[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Making EMACS run faster.
- To: bug-emacs@MIT-XX.ARPA, bug-teco@MIT-XX.ARPA
- Subject: Making EMACS run faster.
- From: William "Chops" Westfield <BILLW@SRI-AI.ARPA>
- Date: Fri 10 Feb 84 16:22:38-PST
This concerns people who are running TENEX EMACS on FOONLY or 2020
TECO currently executes code in the ACs in order to try to make
a string copy loop faster (on KA and KI processors, this was
probably a good idea). On a KL processor with Cache memory,
this probably makes little difference. On a Foonly computer,
and presumably on a 2020, since the technology involved is
similar, this is a VERY BAD idea. The relevant code has been
shown to run 10 to 15 times SLOWER in the ACs of a foonly F4
than the same code run from memory. There are technical reasons
for this, but the bottom line is that unless you have a KA or KI,
the string copy loop should be left in memory.
I have not made the actual patches. the relevant portions of code
are near YANK3 and PCHF2