This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: GCC Release Status (2003-08-22)
- From: Karel Gardas <kgardas at objectsecurity dot com>
- To: Jan Hubicka <hubicka at ucw dot cz>
- Cc: GCC Mailing List <gcc at gcc dot gnu dot org>
- Date: Sat, 30 Aug 2003 23:33:04 +0200 (CEST)
- Subject: Re: GCC Release Status (2003-08-22)
On Sat, 30 Aug 2003, Jan Hubicka wrote:
> > Yes, I know, I should report both issues, but I guess few MBs of
> > preprocessed files are useless for you and on the other hand I've not been
> > able yet to find smaller examples for issues duplication. :-(
>
> What -funit-at-a-time switch does to your compile times?
> In case it does not help, can you send me the preprocessed files
> privately?
Ahoj Honzo,
I've done basic test with -I../include -time -Wall -DPIC -fPIC -c options
and w/ and w/o -funit-at-a-time in MICO's ORB core directory. The results
are in the table below. Sometimes -funit-at-a-time helps sometimes not,
but in the sum it slow down a bit - but it might also be statistical
difference.
File name Normal W/ -fuaat Difference
os-unix.cc 5.81 5.42 0.39
dii.cc 17.08 17.02 0.06
typecode.cc 10.57 10.3 0.27
any.cc 7.86 8.11 -0.25
codec.cc 6.88 7 -0.12
buffer.cc 4.08 4.23 -0.15
context.cc 4.44 4.4 0.04
except.cc 5.84 5.84 0
dispatch.cc 5.76 5.89 -0.13
string.cc 4.28 4.24 0.04
object.cc 5.94 5.99 -0.05
address.cc 6.23 6.36 -0.13
ior.cc 16.67 16.92 -0.25
orb.cc 21.69 21.21 0.48
boa.cc 10.63 10.89 -0.26
dsi.cc 14.29 14.24 0.05
transport.cc 5.44 5.55 -0.11
tr..rt/tcp.cc 5.43 5.35 0.08
tr..rt/udp.cc 5.49 5.5 -0.01
tr..rt/unix.cc 5.29 5.42 -0.13
iop.cc 20.79 20.29 0.5
util.cc 7.06 7.24 -0.18
basic_seq.cc 4.59 4.46 0.13
fast_array.cc 4.83 4.98 -0.15
ssl.cc 13.36 13.55 -0.19
fixed.cc 4.48 4.61 -0.13
intercept.cc 14.26 14.22 0.04
codeset.cc 7.01 7.11 -0.1
queue.cc 5.77 5.82 -0.05
static.cc 24.64 24.74 -0.1
current.cc 12.89 13.02 -0.13
policy_impl.cc 17.18 17.21 -0.03
service_info.cc 12.86 12.81 0.05
ioptypes.cc 14.83 14.75 0.08
ssliop.cc 13.26 13.11 0.15
value.cc 14.95 14.96 -0.01
valuetype.cc 13.88 14.03 -0.15
v..type_impl.cc 16.8 16.89 -0.09
dynany_impl.cc 11.83 11.55 0.28
policy2.cc 13.13 13.14 -0.01
tckind.cc 12.86 12.75 0.11
orb_excepts.cc 13.17 13.12 0.05
policy.cc 13.08 12.99 0.09
poa.cc 17.06 17.1 -0.04
poa_base.cc 14.28 14.42 -0.14
poa_impl.cc 21.91 21.35 0.56
dynany.cc 14.35 14.28 0.07
uni_base64.cc 0.12 0.15 -0.03
uni_unicode.cc 0.22 0.2 0.02
uni_fromuni.cc 0.44 0.46 -0.02
uni_touni.cc 0.42 0.49 -0.07
except2.cc 7.6 7.6 0
pi.cc 14.95 15.03 -0.08
pi_impl.cc 22.8 22.13 0.67
typecode_seq.cc 13.28 13.27 0.01
timebase.cc 12.9 12.91 -0.01
ir.cc 50.98 54.5 -3.52
ir_base.cc 15.24 15.39 -0.15
imr.cc 18.68 19.21 -0.53
mtdebug.cc 4.9 4.9 0
687.34 690.62 -3.28
So just let me know what do you need for better debuging. i.e. do you like
all the files which causes slowdown in preprocessed form?
Thanks,
Karel
--
Karel Gardas kgardas@objectsecurity.com
ObjectSecurity Ltd. http://www.objectsecurity.com