[llvm] [libsycl] Introduce Kernel name based cache (PR #225865)

[via llvm-commits]

================
@@ -185,8 +202,10 @@ ol_symbol_handle_t ProgramAndKernelManager::getOrCreateKernel(
   trackContext(Context);
 
   // Lock order is MDataCollectionMutex -> ContextImpl::MProgramCacheMutex.
-  return Context->getOrCreateKernel(DeviceImage, Device.getOLHandle(),
-                                    KernelInfo.getName());
+  ol_symbol_handle_t Kernel = Context->getOrCreateKernel(
----------------
Robertkq wrote:

Ok, so if I understand correctly:

`MKernels.find(KernelName)` returning successfully is how normally resubmissions get served, with this PR, the cache is queried and will always (for sequential cases) serve the kernel instead of finding it in `MKernels`, however for concurrent cases, it could still end up not existing in the cache yet, for which `MKernel` returning successfully will help. So there is still a path that follows it as far as I understand

https://github.com/llvm/llvm-project/pull/225865