OpenE2K/gcc
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

* reorg.c: Remove comments about dead ports.

Browse Source 
From-SVN: r78280
This commit is contained in:
Kazu Hirata 2004-02-22 22:33:12 +00:00 committed by Kazu Hirata
parent 56d03a4010
commit fcfcb20e67
2 changed files with 6 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
gcc/ChangeLog
View File

@ -1,3 +1,7 @@
2004-02-22 Kazu Hirata <kazu@cs.umass.edu>
* reorg.c: Remove comments about dead ports.
2004-02-22 Christopher Faylor <cgf@redhat.com>
* config.gcc (i[34567]86-*-pe|i[34567]86-*-cygwin*): *Really* specify

13
gcc/reorg.c
View File

@ -36,19 +36,11 @@ Software Foundation, 59 Temple Place - Suite 330, Boston, MA
cycle, and the Branch Penalty is 0. Several RISC machines approach
branch delays differently:
The MIPS and AMD 29000 have a single branch delay slot. Most insns
The MIPS has a single branch delay slot. Most insns
(except other branches) can be used to fill this slot. When the
slot is filled, two insns execute in two cycles, reducing the
branch penalty to zero.
The Motorola 88000 conditionally exposes its branch delay slot,
so code is shorter when it is turned off, but will run faster
when useful insns are scheduled there.
The IBM ROMP has two forms of branch and call insns, both with and
without a delay slot. Much like the 88k, insns not using the delay
slot can be shorted (2 bytes vs. 4 bytes), but will run slowed.
The SPARC always has a branch delay slot, but its effects can be
annulled when the branch is not taken. This means that failing to
find other sources of insns, we can hoist an insn from the branch
@ -84,8 +76,7 @@ Software Foundation, 59 Temple Place - Suite 330, Boston, MA
`fill_eager_delay_slots' tries to guess the direction the branch
will go; if it guesses right 100% of the time, it can reduce the
branch penalty as much as `fill_simple_delay_slots' does. If it
guesses wrong 100% of the time, it might as well schedule nops (or
on the m88k, unexpose the branch slot). When
guesses wrong 100% of the time, it might as well schedule nops. When
`fill_eager_delay_slots' takes insns from the fall-through path of
the jump, usually there is no code expansion; when it takes insns
from the branch target, there is code expansion if it is not the