View Ticket
Not logged in
Ticket UUID: 3364b2f5ebc70589779147a05af7eec2699b725c
Title: Provide double precision
Status: Fixed Type: Feature_Request
Severity: Critical Priority: Immediate
Subsystem: Resolution: Open
Last Modified: 2013-04-04 22:15:30
Version Found In:
A user should be able to compute in double precision in both cpu and/or gpu (if that hardware capability is available). Thus the library should have double and single versions of the routines simultaneously.

A user may choose at installation time to build single only, double only, or both (the default).

gpapamak added on 2013-04-04 22:15:30:
The above was fixed in version 1.5