Some notes from dt-testbed/proto2/README. I haven't committed
changes yet:
For a second effort at dynamic branches, I handcoded an assembly
handler which performs the target address hash function and then
searches each of four entries in the set. If no match is found,
the C handler code is then called.
(DT/native) (for reference)
effort1 effort0
workload microloops native DT factor factor
----------------------------------------------------------------
NOP .52 5.90 11.3 [18.6]
add cascade 5 1.87 7.29 3.9 [ 5.9]
add cascade 10 3.59 9.16 2.6 [ 3.6]
add cascade 100 27.24 32.71 1.2 [ 1.3]
These numbers show a little better performance for dynamic branches.
It took only about 5 loops through the add cascade to average out
the slow-down factor to a similar level as 10 loops in the previous
table.
I'll have a further into other possible enhancements.
-Kevin
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Kevin Lawton [EMAIL PROTECTED]
MandrakeSoft, Inc. Plex86 developer
http://www.linux-mandrake.com/ http://www.plex86.org/
