OpenE2K/binutils-gdb
12
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

sim: bfin: speed up all insn testcases slightly

Browse Source 
The main body of the "all insn" test is executed once per tested insn, and
we test millions of insns here.  Any shrinkage we can do in this loop will
speed things up nicely (since it's multiplied per tested insn).

To that end, simplify the end-of-table test into one less insn, and omit
the SSYNC when we build for the sim.  When we build to run on the hardware,
this insn matters, but the sim doesn't have write store buffers in the chip
that might get in the way (memory writes are atomic).
This commit is contained in:
Mike Frysinger 2013-06-24 01:44:55 +00:00
parent cdaf73c68a
commit 531d5282c0
2 changed files with 8 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
sim/testsuite/sim/bfin/ChangeLog
View File

@ -1,3 +1,8 @@
2013-06-23 Mike Frysinger <vapier@gentoo.org>
* se_allopcodes.h (_match): Simplify register test to one less insn.
Omit the SSYNC insn when compiling for the sim.
2013-06-23 Mike Frysinger <vapier@gentoo.org>
* testutils.inc: Trim trailing whitespace.

5
sim/testsuite/sim/bfin/se_allopcodes.h
View File

@ -102,8 +102,7 @@ _match:
se_all_load_table
/* is this the end of the table? */
R4 = 0;
CC = R4 == R7;
CC = R7 == 0;
IF CC jump _new_instruction;
/* is the opcode (R0) greater than the 2nd entry in the table (R6) */
@ -168,8 +167,10 @@ _legal_instruction:
_next_instruction:
se_all_next_insn
.ifdef BFIN_JTAG
/* Make sure the opcode isn't in a write buffer */
SSYNC;
.endif
R1 = P5;
RETX = R1;