permute and tiled transpose kernel

crates/cubecl-cuda/src/compute/server.rs

@@ -368,7 +368,8 @@ impl ComputeServer for CudaServer {
                     pixels_per_column: _,
                 } => {
                     return Err(LaunchError::Unknown {
-                        context: "CUDA version 12.8 required for tensor map format Im2colWide".into,
+                        context: "CUDA version 12.8 required for tensor map format Im2colWide"
+                            .into(),
                     });
                 }
             };

examples/permute/Cargo.toml

@@ -0,0 +1,16 @@
+[package]
+name = "permute-example"
+version = "0.1.0"
+edition = "2021"
+publish = false
+
+[dependencies]
+cubecl = { path = "../../crates/cubecl", version = "=0.9.0-pre.3" }
+cubecl-core = { path = "../../crates/cubecl-core", default-features = false }
+cubecl-std = { path = "../../crates/cubecl-std", default-features = false }
+
+
+[features]
+default = []
+wgpu = ["cubecl/wgpu"]
+cuda = ["cubecl/cuda"]