Closed zhujun98 closed 6 years ago
you mean getting data from the simulation?
trains were buffered before, that might be the reason.
Yes, I mean getting data from the simulation.
@tmichela's changes to the simulation code in #42 (just merged) include an option to skip generating the random data, which probably makes it quicker. This is --gen zeros
at the command line.
Thank you for letting me know! I tried it but my cpp client test failed (even without --gen zeros). Maybe something (e.g. name of a key) has been changed and I need to check. Indeed, for the integration test of cpp client, I prefer data with none-zero numbers since a chunk of zeros could also mean reading from some other memory space.
Changes in #44 made this quicker again, hopefully back to a level comparable to what we had before. There's also an option to simulate a single detector module, which is quicker because it involves less data.
The time for my cpp client to finish the function call "next()" quadruples with the new master branch.