pub fn thread_idx() -> UVec3 {
 #[inline(always)]
 pub fn block_idx() -> UVec3 {
     unsafe {
-        Vec3::new(
+        UVec3::new(
             __nvvm_block_idx_x(),
             __nvvm_block_idx_y(),
             __nvvm_block_idx_z(),
@@ -181,7 +181,7 @@ pub fn block_idx() -> UVec3 {
 #[inline(always)]
 pub fn block_dim() -> UVec3 {
     unsafe {
-        Vec3::new(
+        UVec3::new(
             __nvvm_block_dim_x(),
             __nvvm_block_dim_y(),
             __nvvm_block_dim_z(),
@@ -195,7 +195,7 @@ pub fn block_dim() -> UVec3 {
 #[inline(always)]
 pub fn grid_dim() -> UVec3 {
     unsafe {
-        Vec3::new(
+        UVec3::new(
             __nvvm_grid_dim_x(),
             __nvvm_grid_dim_y(),
             __nvvm_grid_dim_z(),
@@ -206,7 +206,7 @@ pub fn grid_dim() -> UVec3 {
 /// Gets the overall thread index, accounting for 1d/2d/3d block/grid dimensions. This
 /// value is most commonly used for indexing into data and this index is guaranteed to
 /// be unique for every single thread executing this kernel no matter the launch configuration.
-/// 
+///
 /// For very simple kernels it may be faster to use a more simple index calculation, however,
 /// it will be unsound if the kernel launches in a 2d/3d configuration.
 #[gpu_only]
@@ -218,10 +218,10 @@ pub fn index() -> u32 {
     let block_dim = block_dim();
     let thread_idx = thread_idx();
 
-    let block_id = block_idx.x + block_idx.y * grid_dim.x 
+    let block_id = block_idx.x + block_idx.y * grid_dim.x
                        + grid_dim.x * grid_dim.y * block_idx.z;
 
-    block_id * block_dim.product()
+    block_id * block_dim.element_product()
     + (thread_idx.z * (block_dim.x * block_dim.y))
     + (thread_idx.y * block_dim.x) + thread_idx.x
 }

crates/cust/Cargo.toml

@@ -17,7 +17,7 @@ cust_core = { path = "../cust_core", version = "0.1.0"}
 cust_raw = { path = "../cust_raw", default-features = false, features = ["driver"] }
 bitflags = "2.8"
 cust_derive = { path = "../cust_derive", version = "0.2" }
-glam = { version = "0.29.2", features=["cuda"], optional = true }
+glam = { version = "0.30", features=["cuda"], optional = true }
 mint = { version = "^0.5", optional = true }
 num-complex = { version = "0.4.6", optional = true }
 vek = { version = "0.17.1", optional = true, default-features = false }

crates/cust/src/function.rs

@@ -88,6 +88,31 @@ impl From<vek::Vec3<usize>> for GridSize {
     }
 }
 
+#[cfg(feature = "glam")]
+impl From<glam::UVec2> for GridSize {
+    fn from(vec: glam::UVec2) -> Self {
+        GridSize::xy(vec.x, vec.y)
+    }
+}
+#[cfg(feature = "glam")]
+impl From<glam::UVec3> for GridSize {
+    fn from(vec: glam::UVec3) -> Self {
+        GridSize::xyz(vec.x, vec.y, vec.z)
+    }
+}
+#[cfg(feature = "glam")]
+impl From<glam::USizeVec2> for GridSize {
+    fn from(vec: glam::USizeVec2) -> Self {
+        GridSize::xy(vec.x as u32, vec.y as u32)
+    }
+}
+#[cfg(feature = "glam")]
+impl From<glam::USizeVec3> for GridSize {
+    fn from(vec: glam::USizeVec3) -> Self {
+        GridSize::xyz(vec.x as u32, vec.y as u32, vec.z as u32)
+    }
+}
+
 /// Dimensions of a thread block, or the number of threads in a block.
 ///
 /// Each component of a `BlockSize` must be at least 1. The maximum size depends on your device's
@@ -168,6 +193,31 @@ impl From<vek::Vec3<usize>> for BlockSize {
     }
 }
 
+#[cfg(feature = "glam")]
+impl From<glam::UVec2> for BlockSize {
+    fn from(vec: glam::UVec2) -> Self {
+        BlockSize::xy(vec.x, vec.y)
+    }
+}
+#[cfg(feature = "glam")]
+impl From<glam::UVec3> for BlockSize {
+    fn from(vec: glam::UVec3) -> Self {
+        BlockSize::xyz(vec.x, vec.y, vec.z)
+    }
+}
+#[cfg(feature = "glam")]
+impl From<glam::USizeVec2> for BlockSize {
+    fn from(vec: glam::USizeVec2) -> Self {
+        BlockSize::xy(vec.x as u32, vec.y as u32)
+    }
+}
+#[cfg(feature = "glam")]
+impl From<glam::USizeVec3> for BlockSize {
+    fn from(vec: glam::USizeVec3) -> Self {
+        BlockSize::xyz(vec.x as u32, vec.y as u32, vec.z as u32)
+    }
+}
+
 /// All supported function attributes for [Function::get_attribute](struct.Function.html#method.get_attribute)
 #[repr(u32)]
 #[non_exhaustive]

crates/cust_core/Cargo.toml

@@ -9,7 +9,7 @@ readme = "../../README.md"
 
