From mboxrd@z Thu Jan 1 00:00:00 1970 From: Brad Lucier To: gcc@gcc.gnu.org Cc: lucier@math.purdue.edu (Brad Lucier) Subject: Some statement counts for gcc Date: Sun, 25 Aug 2002 16:04:00 -0000 Message-id: <200208252304.g7PN4Bc18560@banach.math.purdue.edu> X-SW-Source: 2002-08/msg01559.html As far as I know, the algorithm problems in -fnew-ra haven't been fixed yet, and I wanted to help by giving some more data. Here are some statement counts for banach-119% gcc/xgcc -v Reading specs from /home/c/lucier/local/gcc-test/lib/gcc-lib/sparcv9-sun-solaris2.8/3.3/specs Configured with: ../configure --prefix=/home/c/lucier/local/gcc-test --enable-languages=c --enable-checking=no --enable-coverage sparcv9-sun-solaris2.8 Thread model: posix gcc version 3.3 20020823 (experimental) I ran this job banach-113% gcc/cc1 -fnew-ra -m64 -O1 -fschedule-insns2 -fno-strict-aliasing -fno-math-errno -mcpu=ultrasparc -mtune=ultrasparc _repl.i {GC 5344k -> 1146k} ___H__20___repl {GC 27541k -> 9605k} {GC 14111k -> 9356k} {GC 15065k -> 10374k} {GC 16938k -> 10335k} {GC 18900k -> 10618k} {GC 14579k -> 11902k} {GC 20160k -> 12347k} ___init_proc ____20___repl Execution times (seconds) garbage collection : 3.68 ( 1%) usr 0.01 ( 0%) sys 2.00 ( 0%) wall cfg construction : 3.94 ( 1%) usr 0.35 ( 4%) sys 5.00 ( 1%) wall cfg cleanup : 13.53 ( 3%) usr 0.00 ( 0%) sys 13.50 ( 3%) wall trivially dead code : 1.58 ( 0%) usr 0.00 ( 0%) sys 1.50 ( 0%) wall life analysis : 37.20 ( 7%) usr 0.00 ( 0%) sys 37.00 ( 7%) wall life info update : 10.31 ( 2%) usr 0.00 ( 0%) sys 9.50 ( 2%) wall preprocessing : 1.21 ( 0%) usr 1.17 (15%) sys 4.50 ( 1%) wall lexical analysis : 0.70 ( 0%) usr 2.21 (28%) sys 4.00 ( 1%) wall parser : 5.55 ( 1%) usr 1.33 (17%) sys 4.50 ( 1%) wall expand : 1.46 ( 0%) usr 0.05 ( 1%) sys 1.50 ( 0%) wall varconst : 0.69 ( 0%) usr 0.03 ( 0%) sys 0.50 ( 0%) wall integration : 0.37 ( 0%) usr 0.01 ( 0%) sys 0.50 ( 0%) wall jump : 7.99 ( 2%) usr 0.00 ( 0%) sys 8.00 ( 2%) wall CSE : 3.98 ( 1%) usr 0.00 ( 0%) sys 4.50 ( 1%) wall loop analysis : 0.03 ( 0%) usr 0.00 ( 0%) sys 0.00 ( 0%) wall branch prediction : 165.11 (33%) usr 0.07 ( 1%) sys 165.50 (33%) wall flow analysis : 0.72 ( 0%) usr 0.00 ( 0%) sys 1.00 ( 0%) wall combiner : 6.41 ( 1%) usr 0.00 ( 0%) sys 6.00 ( 1%) wall if-conversion : 0.87 ( 0%) usr 0.00 ( 0%) sys 1.00 ( 0%) wall local alloc : 204.74 (41%) usr 2.77 (34%) sys 208.50 (41%) wall global alloc : 5.34 ( 1%) usr 0.00 ( 0%) sys 6.00 ( 1%) wall reload CSE regs : 9.18 ( 2%) usr 0.00 ( 0%) sys 8.50 ( 2%) wall flow 2 : 0.34 ( 0%) usr 0.00 ( 0%) sys 0.50 ( 0%) wall if-conversion 2 : 0.90 ( 0%) usr 0.00 ( 0%) sys 1.00 ( 0%) wall rename registers : 3.23 ( 1%) usr 0.00 ( 0%) sys 4.00 ( 1%) wall scheduling 2 : 3.33 ( 1%) usr 0.00 ( 0%) sys 3.50 ( 1%) wall delay branch sched : 1.89 ( 0%) usr 0.00 ( 0%) sys 2.50 ( 0%) wall shorten branches : 0.24 ( 0%) usr 0.00 ( 0%) sys 0.00 ( 0%) wall final : 1.08 ( 0%) usr 0.01 ( 0%) sys 1.00 ( 0%) wall rest of compilation : 1.97 ( 0%) usr 0.00 ( 0%) sys 1.50 ( 0%) wall TOTAL : 497.61 8.03 507.00 where _repl.i.gz is at http://www.math.purdue.edu/~lucier/_repl.i.gz The complete gcov files are at http://www.math.purdue.edu/~lucier/gcovs.tgz I wanted to compile http://www.math.purdue.edu/~lucier/_io.i.gz but at 5-6 hours per run, using up to 3.8 GB swap, and with me making one mistake or another three times in a row, I thought I'd start with the smaller file ;-). I sorted the executed lines in decreasing order, see http://www.math.purdue.edu/~lucier/sorted-lines.gz A surprising (to me) number of lines in real.c are executed; I might look there to see what's going on. Brad