The embedded version of BLAS and LAPACK is delayed as the code should use a GPU if it is available on your hardware (quite common these days). So, for instance, SAXPY (a*x[i]+y[i]) should use CUDA if available or default to the CPU if not (Torch will do this, for example). I just recently got a Google TPU which will be even faster though much less common. The fun never ends.
On Saturday, December 17, 2022 at 12:00:14 PM UTC-5 [email protected] wrote: > Am astothised . Will respond later > > Le ven. 16 déc. 2022 à 19:20, Tim Daly <[email protected]> a écrit : > >> My email is axiomcas at gmail.com >> >> Domains handling Fortran from NAG generally are found using >> )apropos Asp >> >> Conversions to Fortran work, e.g >> >> outputAsFortran(3*x) ==> R8=3*x >> >> but the code that invokes NAG routines such as f02aff from NagEigenPackage >> compile to a Lisp call 'invokeNagman' which wanders down to a lisp >> 'nagCall'. >> That invokes 'sockSendString' to talk to a NAG co-process. >> >> I don't have a NAG license so I don't have access to NAG libraries to >> test this. >> It would be possible to write such a process and have it attach to a >> socket >> that handled the messages. >> >> Some of the functionality you need may be in BLAS and LAPACK. >> >> https://en.wikipedia.org/wiki/Basic_Linear_Algebra_Subprograms >> >> https://en.wikipedia.org/wiki/LAPACK >> >> Axiom has converted the BLAS and LAPACK source code to Common Lisp. >> Portions of this have been added to Axiom directly (more to come) so there >> is no need for a process/socket. >> >> For example, from BlasLevelOne: >> >> (9) -> t2:Complex DoubleFloat := complex(1.0,1.0) >> >> (9) 1. + %i >> Type: >> Complex(DoubleFloat) >> >> ( >> https://netlib.org/lapack/explore-html/de/da4/group__double__blas__level1_ga53914429b84cb315da483e71e27ed5c1.html >> >> ) >> >> (10) -> dcabs1(t2) >> >> (10) 2. >> Type: >> DoubleFloat >> >> Tim >> >> On Friday, December 16, 2022 at 8:19:57 AM UTC-5 [email protected] wrote: >> >>> Hello Tim, >>> >>> (trying to find your mail) >>> >>> Do you happen to know if "FortranFunctionCategory" was working ? >>> In other words the Fortran interface? >>> >>> Cheers, >>> __ >>> Greg >>> >>> >>> -- >> You received this message because you are subscribed to the Google Groups >> "FriCAS - computer algebra system" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/fricas-devel/f7ae2284-ac19-46a2-80a1-76b845751c27n%40googlegroups.com >> >> <https://groups.google.com/d/msgid/fricas-devel/f7ae2284-ac19-46a2-80a1-76b845751c27n%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "FriCAS - computer algebra system" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/fricas-devel/fb750ac6-2677-4b2f-93e1-f68f8b9bf9c3n%40googlegroups.com.
