[talks] T Jablin preFPO
Melissa M. Lawson
mml at CS.Princeton.EDU
Mon Nov 14 15:18:36 EST 2011
Tom Jablin will present his preFPO on Monday November 21 at 3PM in Room 402.
The members of his committee are: David august, advisor; Doug Clark and
Scott Mahlke (UMichigan), readers; Margaret Martonosi and David Walker, nonreaders.
Everyone is invited to attend his talk. His abstract follows below.
GPUs are flexible parallel processors capable of accelerating real applications.
To exploit them, programmers rewrite programs in new languages using intimate
knowledge of the underlying hardware. This is a step backwards in abstraction
and ease of use from sequential programming. When implementing sequential
applications, programmers focus on programs' high-level algorithmic structure,
allowing the compiler to target the peculiarities of specific
hardware. Automatic parallelization can return ease of use and hardware
abstraction to programmers. This proposal suggests techniques for automatically
parallelizing ordinary sequential C codes for GPUs using DOALL, pipelined, and
speculative techniques. The key contributions are:
the first software-only shared memory system for GPUs,
and the first fully-automatic pipeline parallelization system for GPUs.
Combining these three components yields a fully-automatic parallelizing compiler
for GPUs that is applicable to many general purpose applications.
More information about the talks