Account Options

  1. Sign in
The old Google Groups will be going away soon, but your browser is incompatible with the new version.
Google Groups Home
« Groups Home
fresh build of 5.2 from source on PIII, illegal instruction and wrong answers from ecm (factorisation)
There are currently too many topics in this group that display first. To make this topic appear first, remove this option from another topic.
There was an error processing your request. Please try again.
flag
  6 messages - Collapse all  -  Translate all to Translated (View all originals)
The group you are posting to is a Usenet group. Messages posted to this group will make your email address visible to anyone on the Internet.
Your reply message has not been sent.
Your post was successful
 
From:
To:
Cc:
Followup To:
Add Cc | Add Followup-to | Edit Subject
Subject:
Validation:
For verification purposes please type the characters you see in the picture below or the numbers you hear by clicking the accessibility icon. Listen and type the numbers you hear
 
Andrzej Giniewicz  
View profile  
 More options Aug 4 2012, 9:59 am
From: Andrzej Giniewicz <ggi...@gmail.com>
Date: Sat, 4 Aug 2012 15:59:59 +0200
Local: Sat, Aug 4 2012 9:59 am
Subject: fresh build of 5.2 from source on PIII, illegal instruction and wrong answers from ecm (factorisation)
Hi all,

I'm running one of my Sage instances on quite old machine, 2xPentium
III 1GHz. It doesn't even have sse2, but until 5.2 everything worked
and was quite usable when not used by more than 2 users at a time.
Today I tried to update and it compiled fine, but I got following
doctest failures comming from ecm, all seems to be related to
factorisation. This is first time I see Illegal instruction coming out
when building from source, so it seems quite bizarre to me. I'm trying
to rebuild ecm to see what can be the cause. Any ideas how to proceed
with this one?

Thanks,
Andrzej.

sage -t --long -force_lib "devel/sage/sage/libs/libecm.pyx"
**********************************************************************
File "/usr/local/sage/sage-5.2/devel/sage/sage/libs/libecm.pyx", line 81:
    sage: print "ignore this";  ecmfactor(N, 1e2)  # random
Exception raised:
    Traceback (most recent call last):
      File "/usr/local/sage/sage-5.2/local/bin/ncadoctest.py", line
1231, in run_one_test
        self.run_one_example(test, example, filename, compileflags)
      File "/usr/local/sage/sage-5.2/local/bin/sagedoctest.py", line
38, in run_one_example
        OrigDocTestRunner.run_one_example(self, test, example,
filename, compileflags)
      File "/usr/local/sage/sage-5.2/local/bin/ncadoctest.py", line
1172, in run_one_example
        compileflags, 1) in test.globs
      File "<doctest __main__.example_1[6]>", line 1, in <module>
        print "ignore this";  ecmfactor(N, RealNumber('1e2'))  #
random###line 81:
    sage: print "ignore this";  ecmfactor(N, 1e2)  # random
      File "libecm.pyx", line 152, in sage.libs.libecm.ecmfactor
(sage/libs/libecm.c:1986)
    RuntimeError: Illegal instruction
**********************************************************************
File "/usr/local/sage/sage-5.2/devel/sage/sage/libs/libecm.pyx", line 91:
    sage: ecmfactor(N, 1e3)
Exception raised:
    Traceback (most recent call last):
      File "/usr/local/sage/sage-5.2/local/bin/ncadoctest.py", line
1231, in run_one_test
        self.run_one_example(test, example, filename, compileflags)
      File "/usr/local/sage/sage-5.2/local/bin/sagedoctest.py", line
38, in run_one_example
        OrigDocTestRunner.run_one_example(self, test, example,
filename, compileflags)
      File "/usr/local/sage/sage-5.2/local/bin/ncadoctest.py", line
1172, in run_one_example
        compileflags, 1) in test.globs
      File "<doctest __main__.example_1[9]>", line 1, in <module>
        ecmfactor(N, RealNumber('1e3'))###line 91:
    sage: ecmfactor(N, 1e3)
      File "libecm.pyx", line 152, in sage.libs.libecm.ecmfactor
(sage/libs/libecm.c:1986)
    RuntimeError: Illegal instruction
**********************************************************************
File "/usr/local/sage/sage-5.2/devel/sage/sage/libs/libecm.pyx", line 112:
    sage: ecmfactor(N/11, 100, verbose=True)
Exception raised:
    Traceback (most recent call last):
      File "/usr/local/sage/sage-5.2/local/bin/ncadoctest.py", line
1231, in run_one_test
        self.run_one_example(test, example, filename, compileflags)
      File "/usr/local/sage/sage-5.2/local/bin/sagedoctest.py", line
38, in run_one_example
        OrigDocTestRunner.run_one_example(self, test, example,
filename, compileflags)
      File "/usr/local/sage/sage-5.2/local/bin/ncadoctest.py", line
1172, in run_one_example
        compileflags, 1) in test.globs
      File "<doctest __main__.example_1[16]>", line 1, in <module>
        ecmfactor(N/Integer(11), Integer(100), verbose=True)###line 112:
    sage: ecmfactor(N/11, 100, verbose=True)
      File "libecm.pyx", line 152, in sage.libs.libecm.ecmfactor
(sage/libs/libecm.c:1986)
    RuntimeError: Illegal instruction
**********************************************************************
1 items had failures:
   3 of  22 in __main__.example_1
***Test Failed*** 3 failures.

