thr3ads.net - Speex dev - [Speex-dev] Speex-1.1.7 seems to crash with --enable-sse (on P3/GCC-3.2).. [Apr 2005]

If this information is useful, please help other people find it:
Share via:

jehan Monnier

2005-Apr-20 20:21 UTC

[Speex-dev] Speex-1.1.7 seems to crash with --enable-sse (on P3/GCC-3.2)..

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1"
http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
<b></b>Hi,<br>
<br>
I also tried to build speex 1.1.7&nbsp; with sse enabled on a RAS
3&nbsp; (P4). <br>
Both version 1.1.6 and 1.1.7&nbsp; crash:<br>
<br>
[root@tomb root]# speexenc -n loup.wav loup.spx<br>
Encoding 8000 Hz audio using narrowband mode (stereo)<br>
Segmentation fault (core dumped)<br>
<br>
Stack&nbsp; trace:<br>
Loaded symbols for /lib/ld-linux.so.2<br>
#0&nbsp; fir_mem2_10 (x=0x80552a8, _num=0x80566b4, y=0x80558ac, N=160,
ord=10, _mem=0x80568cc) at xmmintrin.h:790<br>
790&nbsp;&nbsp;&nbsp;&nbsp; {<br>
(gdb) bt<br>
#0&nbsp; fir_mem2_10 (x=0x80552a8, _num=0x80566b4, y=0x80558ac, N=160,
ord=10, _mem=0x80568cc) at xmmintrin.h:790<br>
#1&nbsp; 0xb75dc509 in fir_mem2 (x=0x80552a8, _num=0x80566b4, y=0x80558ac,
N=160, ord=0, _mem=0x1) at filters_sse.h:330<br>
#2&nbsp; 0xb75d2ca7 in nb_encode (state=0x80551b0, vin=0xbfff88e0,
bits=0xbfff9f70) at nb_celp.c:351<br>
#3&nbsp; 0xb75dd560 in speex_encode_int (state=0x80551b0, in=0xb75e9400,
bits=0xbfff9f70) at speex.c:141<br>
#4&nbsp; 0x0804a071 in main (argc=0, argv=0xbfffafe4) at
speexenc.c:688<br>
<br>
<br>
Just a question, how to increase the "encoder/decoder" the stack size
?<br>
Any other suggestion ?<br>
<br>
Regards<br>
<br>
<br>
<br>
<br>
<!--beginarticle-->&nbsp;&nbsp;&nbsp;
&nbsp;&nbsp;&nbsp; Hi,<br>
<blockquote><br>
First question is do you run a 2.4 kernel or later (2.2 doesn't
support<br>
SSE). Next thing to check, does 1.1.6 work? You may also want to<br>
increase the encoder/decoder stack size just in case. Oh, and
don't<br>
compile with -mfpmath=sse .<br>
  <br>
Jean-Marc<br>
  <br>
Le mardi 08 mars 2005 &agrave; 12:02 -0500, Steve Kann a &eacute;crit
:<br>><i> I tried building speex with SSE, and found that it crashes for
me here:</i><br>><i> </i><br>
><i> [This is on RH=8.0]:</i><br>
><i> </i><br>
><i> # libtool gdb ./src/speexenc</i><br>
><i> GNU gdb Red Hat Linux (5.2.1-4)</i><br>
><i> [...]</i><br>
><i> This GDB was configured as
"i386-redhat-linux"...</i><br>
><i> (gdb) r --rate 8000 --16bit -n /dev/zero
/tmp/foo</i><br>
><i> Starting program: /usr/src/speex-1.1.7/src/.libs/lt-speexenc
