Re: FAIL in make check /specfunc on FreeBSD4.3/i386

Re: FAIL in make check /specfunc on FreeBSD4.3/i386

[Brian Gough]

Vladimir Savichev writes:
 > after 891 we have
 > sn = 0.96402778575700188   wanted 0.96402778575700186570 ~ 1.4 10^-17 < 
 > 2*GSL_DBL_EPSILON
 > cn = 0.26580148285600663          0.26580148285600686381 ~ -2.3 10^-16 < 
 > 2*GSL_DBL_EPSILON
 > dn = 0.26580323105264092          0.26580323105264131136 ~ -3.9 10^-16 < 
 > 2*GSL_DBL_EPSILON
 > 
 > but in test_sf_frac_diff we return
 > fabs((x1-x2)/(x1+x2)) so we get about 8 10^-16 at dn : HAS to FAIL !
 > May be it is AMD issue or something  particular about FreeBSD?

Thanks for reporting that... the tolerance in the test seems to be too
small to allow for the small variation between different
platforms/compilers/etc.  We'll increase it before the next release.

 > For example on entry in gsl_sf_elljac_e(u, m, &sn, &cn, &dn)
 > I see like
 > m=0.99999899999999997
 > Shouldn't we see   m=0.99999900000  for declared double m=0.999999 ?

It's ok, there is a conversion from decimal to binary so the result
does not always look exact in decimal.

FAIL in make check /specfunc on FreeBSD4.3/i386

[Vladimir Savichev]

Hi,
I'm here on FreeBSD 4.3-STABLE i386 (AMD  Athlon) gcc 2.95.3
trying to port gls-0.8 to FreeBSD ports collection.
make check in specfunc dep gives 

...
PASS:   gsl_sf_elljac_e(3.0|0.6)
FAIL:   gsl_sf_elljac_e(2.0|0.999999)
PASS:   gsl_sf_elljac_e(1.69695970624443|0.270378013104138)
FAIL: Elliptic Functions (Jacobi)
PASS: gsl_sf_erfc_e(-10.0, &r)
...

here we go to dip into test_sf.c (gdb test)
883   u = 2.0;
884   m = 0.999999;
885   sa = 0;
886   stat_ej = gsl_sf_elljac_e(u, m, &sn, &cn, &dn);
887   sa += ( test_sf_frac_diff( sn, 0.96402778575700186570 ) > TEST_TOL0 );
888   sa += ( test_sf_frac_diff( cn, 0.26580148285600686381 ) > TEST_TOL0 );
889   sa += ( test_sf_frac_diff( dn, 0.26580323105264131136 ) > TEST_TOL0 );
890   gsl_test(sa, "  gsl_sf_elljac_e(2.0|0.999999)");
891   s += sa;
after 891 we have
sn = 0.96402778575700188   wanted 0.96402778575700186570 ~ 1.4 10^-17 < 
2*GSL_DBL_EPSILON
cn = 0.26580148285600663          0.26580148285600686381 ~ -2.3 10^-16 < 
2*GSL_DBL_EPSILON
dn = 0.26580323105264092          0.26580323105264131136 ~ -3.9 10^-16 < 
2*GSL_DBL_EPSILON

but in test_sf_frac_diff we return
fabs((x1-x2)/(x1+x2)) so we get about 8 10^-16 at dn : HAS to FAIL !
May be it is AMD issue or something  particular about FreeBSD?
For example on entry in gsl_sf_elljac_e(u, m, &sn, &cn, &dn)
I see like
m=0.99999899999999997
Shouldn't we see   m=0.99999900000  for declared double m=0.999999 ?
I could just make a guess what else is a matter.

Thanks for your business
Vladimir.

Re: FAIL in make check /specfunc on FreeBSD4.3/i386

[Vladimir Savichev]

didn't get the answer yet and fill   damn to reply on my own message,
but have some remark to be more correct. This FAIL is the ONLY ONE
on the hole test and looks nearly patological, i.e. doesn't look
like real mistake. I was just wondering I saw none like this on
the list, isn't ?

Thanks Vlad.