Thread safety

This MR introduces a simple check for thread safety in the kernel constrains check. If the assignments are not thread safe, optimizations like OpenMP or GPU should fail.

src/pystencils/config.py

@@ -135,14 +135,9 @@ class CreateKernelConfig:
     """
     skip_independence_check: bool = False
     """
-    Don't check that loop iterations are independent. This is needed e.g. for
-    periodicity kernel, that access the field outside the iteration bounds. Use with care!
-    """
-    check_thread_safety: bool = True
-    """
-    Assignments are considered thread safe if read and writes only target the same locations. If this is not the case,
-    multithreaded optimisations will fail. The thread safety check can be deactivated to use e.g. GPUs anyway.
-    Use with care!
+    By default the assignment list is checked for read/write independence. This means fields are only written at
+    locations where they are read. Doing so guarantees thread safety. In some cases e.g. for
+    periodicity kernel, this can not be assured and does the check needs to be deactivated. Use with care!
     """
 
     class DataTypeFactory: