Reduce cache-to-execution latency by using lazy unpickling of instructions.

In order to this, invoker generation needs to be cached.

Performance impact on a warm-cache run of test_fmm.py from sumpy

Before

real 43.77
user 44.58
sys 2.96

After

real 37.08
user 37.73
sys 3.02

Closes #67 (closed)

Depends on pytools!12 (merged)

loopy/target/c/c_execution.py

@@ -385,7 +385,7 @@ class CKernelExecutor(KernelExecutorBase):
         super(CKernelExecutor, self).__init__(kernel)
 
     def get_invoker_uncached(self, kernel, codegen_result):
-        generator = CExecutionWrapperGenerator
+        generator = CExecutionWrapperGenerator()
         return generator(kernel, codegen_result)
 
     @memoize_method