public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
* error in alpha ev56
@ 1998-01-02 12:19 pancho
  1998-01-02 16:09 ` Richard Henderson
  0 siblings, 1 reply; 6+ messages in thread
From: pancho @ 1998-01-02 12:19 UTC (permalink / raw)
  To: egcs

hello again, 

here is the list of errors during compilation of egcs in a alpha ev56

xgcc: Internal compiler error: program cc1 got fatal signal 6
make[2]: *** [cse.o] Error 1
make[2]: Leaving directory `/home/pancho/egcs/objdir/gcc'
make[1]: *** [bootstrap] Error 2
make[1]: Leaving directory `/home/pancho/egcs/objdir/gcc'
make: *** [bootstrap] Error 2

any help is appreciated.

regards,

miguel


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: error in alpha ev56
  1998-01-02 12:19 error in alpha ev56 pancho
@ 1998-01-02 16:09 ` Richard Henderson
  1998-01-02 16:09   ` pancho
                     ` (3 more replies)
  0 siblings, 4 replies; 6+ messages in thread
From: Richard Henderson @ 1998-01-02 16:09 UTC (permalink / raw)
  To: pancho; +Cc: egcs

On Fri, Jan 02, 1998 at 08:19:18PM +0000, pancho@galaxy.lca.uevora.pt wrote:
> here is the list of errors during compilation of egcs in a alpha ev56
> 
> xgcc: Internal compiler error: program cc1 got fatal signal 6

There are some deficiencies wrt ev56 support in the binutils 2.8.1.0.1
that was shipped with RH5.0.  I would suggest first bootstrapping EGCS
as --target=alphaev5-unknown-linux-gnu (rather than alphaev56) just to 
be sure that things are sane.

After that, we can see about possibly upgrading binutils and rebuilding
EGCS to take advantage of the byte/word extension on the ev56.


r~

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: error in alpha ev56
  1998-01-02 16:09 ` Richard Henderson
@ 1998-01-02 16:09   ` pancho
  1998-01-02 16:44   ` pancho
                     ` (2 subsequent siblings)
  3 siblings, 0 replies; 6+ messages in thread
From: pancho @ 1998-01-02 16:09 UTC (permalink / raw)
  To: Richard Henderson; +Cc: egcs

Ok Richard,

I'll do that now.
let you know how it goes.

miguel

On Fri, 2 Jan 1998, Richard Henderson wrote:

> On Fri, Jan 02, 1998 at 08:19:18PM +0000, pancho@galaxy.lca.uevora.pt wrote:
> > here is the list of errors during compilation of egcs in a alpha ev56
> > 
> > xgcc: Internal compiler error: program cc1 got fatal signal 6
> 
> There are some deficiencies wrt ev56 support in the binutils 2.8.1.0.1
> that was shipped with RH5.0.  I would suggest first bootstrapping EGCS
> as --target=alphaev5-unknown-linux-gnu (rather than alphaev56) just to 
> be sure that things are sane.
> 
> After that, we can see about possibly upgrading binutils and rebuilding
> EGCS to take advantage of the byte/word extension on the ev56.
> 
> 
> r~
> 


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: error in alpha ev56
  1998-01-02 16:09 ` Richard Henderson
  1998-01-02 16:09   ` pancho
  1998-01-02 16:44   ` pancho
@ 1998-01-02 16:44   ` pancho
  1998-01-05  9:23   ` Slow processing with g77 in egcs pancho
  3 siblings, 0 replies; 6+ messages in thread
From: pancho @ 1998-01-02 16:44 UTC (permalink / raw)
  To: Richard Henderson; +Cc: egcs

BTW I compiled also the snapshot 971225 for the alpha pc164LX and got:

./xgcc -B./  -DIN_GCC    -g -I./include     -DNO_MEM -DNO_LONG_DOUBLE_IO
-O0 \
  -DSYS_FLOAT_H_WRAP=$SYS_FLOAT_H_WRAP \
  -I. -c /home/pancho/egcs/egcs-971225/gcc/enquire.c
/tmp/cca15777.s: Assembler messages:
/tmp/cca15777.s:4: Error: Unknown pseudo-op:  `.arch'
/tmp/cca15777.s:665: Error: macro requires $at register while noat in
effect
/tmp/cca15777.s:678: Error: macro requires $at register while noat in
effect
/tmp/cca15777.s:688: Error: macro requires $at register while noat in
effect
/tmp/cca15777.s:740: Error: macro requires $at register while noat in
effect

.....

