|
| From: | Paul Eggert |
| Subject: | Re: Changes in GC and in pure space |
| Date: | Wed, 4 Sep 2019 17:51:06 -0700 |
| User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.8.0 |
Angelo Graziosi wrote:
Maybe all I learned from Borland manuals was wrong.
It is indeed wrong, at least for the buffer.h macros that I recently turned into INLINE functions, as that sped up my usual nontrivial benchmark ('cd lisp; make compile-always') by 5.7% on my platform.
Stefan made a suggestion in this thread that I think can get us another ~5% speedup on the same benchmark. I'm looking into that. It also involves using functions not macros.
| [Prev in Thread] | Current Thread | [Next in Thread] |