Opened 12 years ago

Closed 9 years ago

Last modified 7 years ago

#9899 closed defect (invalid)

PARI/GP (2.4.3-svn) self-test fails after self-tuning on Pentium 4 Prescott

Reported by: leif Owned by: leif
Priority: minor Milestone: sage-duplicate/invalid/wontfix
Component: packages: standard Keywords: mathilbert SAGE_TUNE SAGE_CHECK
Cc: jdemeyer Merged in:
Authors: Reviewers: Jeroen Demeyer
Report Upstream: Not yet reported upstream; Will do shortly. Work issues:
Branch: Commit:
Dependencies: Stopgaps:

Status badges

Description (last modified by chapoton)

On Ubuntu 9.04 x86 (Pentium 4 Prescott, gcc 4.3.3) I get the following when (re)installing PARI 2.4.3.svn-12577.p5 with SAGE_TUNE_PARI=yes and SAGE_CHECK=yes: 

...
==========================================================================
Building and tuning PARI (this may take a while)
...
Bye !
Building and installing PARI/GP...
Making gp in Olinux-i686
...
real    20m58.068s
user    20m28.501s
sys    0m20.849s
Successfully installed pari-2.4.3.svn-12577.p5
Running the test suite.
Making test-all in Olinux-i686
...
* Testing analyz     for gp-sta..TIME=28    for gp-dyn..TIME=36
* Testing apply     for gp-sta..TIME=0    for gp-dyn..TIME=8
* Testing aurifeuille     for gp-sta..TIME=4    for gp-dyn..TIME=8
* Testing bezout     for gp-sta..TIME=4    for gp-dyn..TIME=8
* Testing bnfisintnorm     for gp-sta..TIME=1124    for gp-dyn..TIME=1164
* Testing bnr     for gp-sta..TIME=36    for gp-dyn..TIME=44
* Testing charpoly     for gp-sta..TIME=4    for gp-dyn..TIME=8
* Testing combinat     for gp-sta..TIME=36    for gp-dyn..TIME=12
* Testing compat     for gp-sta..TIME=448    for gp-dyn..TIME=432
* Testing contfrac     for gp-sta..TIME=4    for gp-dyn..TIME=8
* Testing debugger     for gp-sta..TIME=8    for gp-dyn..TIME=4
* Testing ell     for gp-sta..TIME=18781    for gp-dyn..TIME=18869
* Testing elliptic     for gp-sta..TIME=44    for gp-dyn..TIME=36
* Testing ellsea     for gp-sta..TIME=27077    for gp-dyn..TIME=27025
* Testing ellweilpairing     for gp-sta..TIME=128    for gp-dyn..TIME=108
* Testing err     for gp-sta..TIME=4    for gp-dyn..TIME=4
* Testing exact0     for gp-sta..TIME=4    for gp-dyn..TIME=4
* Testing extract     for gp-sta..TIME=8    for gp-dyn..TIME=8
* Testing ff     for gp-sta..TIME=884    for gp-dyn..TIME=820
* Testing ffisom     for gp-sta..TIME=700    for gp-dyn..TIME=684
* Testing galois     for gp-sta..TIME=25029    for gp-dyn..TIME=25569
* Testing galoisinit     for gp-sta..TIME=7304    for gp-dyn..TIME=7276
* Testing graph     for gp-sta..TIME=24    for gp-dyn..TIME=32
* Testing ideal     for gp-sta..TIME=8    for gp-dyn..TIME=4
* Testing idealappr     for gp-sta..TIME=4    for gp-dyn..TIME=8
* Testing idealramgroups     for gp-sta..TIME=3576    for gp-dyn..TIME=3580
* Testing intformal     for gp-sta..TIME=4    for gp-dyn..TIME=8
* Testing intnum     for gp-sta..TIME=30797    for gp-dyn..TIME=31073
* Testing ispower     for gp-sta..TIME=7604    for gp-dyn..TIME=7632
* Testing krasner     for gp-sta..TIME=6996    for gp-dyn..TIME=7000
* Testing linear     for gp-sta..BUG [20]    for gp-dyn..BUG [36]
* Testing list     for gp-sta..TIME=96    for gp-dyn..TIME=60
* Testing lll     for gp-sta..TIME=0    for gp-dyn..TIME=4
* Testing mat     for gp-sta..TIME=4    for gp-dyn..TIME=8
* Testing matsnf     for gp-sta..TIME=816    for gp-dyn..TIME=760
