Added compile-time-checked lints against local mem and register spills in CUDA kernels

Author: juntyr

Cargo.lock

@@ -20,7 +20,7 @@ contracts = "0.6.3"
 serde = { version = "1.0", default-features = false, features = ["derive"] }
 
 [target.'cfg(target_os = "cuda")'.dependencies]
-rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "626cd48", features = ["derive"], optional = true }
+rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "6b7c773", features = ["derive"], optional = true }
 
 [target.'cfg(not(target_os = "cuda"))'.dependencies]
-rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "626cd48", features = ["derive", "host"], optional = true }
+rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "6b7c773", features = ["derive", "host"], optional = true }

necsim/core/Cargo.toml

@@ -15,7 +15,7 @@ contracts = "0.6.3"
 serde = { version = "1.0", default-features = false, features = ["derive"] }
 
 [target.'cfg(target_os = "cuda")'.dependencies]
-rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "626cd48", features = ["derive"] }
+rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "6b7c773", features = ["derive"] }
 
 [target.'cfg(not(target_os = "cuda"))'.dependencies]
-rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "626cd48", features = ["derive", "host"] }
+rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "6b7c773", features = ["derive", "host"] }

necsim/impls/cuda/Cargo.toml

@@ -31,7 +31,7 @@ rand_core = "0.6"
 rand_distr = { version = "0.4", default-features = false, features = [] }
 
 [target.'cfg(target_os = "cuda")'.dependencies]
-rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "626cd48", features = ["derive"], optional = true }
+rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "6b7c773", features = ["derive"], optional = true }
 
 [target.'cfg(not(target_os = "cuda"))'.dependencies]
-rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "626cd48", features = ["derive", "host"], optional = true }
+rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "6b7c773", features = ["derive", "host"], optional = true }

necsim/impls/no-std/Cargo.toml

@@ -23,4 +23,4 @@ thiserror = "1.0"
 serde = { version = "1.0", features = ["derive"] }
 serde_state = "0.4"
 serde_derive_state = "0.4"
-rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "626cd48", features = ["host"] }
+rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "6b7c773", features = ["host"] }

rustcoalescence/algorithms/cuda/Cargo.toml

@@ -14,4 +14,4 @@ necsim-impls-no-std = { path = "../../../../necsim/impls/no-std", features = ["c
 necsim-impls-cuda = { path = "../../../../necsim/impls/cuda" }
 rustcoalescence-algorithms-cuda-gpu-kernel = { path = "../gpu-kernel" }
 
-rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "626cd48", features = ["host"] }
+rust-cuda = { git = "https://github.com/juntyr/rust-cuda", rev = "6b7c773", features = ["host"] }

rustcoalescence/algorithms/cuda/cpu-kernel/Cargo.toml

@@ -27,11 +27,14 @@ use rust_cuda::{
         error::{CudaError, CudaResult},
         function::{BlockSize, Function, GridSize},
     },
+    safety::NoAliasing,
 };
 
 use rustcoalescence_algorithms_cuda_gpu_kernel::{
-    BitonicGlobalSortSteppableKernel, BitonicSharedSortPreparableKernel,
-    BitonicSharedSortSteppableKernel, EvenOddSortableKernel, SimulatableKernel,
+    BitonicGlobalSortStepKernelPtx, BitonicGlobalSortSteppableKernel,
+    BitonicSharedSortPrepKernelPtx, BitonicSharedSortPreparableKernel,
+    BitonicSharedSortStepKernelPtx, BitonicSharedSortSteppableKernel, EvenOddSortKernelPtx,
+    EvenOddSortableKernel, SimulatableKernel, SimulationKernelPtx,
 };
 
 mod link;
@@ -83,17 +86,17 @@ pub struct SimulationKernel<
 
 impl<
         M: MathsCore,
-        H: Habitat<M> + RustToCuda,
-        G: Rng<M, Generator: PrimeableRng> + RustToCuda,
-        S: LineageStore<M, H> + RustToCuda,
-        X: EmigrationExit<M, H, G, S> + RustToCuda,
-        D: DispersalSampler<M, H, G> + RustToCuda,
-        C: CoalescenceSampler<M, H, S> + RustToCuda,
-        T: TurnoverRate<M, H> + RustToCuda,
-        N: SpeciationProbability<M, H> + RustToCuda,
-        E: MinSpeciationTrackingEventSampler<M, H, G, S, X, D, C, T, N> + RustToCuda,
-        I: ImmigrationEntry<M> + RustToCuda,
-        A: SingularActiveLineageSampler<M, H, G, S, X, D, C, T, N, E, I> + RustToCuda,
+        H: Habitat<M> + RustToCuda + NoAliasing,
+        G: Rng<M, Generator: PrimeableRng> + RustToCuda + NoAliasing,
+        S: LineageStore<M, H> + RustToCuda + NoAliasing,
+        X: EmigrationExit<M, H, G, S> + RustToCuda + NoAliasing,
+        D: DispersalSampler<M, H, G> + RustToCuda + NoAliasing,
+        C: CoalescenceSampler<M, H, S> + RustToCuda + NoAliasing,
+        T: TurnoverRate<M, H> + RustToCuda + NoAliasing,
+        N: SpeciationProbability<M, H> + RustToCuda + NoAliasing,
+        E: MinSpeciationTrackingEventSampler<M, H, G, S, X, D, C, T, N> + RustToCuda + NoAliasing,
+        I: ImmigrationEntry<M> + RustToCuda + NoAliasing,
+        A: SingularActiveLineageSampler<M, H, G, S, X, D, C, T, N, E, I> + RustToCuda + NoAliasing,
         ReportSpeciation: Boolean,
         ReportDispersal: Boolean,
     > SimulationKernel<M, H, G, S, X, D, C, T, N, E, I, A, ReportSpeciation, ReportDispersal>