and

sage -t --long -force_lib "devel/sage/sage/interfaces/ecm.py"
**********************************************************************
File "/usr/local/sage/sage-5.2/devel/sage/sage/interfaces/ecm.py", line 173:
    sage: f.one_curve(n, B1=10000, sigma=1022170541)
Expected:
    [79792266297612017, 6366805760909027985741435139224233]
Got:
    [1, 508021860739623467191080372196682785441177798407961]
**********************************************************************
File "/usr/local/sage/sage-5.2/devel/sage/sage/interfaces/ecm.py", line 176:
    sage: f.one_curve(n, B1=500000, algorithm="P-1")
Expected:
    [67872792749091946529, 6366805760909027985741435139224233]
Got:
    [1, 432132887883903108009802143314445113500016816977037257]
**********************************************************************
File "/usr/local/sage/sage-5.2/devel/sage/sage/interfaces/ecm.py", line 179:
    sage: f.one_curve(n, B1=2000, algorithm="P+1", x0=5)
Expected:
    [328006342451, 6366805760909027985741435139224233]
Got:
    [1, 2088352670731726262548647919416588631875815083]
**********************************************************************
File "/usr/local/sage/sage-5.2/devel/sage/sage/interfaces/ecm.py", line 243:
    sage: f.find_factor(n)
Expected:
    [79792266297612017, 6366805760909027985741435139224233]
Got:
    [508021860739623467191080372196682785441177798407961]
**********************************************************************
File "/usr/local/sage/sage-5.2/devel/sage/sage/interfaces/ecm.py", line 333:
    sage: ecm.factor(602400691612422154516282778947806249229526581)
Expected:
    [45949729863572179, 13109994191499930367061460439]
Got:
    [602400691612422154516282778947806249229526581]
**********************************************************************
File "/usr/local/sage/sage-5.2/devel/sage/sage/interfaces/ecm.py", line 336:
    sage: ecm.factor((2^197 + 1)/3)           # takes a long time
Expected:
    [197002597249, 1348959352853811313, 251951573867253012259144010843]
Got:
    [66955751844124594814248420514215108438425124740949701470891]
**********************************************************************
File "/usr/local/sage/sage-5.2/devel/sage/sage/interfaces/ecm.py", line 377:
    sage: ecm.factor((2^197 + 1)/3)             # long time
Expected:
    [197002597249, 1348959352853811313, 251951573867253012259144010843]
Got:
    [66955751844124594814248420514215108438425124740949701470891]
**********************************************************************
4 items had failures:
   3 of  11 in __main__.example_6
   1 of   8 in __main__.example_7
   2 of   5 in __main__.example_8
   1 of   5 in __main__.example_9
***Test Failed*** 7 failures.


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Jeroen Demeyer  
View profile  
 More options Aug 4 2012, 10:01 am
From: Jeroen Demeyer <jdeme...@cage.ugent.be>
Date: Sat, 04 Aug 2012 16:01:38 +0200
Local: Sat, Aug 4 2012 10:01 am
Subject: Re: [sage-devel] fresh build of 5.2 from source on PIII, illegal instruction and wrong answers from ecm (factorisation)
On 2012-08-04 15:59, Andrzej Giniewicz wrote:
> until 5.2 everything worked
> and was quite usable when not used by more than 2 users at a time.

