Re: Code works with real but does not with double precision. Why??
- From: "Rafael" <rafael.doig@xxxxxxxxx>
- Date: 7 Feb 2006 20:02:49 -0800
yes, it wasn't enough information. Actually, it's difficult to give the
right information in this case.
I copied the dump of the program by using real and then double
precision.
I meant with numerical errors that the time step (dt) becomes rapidly
very small (~D-43).
I also checked all 52 files for consistency and no progress. Thanks
anyway, I will keep checking and checking...
##########CORRECT (USING
REAL)#############################################
runtime: 2.666664E-03 ; dt: 1.333332E-03
runtime: 3.988468E-03 ; dt: 1.321803E-03
runtime: 3.988670E-03 ; dt: 2.027603E-07
runtime: 3.988670E-03 ; dt: 0.000000E+00
runtime: 5.293651E-03 ; dt: 1.304980E-03
runtime: 6.593200E-03 ; dt: 1.299549E-03
runtime: 7.882052E-03 ; dt: 1.288852E-03
runtime: 9.155320E-03 ; dt: 1.273268E-03
runtime: 1.041420E-02 ; dt: 1.258883E-03
runtime: 1.165897E-02 ; dt: 1.244763E-03
runtime: 1.287174E-02 ; dt: 1.212775E-03
runtime: 1.405423E-02 ; dt: 1.182487E-03
runtime: 1.522362E-02 ; dt: 1.169390E-03
runtime: 1.638772E-02 ; dt: 1.164101E-03
....
##########GOES INTO NIRVANA (USING DOUBLE
PRECISION)######################
runtime: 2.66666666666665D-003 ; dt: 1.33333333333333D-003
runtime: 3.98847134828545D-003 ; dt: 1.32180468161880D-003
runtime: 3.98867410894106D-003 ; dt: 2.02760655606754D-007
runtime: 3.98867410894106D-003 ; dt: 3.11396296121217D-027
runtime: 3.98867410894106D-003 ; dt: 1.20896085030152D-043
runtime: 3.98867410894106D-003 ; dt: 1.20895674648333D-043
runtime: 3.98867410894106D-003 ; dt: 1.20892458239154D-043
runtime: 3.98867410894106D-003 ; dt: 1.20889230865314D-043
runtime: 3.98867410894106D-003 ; dt: 1.20886003714134D-043
runtime: 3.98867410894106D-003 ; dt: 1.20882779609914D-043
runtime: 3.98867410894106D-003 ; dt: 1.20879559682467D-043
runtime: 3.98867410894106D-003 ; dt: 1.20876344494483D-043
runtime: 3.98867410894106D-003 ; dt: 1.20873134363264D-043
runtime: 3.98867410894106D-003 ; dt: 1.20869929484410D-043
....
#################################################################
.
- Follow-Ups:
- Re: Code works with real but does not with double precision. Why??
- From: David Jones
- Re: Code works with real but does not with double precision. Why??
- From: glen herrmannsfeldt
- Re: Code works with real but does not with double precision. Why??
- References:
- Code works with real but does not with double precision. Why??
- From: Rafael
- Re: Code works with real but does not with double precision. Why??
- From: Richard Maine
- Code works with real but does not with double precision. Why??
- Prev by Date: Re: Fortran Features
- Next by Date: Re: Fortran Features
- Previous by thread: Re: Code works with real but does not with double precision. Why??
- Next by thread: Re: Code works with real but does not with double precision. Why??
- Index(es):
Relevant Pages
|