* Testing member     for gp-sta..TIME=136    for gp-dyn..TIME=108
* Testing modpr     for gp-sta..TIME=0    for gp-dyn..TIME=4
* Testing multivar-mul     for gp-sta..TIME=17417    for gp-dyn..TIME=19153
* Testing nf     for gp-sta..TIME=4128    for gp-dyn..TIME=4072
* Testing nffactor     for gp-sta..TIME=20721    for gp-dyn..TIME=20745
* Testing nfhilbert     for gp-sta..TIME=8    for gp-dyn..TIME=12
* Testing nfields     for gp-sta..TIME=248    for gp-dyn..TIME=216
* Testing nfrootsof1     for gp-sta..TIME=45166    for gp-dyn..TIME=45766
* Testing number     for gp-sta..TIME=76    for gp-dyn..TIME=48
* Testing objets     for gp-sta..TIME=4    for gp-dyn..TIME=4
* Testing partition     for gp-sta..TIME=60831    for gp-dyn..TIME=62931
* Testing polchebyshev     for gp-sta..TIME=16    for gp-dyn..TIME=20
* Testing polmod     for gp-sta..TIME=4    for gp-dyn..TIME=8
* Testing polred     for gp-sta..TIME=6104    for gp-dyn..TIME=6136
* Testing polyser     for gp-sta..TIME=16    for gp-dyn..TIME=28
* Testing printf     for gp-sta..TIME=4    for gp-dyn..TIME=16
* Testing program     for gp-sta..TIME=44    for gp-dyn..TIME=28
* Testing qf     for gp-sta..TIME=4    for gp-dyn..TIME=4
* Testing qfbsolve     for gp-sta..TIME=3576    for gp-dyn..TIME=3756
* Testing quad     for gp-sta..TIME=0    for gp-dyn..TIME=8
* Testing quadclassunit     for gp-sta..TIME=17357    for gp-dyn..TIME=18169
* Testing quadray     for gp-sta..TIME=1232    for gp-dyn..TIME=1256
* Testing random     for gp-sta..TIME=12    for gp-dyn..TIME=12
* Testing resultant     for gp-sta..TIME=22545    for gp-dyn..TIME=23341
* Testing rfrac     for gp-sta..TIME=6680    for gp-dyn..TIME=7128
* Testing rnf     for gp-sta..TIME=564    for gp-dyn..TIME=548
* Testing rnfkummer     for gp-sta..TIME=87353    for gp-dyn..TIME=87361
* Testing round4     for gp-sta..TIME=13872    for gp-dyn..TIME=13904
* Testing select     for gp-sta..TIME=8    for gp-dyn..TIME=4
* Testing stark     for gp-sta..TIME=47794    for gp-dyn..TIME=47862
* Testing subcyclo     for gp-sta..TIME=4    for gp-dyn..TIME=8
* Testing subfields     for gp-sta..TIME=24165    for gp-dyn..TIME=24129
* Testing sumiter     for gp-sta..TIME=52    for gp-dyn..TIME=44
* Testing thue     for gp-sta..TIME=3756    for gp-dyn..TIME=3932
* Testing trans     for gp-sta..TIME=112    for gp-dyn..TIME=84
* Testing zetak     for gp-sta..TIME=5560    for gp-dyn..TIME=5520
* Testing zn     for gp-sta..TIME=8    for gp-dyn..TIME=8
+++ [BUG] Total bench for gp-sta is 521185
+++ [BUG] Total bench for gp-dyn is 528713

PROBLEMS WERE NOTED. The following files list them in diff format: 
Directory: /home/leif/Sage/sage-4.6.prealpha4/spkg/build/pari-2.4.3.svn-12577.p5/src/Olinux-i686
    linear-sta.dif
    linear-dyn.dif
make[1]: *** [test-all] Error 1
make[1]: Leaving directory `/home/leif/Sage/sage-4.6.prealpha4/spkg/build/pari-2.4.3.svn-12577.p5/src/Olinux-i686'
make: *** [test-all] Error 2
Error: PARI failed the self-tests when running 'make -j8 test-all'
*************************************
Error testing package ** pari-2.4.3.svn-12577.p5 **
*************************************
sage: An error occurred while testing pari-2.4.3.svn-12577.p5
...

The attached diffs generated by PARI look weird since PARI breaks the output into fixed width lines (and compares these) - even within numbers that would fit on a single line.