/tmp/cca15777.s:5289: Error: macro requires $at register while noat in
effect
/tmp/cca15777.s:5307: Error: macro requires $at register while noat in
effect
/tmp/cca15777.s:5326: Error: macro requires $at register while noat in
effect
make[2]: *** [enquire.o] Error 1
make[2]: Leaving directory `/home/pancho/egcs/objdir/gcc'
make[1]: *** [bootstrap] Error 2
make[1]: Leaving directory `/home/pancho/egcs/objdir/gcc'
make: *** [bootstrap] Error 2

miguel


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: error in alpha ev56
  1998-01-02 16:09 ` Richard Henderson
  1998-01-02 16:09   ` pancho
@ 1998-01-02 16:44   ` pancho
  1998-01-02 16:44   ` pancho
  1998-01-05  9:23   ` Slow processing with g77 in egcs pancho
  3 siblings, 0 replies; 6+ messages in thread
From: pancho @ 1998-01-02 16:44 UTC (permalink / raw)
  To: Richard Henderson; +Cc: egcs

Ok, I recompiled egcs using --target==alphaev5-unknown-linux-gnu
and got the following error:

In file included from tconfig.h:1,
                 from /home/pancho/egcs/egcs-1.0/gcc/libgcc2.c:33:
/home/pancho/egcs/egcs-1.0/gcc/config/alpha/xm-alpha.h:71: string.h: No
such file or directory
make[2]: *** [libgcc2.a] Error 1
make[2]: Leaving directory `/home/pancho/egcs/objdir/gcc'
make[1]: *** [bootstrap] Error 2
make[1]: Leaving directory `/home/pancho/egcs/objdir/gcc'
make: *** [bootstrap] Error 2


miguel


On Fri, 2 Jan 1998, Richard Henderson wrote:

> On Fri, Jan 02, 1998 at 08:19:18PM +0000, pancho@galaxy.lca.uevora.pt wrote:
> > here is the list of errors during compilation of egcs in a alpha ev56
> > 
> > xgcc: Internal compiler error: program cc1 got fatal signal 6
> 
> There are some deficiencies wrt ev56 support in the binutils 2.8.1.0.1
> that was shipped with RH5.0.  I would suggest first bootstrapping EGCS
> as --target=alphaev5-unknown-linux-gnu (rather than alphaev56) just to 
> be sure that things are sane.
> 
> After that, we can see about possibly upgrading binutils and rebuilding
> EGCS to take advantage of the byte/word extension on the ev56.
> 
> 
> r~
> 


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Slow processing with g77 in egcs
  1998-01-02 16:09 ` Richard Henderson
                     ` (2 preceding siblings ...)
  1998-01-02 16:44   ` pancho
@ 1998-01-05  9:23   ` pancho
  3 siblings, 0 replies; 6+ messages in thread
From: pancho @ 1998-01-05  9:23 UTC (permalink / raw)
  To: Richard Henderson, egcs

Hi all,

I installed egcs-1.0 in a PC164-LX having a cpu 21164a (600 MHz) cache=4
MB ans RAM=512 MB.


I teste the machine with a f77 code that ran in several platforms.
For comparison this code was ran in a axp 255. and a PC164 LX with a cpu
500 MHz, running DU a DEC fortran. Results:

code in PC164LX with cpu 500 MHz was 7 times faster that in the axp255.
code in the PC164LX with cpu 600 MHz running linux and egcs is only 2.5
times faster than axp255.

So my question is: Whyt this difference? shouldn't the code be faster in
the 600 MHz? 

I tryed all the flags tu make the code faster in the PC164LX with cpu 600
MHz.

I would appreciate any suggestions.

regards,

miguel







^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~1998-01-05  9:23 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1998-01-02 12:19 error in alpha ev56 pancho
1998-01-02 16:09 ` Richard Henderson
1998-01-02 16:09   ` pancho
1998-01-02 16:44   ` pancho
1998-01-02 16:44   ` pancho
1998-01-05  9:23   ` Slow processing with g77 in egcs pancho

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).