--rate 8000 </i><br>><i> --16bit -n /dev/zero /tmp/foo</i><br>
><i> Encoding 8000 Hz audio using narrowband mode
(mono)</i><br>
><i> </i><br>
><i> Program received signal SIGSEGV, Segmentation
fault.</i><br>
><i> fir_mem2_10 (x=0x8054a88, _num=0x8055e94, y=0x805508c, N=160,
ord=10, </i><br>><i> _mem=0x80560ac)</i><br>
><i> at
/usr/lib/gcc-lib/i386-redhat-linux/3.2/include/xmmintrin.h:774</i><br>><i> 774 {</i><br>
><i> (gdb)</i><br>
><i> (gdb) bt</i><br>
><i> #0 fir_mem2_10 (x=0x8054a88, _num=0x8055e94, y=0x805508c,
N=160, </i><br>><i> ord=10, _mem=0x80560ac)</i><br>
><i> at
/usr/lib/gcc-lib/i386-redhat-linux/3.2/include/xmmintrin.h:774</i><br>><i> #1 0x40020add in fir_mem2 (x=0x8054a88, _num=0x8055e94,
y=0x805508c, </i><br>><i> N=160, ord=0, _mem=0x1) at filters_sse.h:330</i><br>
><i> #2 0x40016c23 in nb_encode (state=0x8054990, vin=0xbfffd420,
</i><br>
><i> bits=0xbfffeaa0) at nb_celp.c:351</i><br>
><i> #3 0x400215f4 in speex_encode_int (state=0x8054990,
in=0xbfffead0, </i><br>><i> bits=0xbfffeaa0) at speex.c:141</i><br>
><i> #4 0x0804a089 in main (argc=-80, argv=0xbffffad4) at
speexenc.c:688</i><br>><i> #5 0x420158d4 in __libc_start_main () from
/lib/i686/libc.so.6</i><br>
><i> (gdb) q</i><br>
><i> The program is running. Exit anyway? (y or n)
y</i><br>
><i> </i><br>
><i> </i><br>
><i> If I build without sse, things are OK.</i><br>
><i> </i><br>
><i> This box is running RH-8 on a PIII (katmai), which seems to
support sse:</i><br>><i> flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca
</i><br>
><i> cmov pat pse36 mmx fxsr sse</i><br>
><i> </i><br>
><i> If I compile the same thing on another box with an Athlon CPU
instead, </i><br>><i> it seems to work without crashing..</i><br>
><i> </i><br>
><i> I guess it's not critical in this case, but does anyone know
if
this is </i><br>><i> an issue with the SSE intrinsics in gcc-3.2, or is it a
CPU-support </i><br>><i> issue (or something else?).</i><br>
><i> </i><br>
><i> -SteveK</i><br>
><i> </i><br>
><i> </i><br>
><i>
_______________________________________________</i><br>
><i> Speex-dev mailing list</i><br>
><i> <a
href="http://lists.xiph.org/mailman/listinfo/speex-dev">Speex-devat xiph.org</a></i><br>
  <i></i>><i> <a
 href="http://lists.xiph.org/mailman/listinfo/speex-dev">http://lists.xiph.org/mailman/listinfo/speex-dev</a></i><br>
  <i></i>><i> </i><br>
-- <br>
Jean-Marc Valin <<a
 href="http://lists.xiph.org/mailman/listinfo/speex-dev">Jean-Marc.Valin
at USherbrooke.ca</a>><br>
Universit&eacute; de Sherbrooke<br>
</blockquote>
<pre>
</pre>
</body>
</html>

