On 2021-06-28 21:18, Alessandro Motta wrote:> One thing that is still puzzling / worrying me: I completely failed to > identify the actual bottleneck when profiling the code. > > Did I simply misinterpret the profiling output / flame graph? Or is the > problem rather that memory allocations and garbage collection are not > tracked by the profiler?
Some things that might help (at least they helped me :-) ) ... - Use the `profile-thunk` function described at https://docs.racket-lang.org/profile/ (most likely that's how you got your profile data to begin with). - Use the `errortrace` option, i. e. run the instrumented program with `racket -l errortrace -t program.rkt`. The errortrace functionality is described as: When using Errortrace, profiles are more precise and more fine-grained (expression-level instead of function-level) but profiling has higher overhead and recompilation may be necessary. - Note the remark about recompilation. It took me a bit to realize that I had to delete the `.zo` files. :-) In practice, I remove the `compiled` folders for the program I want to run. If you don't remove the compiled code, the old code - without errortrace information - will be used, but as far as I remember you'll get no error message or warning. - There's also an option for the sample interval, so you can lower it to get finer-grained measurements. I don't find this now. Maybe someone else can point to more information. - Learn to read the text output of the profiler as described under https://docs.racket-lang.org/profile/#%28part._.Textual_.Rendering%29 Especially the distinction between the times with and without calls to nested functions is helpful. The profiler output may seem a bit overwhelming first, but with some time and experimentation it becomes better (like with so many things regarding Racket ;-) ). Stefan -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/racket-users/08b4bcb1-145f-0824-55c0-05bf6641bf06%40sschwarzer.net.
