OK, I've pushed a few more relatively minor cleanups to the ARM's FMC I/O code, most significant of which was converting a few heavily used functions to `static inline` to remove the function call overhead. Collectively, these seem to be worth about a 10% improvement in overall RSA signature throughput, not spectacular but worth grabbing.
I have another set of throughput and profiling tests running which I'm happy to share if others want to see them, but after that I think I'm done for a little while, since I think the next step is for Pavel et al to look into synchronized FMC clocking idea we've been discussing. Thanks, all. _______________________________________________ Tech mailing list [email protected] https://lists.cryptech.is/listinfo/tech
