math/blis: update to release 0.5.0.

Most important addition to this release are mixed precision interfaces. This
allows one to, e.g., put in single precision data but carry the computation
out in double precision accuracy.

Note that flame/blis is the upstream for the BLIS library. AMD has forked
BLIS and releases independently from the FLAME project but is strightly
downstream.
This commit is contained in:
Johannes M Dieterich 2018-11-05 01:09:48 +00:00
parent 004701fd55
commit 235d33c0da
Notes: svn2git 2021-03-31 03:12:20 +00:00
svn path=/head/; revision=484161
3 changed files with 9 additions and 6 deletions

View file

@ -2,7 +2,7 @@
# $FreeBSD$
PORTNAME= blis
PORTVERSION= 0.4.0
PORTVERSION= 0.5.0
CATEGORIES= math
MAINTAINER= jmd@FreeBSD.org
@ -51,11 +51,11 @@ CONFIGURE_ARGS+= --enable-cblas
.if ${ARCH} == amd64
LIBNAME= x86_64
CONFIGURE_ARGS+= x86_64
PLIST_SUB+= X8664=""
PLIST_SUB+= ARCH="x86_64"
.else
LIBNAME= generic
CONFIGURE_ARGS+= generic
PLIST_SUB+= X8664="@comment "
PLIST_SUB+= ARCH="generic"
.endif
CC= ${LOCALBASE}/bin/clang60

View file

@ -1,3 +1,3 @@
TIMESTAMP = 1532834178
SHA256 (flame-blis-0.4.0_GH0.tar.gz) = 9c7efd75365a833614c01b5adfba93210f869d92e7649e0b5d9edc93fc20ea76
SIZE (flame-blis-0.4.0_GH0.tar.gz) = 1871027
TIMESTAMP = 1541177732
SHA256 (flame-blis-0.5.0_GH0.tar.gz) = 1a004d69c139e8a0448c6a6007863af3a8c3551b8d9b8b73fe08e8009f165fa8
SIZE (flame-blis-0.5.0_GH0.tar.gz) = 2262591

View file

@ -4,3 +4,6 @@ lib/libblis.a
lib/libblis.so
lib/libblis.so.1
lib/libblis.so.1.0.0
%%DATADIR%%/common.mk
%%DATADIR%%/config.mk
%%DATADIR%%/config/%%ARCH%%/make_defs.mk