| Summary: | [3.4/4.0 Regression] useless copies of floating point operands | ||
|---|---|---|---|
| Product: | gcc | Reporter: | Sylvain Pion <pion.sylvain> |
| Component: | middle-end | Assignee: | Richard Henderson <rth> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | gcc-bugs |
| Priority: | P2 | Keywords: | missed-optimization |
| Version: | 3.4.0 | ||
| Target Milestone: | 3.4.1 | ||
| Host: | i686-pc-linux-gnu | Target: | i686-pc-linux-gnu |
| Build: | i686-pc-linux-gnu | Known to work: | 2.95.3 |
| Known to fail: | 3.0.4 3.2.3 3.3.3 3.4.0 4.0.0 | Last reconfirmed: | 2004-04-30 15:57:08 |
I still have gcc 2.95 around, so I have tested with it as well.
It produces the better code that I expected :
force2memdouble:
#APP
#NO_APP
fldl 4(%esp)
ret
So, if the code is correct, then this problem is a regression on code quality
compared to 2.95.
I want to mention that this asm() is critical when implementing interval
arithmetic. Also, GMP uses that kind of asm() in mpz_get_d_2exp.
Note before 3.3 and after 2.95.3, the following sequence was used to copy the FP:
subl $12, %esp
movl 16(%esp), %edx
movl 20(%esp), %ecx
movl %edx, (%esp)
movl %ecx, 4(%esp)
Confirmed as a regression.
Will not fix in 3.3.x. -- Gaby Richard -- Would you please take a look at this? Thanks, -- Mark Testing a patch. Subject: Bug 15228 CVSROOT: /cvs/gcc Module name: gcc Branch: gcc-3_4-branch Changes by: rth@gcc.gnu.org 2004-06-08 16:44:56 Modified files: gcc : ChangeLog function.c Log message: PR middle-end/15228 * function.c (assign_parms): Always set_mem_align with the computed FUNCTION_ARG_BOUNDARY. Don't clear stack_parm if !STRICT_ALIGNMENT. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=2.2326.2.481&r2=2.2326.2.482 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/function.c.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.483.4.13&r2=1.483.4.14 Subject: Bug 15228 CVSROOT: /cvs/gcc Module name: gcc Changes by: rth@gcc.gnu.org 2004-06-09 23:52:15 Modified files: gcc : ChangeLog function.c Log message: PR middle-end/15228 * function.c (assign_parms): Always set_mem_align with the computed FUNCTION_ARG_BOUNDARY. Don't clear stack_parm if !STRICT_ALIGNMENT. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.3901&r2=2.3902 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/function.c.diff?cvsroot=gcc&r1=1.520&r2=1.521 Fixed. |
Consider the following function : // Forces a double to memory double force2memdouble(double x) { asm("" : "=m"(x) : "m"(x)); return x; } This is used to force a double value in memory for platforms like x86 where the FP registers have too much precision. It gets compiled by GCC (3.4 and 3.3.3, with -O2 -fomit-frame-pointer) to : force2memdouble: subl $12, %esp fldl 16(%esp) fstpl (%esp) fldl (%esp) addl $12, %esp ret I believe that the store/load is useless, and that the function could be compiled as simply : force2memdouble: fldl 4(%esp) ret It is interesting to note that when using "int", "float" or "long double" instead of "double", then there is no such useless store/load. So, is this something that could be improved in GCC ? Maybe the code responsible for asm() adds too many requirements on the operands for doubles ?