|
Size: 1277
Comment:
|
← Revision 23 as of 2008-11-14 13:42:16 ⇥
Size: 2637
Comment: converted to 1.6 markup
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 7: | Line 7: |
| At a minimum, make sure you are completely familiar with the following before the workshop, in order: | At a minimum, make sure you read the following before the workshop, in order: |
| Line 9: | Line 9: |
| 1. [http://www.llnl.gov/computing/tutorials/parallel_comp/ Introduction to Parallel Computing] | 1. [[http://www.llnl.gov/computing/tutorials/parallel_comp/|Introduction to Parallel Computing]] |
| Line 11: | Line 11: |
| 2. [http://www.llnl.gov/computing/tutorials/pthreads/ POSIX Threads Programming] | 2. [[http://www.llnl.gov/computing/tutorials/pthreads/|POSIX Threads Programming]] |
| Line 13: | Line 13: |
| 3. [http://www.llnl.gov/computing/tutorials/openMP/ OpenMP tutorial] 4. [http://www.llnl.gov/computing/tutorials/mpi/ MPI Tutorial] == Optional Reading List == 1. [http://mail.python.org/pipermail/python-list/2006-May/382955.html List post about OpenMp and Python] 2. [http://aspn.activestate.com/ASPN/search?query=parallel&x=0&y=0§ion=PYTHONCKBK&type=Subsection Python Parallel Recipes] |
3. [[http://www.llnl.gov/computing/tutorials/mpi/|MPI Tutorial]] |
| Line 25: | Line 18: |
| Skim or read these as time permits: | Skim or read as time permits: |
| Line 27: | Line 20: |
| * [http://ocw.mit.edu/OcwWeb/Mathematics/18-337JSpring-2005/LectureNotes/index.htm Applied Parallel Computing Lectures] | 1. [[http://modular.math.washington.edu/msri07/read/|Miscellaneous Papers and Files]] 1. [[http://www.llnl.gov/computing/tutorials/openMP/|OpenMP tutorial]] and [[http://gcc.gnu.org/projects/gomp/|GOMP]] 2. [[http://ocw.mit.edu/OcwWeb/Mathematics/18-337JSpring-2005/LectureNotes/index.htm|Applied Parallel Computing Lectures]] 3. [[http://www.eecs.berkeley.edu/Pubs/TechRpts/2006/EECS-2006-1.pdf|Alternatives to threads]] 4. [[http://www.erlang.org/|Erlang programming language]] 5. [[http://erights.org/|E programming language]] 6. [[http://twistedmatrix.com/projects/core/documentation/howto/async.html|About Twisted's Deferred Objects]] == SAGE-Related Reading List == Because of something called the ''Global Interpreter Lock'' (or GIL) in the Python ''interpreter'', threading is pretty useless for speeding up serious mathematical compute applications. For *multi-threaded* programs to be truly useful for SAGE computation, they will have to be implemented entirely in C/C++/Fortran libraries and SageX extension modules in blocks of code that do not make any Python/C API calls. 1. [[http://ldp.paradoxical.co.uk/LDP/LGNET/107/pai.html|Python threading tutorial]] 2. [[http://docs.python.org/api/threads.html|Discussion About the GIL in Python]] 3. [[http://mail.python.org/pipermail/python-list/2006-May/382955.html|List post about OpenMp and Python]] 4. [[http://aspn.activestate.com/ASPN/search?query=parallel&x=0&y=0§ion=PYTHONCKBK&type=Subsection|Python Parallel Recipes]] 5. [[http://poshmodule.sourceforge.net/|POSH -- Python object sharing; a way to get around using threads for what we need]] 6. [[http://www.parallelpython.com/|Parallel Python -- is this useful for something? (or is it only stupid closed-source poppycock?) -- it's another way to get around threads]] |
Preworkshop Reading Materials
All participants at the MSRI Parallel Computation Workshop will be assumed to have a basic background in parallel computation already -- there will be no basic background tutorials or lectures. Fortunately, anybody can get such a background in a day by reading the documents listed below under required reading.
Required Reading List
At a minimum, make sure you read the following before the workshop, in order:
Other Relevant Documents
Skim or read as time permits:
OpenMP tutorial and GOMP
SAGE-Related Reading List
Because of something called the Global Interpreter Lock (or GIL) in the Python interpreter, threading is pretty useless for speeding up serious mathematical compute applications. For *multi-threaded* programs to be truly useful for SAGE computation, they will have to be implemented entirely in C/C++/Fortran libraries and SageX extension modules in blocks of code that do not make any Python/C API calls.