Par4all: Auto-Parallelizing C and Fortran for the CUDA Architecture
Par4all: Auto-Parallelizing C and Fortran for the CUDA Architecture
Par4all: Auto-Parallelizing C and Fortran for the CUDA Architecture
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
•Par4All ◮<br />
• <strong>Auto</strong>matic parallelization (C & <strong>Fortran</strong> to OpenMP)<br />
• Distributed memory computing with OpenMP-to-MPI translation<br />
(STEP project)<br />
• Generic vectorization <strong>for</strong> SIMD instructions (SSE, VMX...) (SAC<br />
project)<br />
• Parallelization <strong>for</strong> embedded systems (SCALOPES)<br />
• Compilation <strong>for</strong> hardware accelerators (Ter@PIX, SPoC, SIMD,<br />
FPGA...)<br />
• High-level hardware accelerators syn<strong>the</strong>sis generation <strong>for</strong> FPGA<br />
• Reverse engineering & decompiler (reconstruction from binary to<br />
C)<br />
Logical next stop<br />
GPU! �<br />
�Par4All in <strong>CUDA</strong> — GPU conference 10/1/2009<br />
HPC Project, Mines ParisTech, TÉLÉCOM Bretagne, RPI Ronan KERYELL et al. 9 / 46