We are writing an llvm pass that analyzes huge bitcode files. One problem we faced was that it took a lot of time to parse bitcode files, hindering our debugging time. The following snippet shows the bottleneck and time-consuming task.
llvm_module=parseIRFile(“llvm_bitcode.bc”, error, context);
Can we cache llvm_module, and for the subsequent run, we avoid parsing and reading from the file?