Jean-Marc Valin

2005-Apr-20 20:27 UTC

head link

[Speex-dev] Speex-1.1.7 seems to crash with --enable-sse (on P3/GCC-3.2)..

Hi,

To increase the stack size, you can increase the value in the
speex_alloc() call in nb_celp.c. Let me know if it solves the problem.
Also, what compiler are you using. I think some older versions of gcc
may have problems with SSE intrinsics.

	Jean-Marc

Le mardi 19 avril 2005 ? 16:02 +0200, jehan Monnier a ?crit
:> Hi,
> 
> I also tried to build speex 1.1.7  with sse enabled on a RAS 3  (P4). 
> Both version 1.1.6 and 1.1.7  crash:
> 
> [root@tomb root]# speexenc -n loup.wav loup.spx
> Encoding 8000 Hz audio using narrowband mode (stereo)
> Segmentation fault (core dumped)
> 
> Stack  trace:
> Loaded symbols for /lib/ld-linux.so.2
> #0  fir_mem2_10 (x=0x80552a8, _num=0x80566b4, y=0x80558ac, N=160,
> ord=10, _mem=0x80568cc) at xmmintrin.h:790
> 790     {
> (gdb) bt
> #0  fir_mem2_10 (x=0x80552a8, _num=0x80566b4, y=0x80558ac, N=160,
> ord=10, _mem=0x80568cc) at xmmintrin.h:790
> #1  0xb75dc509 in fir_mem2 (x=0x80552a8, _num=0x80566b4, y=0x80558ac,
> N=160, ord=0, _mem=0x1) at filters_sse.h:330
> #2  0xb75d2ca7 in nb_encode (state=0x80551b0, vin=0xbfff88e0,
> bits=0xbfff9f70) at nb_celp.c:351
> #3  0xb75dd560 in speex_encode_int (state=0x80551b0, in=0xb75e9400,
> bits=0xbfff9f70) at speex.c:141
> #4  0x0804a071 in main (argc=0, argv=0xbfffafe4) at speexenc.c:688
> 
> 
> Just a question, how to increase the "encoder/decoder" the stack
> size ?
> Any other suggestion ?
> 
> Regards
> 
> 
> 
> 
>         Hi,
>         
>         First question is do you run a 2.4 kernel or later (2.2
>         doesn't support
>         SSE). Next thing to check, does 1.1.6 work? You may also want
>         to
>         increase the encoder/decoder stack size just in case. Oh, and
>         don't
>         compile with -mfpmath=sse .
>         
>         Jean-Marc
>         
>         Le mardi 08 mars 2005 ? 12:02 -0500, Steve Kann a ?crit :
>         > I tried building speex with SSE, and found that it crashes
>         for me here:
>         > 
>         > [This is on RH=8.0]:
>         > 
>         > # libtool gdb ./src/speexenc
>         > GNU gdb Red Hat Linux (5.2.1-4)
>         > [...]
>         > This GDB was configured as "i386-redhat-linux"...
>         > (gdb) r --rate 8000 --16bit -n /dev/zero /tmp/foo
>         > Starting program: /usr/src/speex-1.1.7/src/.libs/lt-speexenc
>         --rate 8000 
>         > --16bit -n /dev/zero /tmp/foo
>         > Encoding 8000 Hz audio using narrowband mode (mono)
>         > 
>         > Program received signal SIGSEGV, Segmentation fault.
>         > fir_mem2_10 (x=0x8054a88, _num=0x8055e94, y=0x805508c,
>         N=160, ord=10, 
>         > _mem=0x80560ac)
>         >
>         at /usr/lib/gcc-lib/i386-redhat-linux/3.2/include/xmmintrin.h:774
>         > 774 {
>         > (gdb)
>         > (gdb) bt
>         > #0 fir_mem2_10 (x=0x8054a88, _num=0x8055e94, y=0x805508c,
>         N=160, 
>         > ord=10, _mem=0x80560ac)
>         >
>         at /usr/lib/gcc-lib/i386-redhat-linux/3.2/include/xmmintrin.h:774
>         > #1 0x40020add in fir_mem2 (x=0x8054a88, _num=0x8055e94,
>         y=0x805508c, 
>         > N=160, ord=0, _mem=0x1) at filters_sse.h:330
>         > #2 0x40016c23 in nb_encode (state=0x8054990,
>         vin=0xbfffd420, 
>         > bits=0xbfffeaa0) at nb_celp.c:351
>         > #3 0x400215f4 in speex_encode_int (state=0x8054990,
>         in=0xbfffead0, 
>         > bits=0xbfffeaa0) at speex.c:141
>         > #4 0x0804a089 in main (argc=-80, argv=0xbffffad4) at
>         speexenc.c:688
>         > #5 0x420158d4 in __libc_start_main ()
>         from /lib/i686/libc.so.6
>         > (gdb) q
>         > The program is running. Exit anyway? (y or n) y
>         > 
>         > 
>         > If I build without sse, things are OK.
>         > 
>         > This box is running RH-8 on a PIII (katmai), which seems to
>         support sse:
>         > flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca 
>         > cmov pat pse36 mmx fxsr sse
>         > 
>         > If I compile the same thing on another box with an Athlon
>         CPU instead, 
>         > it seems to work without crashing..
>         > 
>         > I guess it's not critical in this case, but does anyone
know
>         if this is 
>         > an issue with the SSE intrinsics in gcc-3.2, or is it a
>         CPU-support 
>         > issue (or something else?).
>         > 
>         > -SteveK
>         > 
>         > 
>         > _______________________________________________
>         > Speex-dev mailing list
>         > Speex-dev at xiph.org
>         > http://lists.xiph.org/mailman/listinfo/speex-dev
>         > 
>         -- 
>         Jean-Marc Valin <Jean-Marc.Valin at USherbrooke.ca>
>         Universit? de Sherbrooke
> _______________________________________________
> Speex-dev mailing list
> Speex-dev@xiph.org
> http://lists.xiph.org/mailman/listinfo/speex-dev-- 
Jean-Marc Valin <Jean-Marc.Valin@USherbrooke.ca>
Universit? de Sherbrooke

jehan Monnier

2005-Apr-22 04:29 UTC

head link

[Speex-dev] Speex-1.1.7 seems to crash with --enable-sse (on P3/GCC-3.2)..

