3 msgNewbie questions about gcc / glibc
3 msgTwo quick C++ performace inquiries
4 msgSigned division with rounding towards -infinity...
72 msgLink tests after GCC_NO_EXECUTABLES
10 msgporting problem again: ICE in add_clobbers
1 msggcc-4.1-20070917 is now available
9 msgbootstrap failure on ppc64-linux: ICE in set_va...
9 msgAda rts fails to build on ppc64?
4 msgOnline Job Representative Staff
3 msgDefine __NO_MATH_INLINES for x86 -ffast-math?
1 msglibstdc++-v3/include/precompiled weirdness
2 msgShould -ftrapv check type conversion?
2 msgTrap floating-point exceptions?
6 msgCoding conventions -- command line option vs co...
1 msgLegal Experts Directory
2 msgQuestion about rs6000 zero_fp_constant predicate

Nvidia PTX backend
\ Justin Bennett (15 Sep 2007)

1 msgNew GCC ICI (with plug-ins for dynamic pass reo...
1 msg__if_exists __if_not_exists alternative
1 msgHaunted Halloween fun at Mountain Village 1890!
Subject:Nvidia PTX backend
Group:Gcc
From:Justin Bennett
Date:15 Sep 2007


I'm thinking about starting a project to add a backend for the Nvidia
CUDA PTX instruction set, and eventually a front end that understands
the .cu files. Has anybody taken a look at doing this? Just wanted
to check to make sure I'm not duplicating work before I begin planning
out the changes needed.

Also, I expect there will be a lot changes needed to support this, so
if anybody else is interested let me know.

Justin


© 2004-2008 readlist.com