Optimize the cat operation on contiguous tensors (#1855)

* Add a specialized kernel for copy2d.

* Move the cat operations.

* Avoid transpositions in cat.

* Bugfix.

* Bugfix for the cuda kernel.

* Add a benchmark.

* Add more testing.

* Test fix.

* Faster kernel.

* Add the missing kernel.

* Tweak the test.

* Add a metal kernel.

* Fix for the metal kernel.

* Get the tests to pass on metal.

* Also use this opportunity to fix the metal kernel for ELU.

* Add some bf16 kernels.

* Clippy fixes.

candle-core/src/dummy_cuda_backend.rs

@@ -154,6 +154,19 @@ impl crate::backend::BackendStorage for CudaStorage {
         Err(Error::NotCompiledWithCudaSupport)
     }
 
+    fn copy2d(
+        &self,
+        _: &mut Self,
+        _: usize,
+        _: usize,
+        _: usize,
+        _: usize,
+        _: usize,
+        _: usize,
+    ) -> Result<()> {
+        Err(Error::NotCompiledWithCudaSupport)
+    }
+
     fn avg_pool2d(&self, _: &Layout, _: (usize, usize), _: (usize, usize)) -> Result<Self> {
         Err(Error::NotCompiledWithCudaSupport)
     }