| |
X10 Code Examples
The examples listed here are either small applications extracted from computational kernels or scientific computing benchmarks. Most of them are covered by our presentations:
- HPC Challenge Submission for X10 (SC 2007)
pdf (won Class 2 award: Most Productive Research Implementation)
- Programming Scientific Computations in X10 (PGAS 2006, X10 Demo SC 2006)
pdf
- X10: a High-Productivity Approach to High Performance Programming (HPC Challenge Class 2 Award Submission SC 2006)
pdf
More examples (not limited to the domain of scientific computing) will be added.
|
The following examples demonstrate in X10 how to build and iterate over distributed arrays, and how to coordinate activities. Each of them either has a dependent-type version or itself uses dependent types. Note that our implementation of dependent types is still an ongoing effort. More support and improvements are on the way.
|
| |
We have been implementing benchmarks in X10 (literal translation to X10 or implementation from scratch). With the new x10c++ compiler which translates X10 programs into C++ SPMD programs communicating using an active messaging infrastructure and the runtime system based on IBM's LAPI (Low-level API), performance comparable to hand-written MPI implementations is realized for well-known benchmarks such as FT in NPB, and Random Access, Stream, and FFT in HPCC. Some of the X10 programs posted below also have a "static" version which is the one to be compiled with x10c++. Note that the current x10c++ compiler and runtime system (to be released soon) only implements a subset of X10 functionalities.
|
|
