[PYTHON][EINSUM] Now handling reduction sizes that are not a multiple of

TK

lib/runtime/function.cc

@@ -255,6 +255,7 @@ std::unique_ptr<driver::module> function::make_bin(ir::module &module, driver::c
   if(allocation.allocated_size() > context->device()->max_shared_memory())
     return std::unique_ptr<driver::module>();
   barriers.run(module);
+//  ir::print(module, std::cout);
   isel.visit(module, *llvm);
   // return binary
   std::unique_ptr<driver::module> res(driver::module::create(context, std::move(llvm)));