 [dependencies]
 vek = { version = "0.17.1", default-features=false, features=["libm"], optional = true }
-glam = { version = "0.29.2", features=["cuda", "libm"], default-features=false, optional=true }
+glam = { version = "0.30", features=["cuda", "libm"], default-features=false, optional=true }
 mint = { version = "^0.5", optional = true }
 half = { version = "2.4.1", optional = true }
 num-complex = { version = "0.4.6", optional = true }

crates/cust_core/src/lib.rs

@@ -1,5 +1,3 @@
-#![no_std]
-
 pub use _hidden::*;
 pub use cust_derive::DeviceCopyCore as DeviceCopy;
 
@@ -143,6 +141,7 @@ pub mod _hidden {
     {
     }
 
+    #[allow(unused_macros)]
     macro_rules! impl_device_copy_generic {
     ($($($strukt:ident)::+),* $(,)?) => {
         $(
@@ -151,6 +150,7 @@ pub mod _hidden {
     }
 }
 
+    #[allow(unused_macros)]
     macro_rules! impl_device_copy {
     ($($strukt:ty),* $(,)?) => {
         $(
@@ -172,7 +172,22 @@ pub mod _hidden {
 
     #[cfg(feature = "glam")]
     impl_device_copy! {
-        glam::Vec2, glam::Vec3, glam::Vec4, glam::IVec2, glam::IVec3, glam::IVec4,
+        glam::BVec2, glam::BVec3, glam::BVec3A, glam::BVec4, glam::BVec4A,
+        glam::U8Vec2, glam::U8Vec3, glam::U8Vec4,
+        glam::I8Vec2, glam::I8Vec3, glam::I8Vec4,
+        glam::U16Vec2, glam::U16Vec3, glam::U16Vec4,
+        glam::I16Vec2, glam::I16Vec3, glam::I16Vec4,
+        glam::UVec2, glam::UVec3, glam::UVec4,
+        glam::IVec2, glam::IVec3, glam::IVec4,
+        glam::U64Vec2, glam::U64Vec3, glam::U64Vec4,
+        glam::I64Vec2, glam::I64Vec3, glam::I64Vec4,
+        glam::USizeVec2, glam::USizeVec3, glam::USizeVec4,
+        glam::Vec2, glam::Vec3, glam::Vec3A, glam::Vec4,
+        glam::DVec2, glam::DVec3, glam::DVec4,
+        glam::Mat2, glam::Mat3, glam::Mat3A, glam::Mat4,
+        glam::DMat2, glam::DMat3, glam::DMat4,
+        glam::Quat, glam::DQuat,
+        glam::Affine2, glam::Affine3A,
     }
 
     #[cfg(feature = "mint")]

crates/gpu_rand/src/lib.rs

@@ -12,7 +12,6 @@
 #![deny(missing_docs)]
 #![deny(missing_debug_implementations)]
 #![allow(clippy::unreadable_literal)]
-#![cfg_attr(target_os = "cuda", no_std)]
 #![feature(doc_cfg)]
 
 pub mod xoroshiro;

crates/optix/Cargo.toml

@@ -12,7 +12,7 @@ cust = { version = "0.3", path = "../cust", features=["impl_mint"] }
 cust_raw = { path = "../cust_raw", features=["driver"] }
 cfg-if = "1.0.0"
 bitflags = "2.9.0"
-glam = { version = "0.29", features=["cuda", "libm"], default-features=false, optional=true }
+glam = { version = "0.30", features=["cuda", "libm"], default-features=false, optional=true }
 half = { version = "2.4.1", optional = true }
 memoffset = "0.9.1"
 mint = "0.5.9"

crates/optix/examples/ex04_mesh/Cargo.toml

@@ -12,7 +12,7 @@ anyhow = "1.0.44"
 glfw = "0.42.0"
 gl = "0.14.0"
 num-traits = "0.2.14"
-glam = { version = "0.29.2", features=["cuda"] }
+glam = { version = "0.30", features=["cuda"] }
 
 [build-dependencies]
 cuda_builder = { version = "0.3", path = "../../../cuda_builder" }

crates/optix_device/Cargo.toml

@@ -13,6 +13,3 @@ cuda_std = { version = "0.2", path = "../cuda_std" }
 paste = "1.0.15"
 seq-macro = "0.3.5"
 cust_core = { version = "0.1", path = "../cust_core" }
-
-[target.'cfg(not(target_os = "cuda"))'.dependencies]
-glam = { version = "0.29", features = ["cuda"], default-features = false }

crates/optix_device/src/lib.rs

@@ -14,7 +14,8 @@ pub mod trace;
 pub mod transform;
 pub mod util;
 
-use cuda_std::{glam, *};
+pub use cuda_std::glam;
+use cuda_std::*;
 use glam::UVec3;
 pub use misc::*;