And which is the last version of Sage which worked for sure?

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Andrzej Giniewicz  
View profile  
 More options Aug 4 2012, 10:07 am
From: Andrzej Giniewicz <ggi...@gmail.com>
Date: Sat, 4 Aug 2012 16:07:50 +0200
Local: Sat, Aug 4 2012 10:07 am
Subject: Re: [sage-devel] fresh build of 5.2 from source on PIII, illegal instruction and wrong answers from ecm (factorisation)

On Sat, Aug 4, 2012 at 4:01 PM, Jeroen Demeyer <jdeme...@cage.ugent.be> wrote:
> On 2012-08-04 15:59, Andrzej Giniewicz wrote:
>> until 5.2 everything worked
>> and was quite usable when not used by more than 2 users at a time.
> And which is the last version of Sage which worked for sure?

Currently it runs 5.1 without issues, it is too old to try every
beta/rc (build times are looong), but I try to keep up to date with
stable releases.

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Jeroen Demeyer  
View profile  
 More options Aug 4 2012, 10:10 am
From: Jeroen Demeyer <jdeme...@cage.ugent.be>
Date: Sat, 04 Aug 2012 16:10:19 +0200
Local: Sat, Aug 4 2012 10:10 am
Subject: Re: [sage-devel] fresh build of 5.2 from source on PIII, illegal instruction and wrong answers from ecm (factorisation)
On 2012-08-04 16:07, Andrzej Giniewicz wrote:
> On Sat, Aug 4, 2012 at 4:01 PM, Jeroen Demeyer <jdeme...@cage.ugent.be> wrote:
>> On 2012-08-04 15:59, Andrzej Giniewicz wrote:
>>> until 5.2 everything worked
>>> and was quite usable when not used by more than 2 users at a time.
>> And which is the last version of Sage which worked for sure?

> Currently it runs 5.1 without issues, it is too old to try every
> beta/rc (build times are looong), but I try to keep up to date with
> stable releases.

Could you diff the logfiles of ecm for both versions and see what's the
difference?

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Jeroen Demeyer  
View profile  
 More options Aug 4 2012, 10:11 am
From: Jeroen Demeyer <jdeme...@cage.ugent.be>
Date: Sat, 04 Aug 2012 16:11:41 +0200
Local: Sat, Aug 4 2012 10:11 am
Subject: Re: [sage-devel] fresh build of 5.2 from source on PIII, illegal instruction and wrong answers from ecm (factorisation)
On 2012-08-04 16:07, Andrzej Giniewicz wrote:

> On Sat, Aug 4, 2012 at 4:01 PM, Jeroen Demeyer <jdeme...@cage.ugent.be> wrote:
>> On 2012-08-04 15:59, Andrzej Giniewicz wrote:
>>> until 5.2 everything worked
>>> and was quite usable when not used by more than 2 users at a time.
>> And which is the last version of Sage which worked for sure?

> Currently it runs 5.1 without issues, it is too old to try every
> beta/rc (build times are looong), but I try to keep up to date with
> stable releases.

Also: could you rebuild sgae-5.1 from scratch to make sure the failure
isn't caused by some change on your system?

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Andrzej Giniewicz  
View profile  
 More options Aug 4 2012, 10:52 am
From: Andrzej Giniewicz <ggi...@gmail.com>
Date: Sat, 4 Aug 2012 16:52:17 +0200
Local: Sat, Aug 4 2012 10:52 am
Subject: Re: [sage-devel] fresh build of 5.2 from source on PIII, illegal instruction and wrong answers from ecm (factorisation)
I could have sworn that it worked before, but when I rebuilt ecm on
5.1 it no longer works... strange... diffing logs showed noting
interesting, except for variable ECM_EXTRA_OPTS was renamed to
ECM_CONFIGURE.

Anyway, setting ECM_CONFIGURE=--disable-sse2 manually fixed the issue,
seems that before it was able to detect it, but no longer - maybe
because of gcc update (I'm using system wide gcc and fortran, version
4.7.1)? Or maybe I did not noticed it failing before, if yes then
sorry for noise. Anyway, it's working now with this environment
variable set. Seems it was local issue.

Thanks,
Andrzej.


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
End of messages
« Back to Discussions « Newer topic     Older topic »