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1"
http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Hi Jean-Marc,<br>
<br>
I have increased the stack size from 8000 to 16000:<br>
<blockquote>&nbsp;&nbsp; /*st
(EncState*)speex_alloc(sizeof(EncState)+8000*sizeof(spx_sig_t));*/<br>
&nbsp;&nbsp; st
(EncState*)speex_alloc(sizeof(EncState)+16000*sizeof(spx_sig_t));<br>
</blockquote>
But the result is the same (core dump) :-(<br>
<br>
The compiler version I used is the following:<br>
<br>
<blockquote>[root@tomb speex-1.1.7]# gcc -v<br>
Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/3.2.3/specs<br>
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
--infodir=/usr/share/info --enable-shared --enable-threads=posix
--disable-checking --with-system-zlib --enable-__cxa_atexit
--host=i386-redhat-linux<br>
Thread model: posix<br>
gcc version 3.2.3 20030502 (Red Hat Linux 3.2.3-34)<br>
</blockquote>
<br>
<br>
Compiled with gcc 3.4.3, speex 1.1.7 works fine&nbsp; (both with and without
SSE)<br>
So this issue seems to be in gcc 3.2.3<br>
<br>
Best Regards,<br>
<br>
Jehan<br>
<br>
Jean-Marc Valin wrote:
<blockquote cite="mid1114053926.7347.1.camel@localhost"
type="cite">
  <pre wrap="">Hi,

To increase the stack size, you can increase the value in the
speex_alloc() call in nb_celp.c. Let me know if it solves the problem.
Also, what compiler are you using. I think some older versions of gcc
may have problems with SSE intrinsics.

&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbsp&nbspJean-Marc

Le mardi 19 avril 2005 &agrave; 16:02 +0200, jehan Monnier a
&eacute;crit :
  </pre>
  <blockquote type="cite">
    <pre wrap="">Hi,

I also tried to build speex 1.1.7  with sse enabled on a RAS 3  (P4). 
Both version 1.1.6 and 1.1.7  crash:

[root@tomb root]# speexenc -n loup.wav loup.spx
Encoding 8000 Hz audio using narrowband mode (stereo)
Segmentation fault (core dumped)

Stack  trace:
Loaded symbols for /lib/ld-linux.so.2
#0  fir_mem2_10 (x=0x80552a8, _num=0x80566b4, y=0x80558ac, N=160,
ord=10, _mem=0x80568cc) at xmmintrin.h:790
790     {
(gdb) bt
#0  fir_mem2_10 (x=0x80552a8, _num=0x80566b4, y=0x80558ac, N=160,
ord=10, _mem=0x80568cc) at xmmintrin.h:790
#1  0xb75dc509 in fir_mem2 (x=0x80552a8, _num=0x80566b4, y=0x80558ac,
N=160, ord=0, _mem=0x1) at filters_sse.h:330
#2  0xb75d2ca7 in nb_encode (state=0x80551b0, vin=0xbfff88e0,
bits=0xbfff9f70) at nb_celp.c:351
#3  0xb75dd560 in speex_encode_int (state=0x80551b0, in=0xb75e9400,
bits=0xbfff9f70) at speex.c:141
#4  0x0804a071 in main (argc=0, argv=0xbfffafe4) at speexenc.c:688


Just a question, how to increase the "encoder/decoder" the stack
size ?
Any other suggestion ?

Regards




        Hi,
        
        First question is do you run a 2.4 kernel or later (2.2
        doesn't support
        SSE). Next thing to check, does 1.1.6 work? You may also want
        to
        increase the encoder/decoder stack size just in case. Oh, and
        don't
        compile with -mfpmath=sse .
        