I've reformatted both the expected and the actual output s.t. the differences are more readable; now more obviously this is just numerical noise:

  • pari-test-linear.

    old new  
    11? (1.*mathilbert(7))^(-1)
    22
    33[       49.000000000000000000000000000001579425
    4      -1176.0000000000000000000000000000621879
     4     -1176.0000000000000000000000000000621902
    55      8820.0000000000000000000000000005948011
    66    -29400.000000000000000000000000002301748
    77     48510.000000000000000000000000004207809
    … …  
    99     12012.000000000000000000000000001189528]
    1010
    1111[    -1176.0000000000000000000000000000626175
    12      37632.000000000000000000000000002469695
     12     37632.000000000000000000000000002469681
    1313   -317520.00000000000000000000000002364675
    1414   1128960.0000000000000000000000000915946
    1515  -1940400.0000000000000000000000001675308
    … …  
    1717   -504504.00000000000000000000000004740575]
    1818
    1919[     8820.0000000000000000000000000006015186
    20    -317520.00000000000000000000000002375151
     20   -317520.00000000000000000000000002375139
    2121   2857680.0000000000000000000000002275919
    2222 -10584000.000000000000000000000000882085
    2323  18711000.000000000000000000000001614074
    … …  
    2525   5045040.0000000000000000000000004570220]
    2626
    2727[   -29400.000000000000000000000000002335330
    28    1128960.0000000000000000000000000922894
     28   1128960.0000000000000000000000000922890
    2929 -10584000.000000000000000000000000884843
    3030  40320000.000000000000000000000003430862
    3131 -72765000.000000000000000000000006279937
    … …  
    3333 -20180160.000000000000000000000001778972]
    3434
    3535[    48510.000000000000000000000000004278611
    36   -1940400.0000000000000000000000001691936
     36  -1940400.0000000000000000000000001691929
    3737  18711000.000000000000000000000001622888
    3838 -72765000.000000000000000000000006294553
    3939 133402500.00000000000000000000001152452
    … …  
    4141  37837800.000000000000000000000003265792]
    4242
    4343[   -38808.000000000000000000000000003696181
    44    1596672.0000000000000000000000001462360
     44   1596672.0000000000000000000000001462353
    4545 -15717240.000000000000000000000001403167
    4646  62092800.000000000000000000000005443713
    4747-115259760.00000000000000000000000996866
    … …  
    4949 -33297264.000000000000000000000002825670]
    5050
    5151[    12012.000000000000000000000000001213544
    52    -504504.00000000000000000000000004803250
     52   -504504.00000000000000000000000004803228
    5353   5045040.0000000000000000000000004610120
    5454 -20180160.000000000000000000000001788903
    5555  37837800.000000000000000000000003276393

At least in Sage 4.6.alpha0 the slight deviation doesn't affect ptestlong, i.e. all tests passed regardless of PARI's self-tuning being enabled or not.

Nevertheless it's odd PARI's test suite fails with tuning enabled, since

  • Sage will reject to install the package when SAGE_CHECK=yes,
  • analyzing the failures is quite tedious, and
  • disabling the self-tests is potentially dangerous.

On the same machine running Fedora 13 x86 (gcc 4.4.4) PARI's self-tuning hangs when tuning REMIIMUL_LIMIT, i.e. the output stops, but tune doesn't terminate. This should be reported upstream as well.

Attachments (2)

linear-dyn.dif​ (4.5 KB) - added by leif 12 years ago.
linear-sta.dif​ (4.5 KB) - added by leif 12 years ago.

Download all attachments as: .zip

Change History (7)

Changed 12 years ago by leif

Attachment: linear-dyn.dif​ added

Changed 12 years ago by leif

Attachment: linear-sta.dif​ added

comment:1 Changed 9 years ago by jdemeyer

Milestone: sage-5.11 → sage-5.12

comment:2 Changed 9 years ago by jdemeyer

Milestone: sage-5.13 → sage-duplicate/invalid/wontfix
Reviewers: → Jeroen Demeyer
Status: new → needs_review

This refers to an old version of PARI...

comment:3 Changed 9 years ago by jdemeyer

Status: needs_review → positive_review

comment:4 Changed 9 years ago by jdemeyer

Resolution: → invalid
Status: positive_review → closed

comment:5 Changed 7 years ago by chapoton

Description: modified (diff)
Note: See TracTickets for help on using tickets.