On Tuesday, June 30, 2015 at 1:11:55 PM UTC-4, jean-pierre.muench wrote: > > Hey Jeff, > > 1. Shouldn't CRYPTOPP_RW_USE_OMP live in config.h and be a #define > rather than a "static const bool"? > > Good points here. Use of "static const bool" vs #define is just the C++ way to do things. OpenMP uses a "bool" on the if clause, so we can include the type. If it was guarded in a block, I would have used a define.
The config.h question is open. Since the modification only affected this class, I wanted to move into into the implementation to keep the change as local as possible. Maybe we need to think about it some more.... > > 1. Why do you default-*enable* OMP, if it gives us worse performance > than non-OMP (or did you fix it without telling us?)? > > The indecision here is two part: (1) the library's existing behavior is to use OMP by default, and (2) I was only able to test on a handful of platforms. I think we should optimize for the common case, so I think that means we should disable it by default on this class. > > 1. May you please change my name to Muench rather than Munch? > > Yes, will do. Sorry about that. The idea is to keep the source files 8-bit clean so "things just work". I know the diacritics in Jean-Pierre Münch causes problems in two editors and the Java compiler (javac tries to interpret comments), so I dropped the diacritics. Sorry again. -- -- You received this message because you are subscribed to the "Crypto++ Users" Google Group. To unsubscribe, send an email to [email protected]. More information about Crypto++ and this group is available at http://www.cryptopp.com. --- You received this message because you are subscribed to the Google Groups "Crypto++ Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