        Jean-Marc
        
        Le mardi 08 mars 2005 &agrave; 12:02 -0500, Steve Kann a
&eacute;crit :
        > I tried building speex with SSE, and found that it crashes
        for me here:
        > 
        > [This is on RH=8.0]:
        > 
        > # libtool gdb ./src/speexenc
        > GNU gdb Red Hat Linux (5.2.1-4)
        > [...]
        > This GDB was configured as "i386-redhat-linux"...
        > (gdb) r --rate 8000 --16bit -n /dev/zero /tmp/foo
        > Starting program: /usr/src/speex-1.1.7/src/.libs/lt-speexenc
        --rate 8000 
        > --16bit -n /dev/zero /tmp/foo
        > Encoding 8000 Hz audio using narrowband mode (mono)
        > 
        > Program received signal SIGSEGV, Segmentation fault.
        > fir_mem2_10 (x=0x8054a88, _num=0x8055e94, y=0x805508c,
        N=160, ord=10, 
        > _mem=0x80560ac)
        >
        at /usr/lib/gcc-lib/i386-redhat-linux/3.2/include/xmmintrin.h:774
        > 774 {
        > (gdb)
        > (gdb) bt
        > #0 fir_mem2_10 (x=0x8054a88, _num=0x8055e94, y=0x805508c,
        N=160, 
        > ord=10, _mem=0x80560ac)
        >
        at /usr/lib/gcc-lib/i386-redhat-linux/3.2/include/xmmintrin.h:774
        > #1 0x40020add in fir_mem2 (x=0x8054a88, _num=0x8055e94,
        y=0x805508c, 
        > N=160, ord=0, _mem=0x1) at filters_sse.h:330
        > #2 0x40016c23 in nb_encode (state=0x8054990,
        vin=0xbfffd420, 
        > bits=0xbfffeaa0) at nb_celp.c:351
        > #3 0x400215f4 in speex_encode_int (state=0x8054990,
        in=0xbfffead0, 
        > bits=0xbfffeaa0) at speex.c:141
        > #4 0x0804a089 in main (argc=-80, argv=0xbffffad4) at
        speexenc.c:688
        > #5 0x420158d4 in __libc_start_main ()
        from /lib/i686/libc.so.6
        > (gdb) q
        > The program is running. Exit anyway? (y or n) y
        > 
        > 
        > If I build without sse, things are OK.
        > 
        > This box is running RH-8 on a PIII (katmai), which seems to
        support sse:
        > flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca 
        > cmov pat pse36 mmx fxsr sse
        > 
        > If I compile the same thing on another box with an Athlon
        CPU instead, 
        > it seems to work without crashing..
        > 
        > I guess it's not critical in this case, but does anyone know
        if this is 
        > an issue with the SSE intrinsics in gcc-3.2, or is it a
        CPU-support 
        > issue (or something else?).
        > 
        > -SteveK
        > 
        > 
        > _______________________________________________
        > Speex-dev mailing list
        > Speex-dev at xiph.org
        > <a class="moz-txt-link-freetext"
href="http://lists.xiph.org/mailman/listinfo/speex-dev">http://lists.xiph.org/mailman/listinfo/speex-dev</a>
        > 
        -- 
        Jean-Marc Valin <Jean-Marc.Valin at USherbrooke.ca>
        Universit&eacute; de Sherbrooke
_______________________________________________
Speex-dev mailing list
<a class="moz-txt-link-abbreviated"
href="mailto:Speex-dev@xiph.org">Speex-dev@xiph.org</a>
<a class="moz-txt-link-freetext"
href="http://lists.xiph.org/mailman/listinfo/speex-dev">http://lists.xiph.org/mailman/listinfo/speex-dev</a>
    </pre>
  </blockquote>
</blockquote>
<br>
</body>
</html>

Apparently Analagous Threads

Search for more reasonably related threads

Speex dev - Apr 2005 - Speex-1.1.7 seems to crash with --enable-sse (on P3/GCC-3.2)..

[Speex-dev] Speex-1.1.7 seems to crash with --enable-sse (on P3/GCC-3.2)..

[Speex-dev] Speex-1.1.7 seems to crash with --enable-sse (on P3/GCC-3.2)..

[Speex-dev] Speex-1.1.7 seems to crash with --enable-sse (on P3/GCC-3.2)..

Apparently Analagous Threads