Differenze

Queste sono le differenze tra la revisione selezionata e la versione attuale della pagina.

--- magistraleinformaticanetworking:spm:recupero17 [14/11/2017 alle 14:11 (7 anni fa)] – [14 Nov] Marco Danelutto
+++ magistraleinformaticanetworking:spm:recupero17 [13/12/2017 alle 18:08 (7 anni fa)] (versione attuale) – [December 12] Marco Danelutto
@@ Linea 1: / Linea 1: @@
 ===== SPM 2017 recovery edition lessons =====
@@ Linea 103: / Linea 102: @@
 ==== Nov 7 ====
-Introduction to parallel patterns/skeletons
+h: Introduction to parallel patterns/skeletons
   * using [[http://www.openmp.org/|OpenMP]] to implement parallel for (example of simple data parallel pattern) ([[http://www.openmp.org/wp-content/uploads/OpenMP-4.0-C.pdf|Summary of 4.0]], Old [[http://www.openmp.org/wp-content/uploads/omp-hands-on-SC08.pdf|Tutorial]] slides). OpenMP is available in g++ and in icc by specifying the flag **-fopenmp**.
@@ Linea 123: / Linea 122: @@
 ==== 14 Nov ====
+h:
 Sample code
   * [[traduttorecpp|code]] with different implementation of a "map" ...
@@ Linea 137: / Linea 138: @@
 **Assignment**:
   * implement a divide and conquer computation using c++ threads. The computation should look for a word in a text. Recursively divide the text up to the point you either get something of the a given length (parameter), and in this case you look for the string sequentially. Return an integer value counting the occurrences of the string.  (//sub assignment: consider programming the solution in such a  way the code for the divide and conquer strategy may be used to program another divide and conquer application//)
+==== Nov 21 ====
+h
+  * data parallel patterns (map, reduce, google map reduce, stencil, stencil-reduce)
+  * implementation of pattern frameworks (supporting composition)
+    * template based
+    * data flow based
+{{ :magistraleinformaticanetworking:spm:spm21nova.mp4 |H1}}
+{{ :magistraleinformaticanetworking:spm:spm21novb.mp4 |H2}}
+{{ :magistraleinformaticanetworking:spm:spm21no17.pdf |PDF}}
+**Assignments**
+  * read Chapter 4 of the notes
+  * implement the word count with your own google mapreduce in C++ (//use threads to compute in parallel the map phase and again threads to compute in parallel the reduce phase//)
+==== Nov 28 ====
+h
+  * FastFlow programming framework: Introduction, core patterns.
+  * documentation may be found at [[http://calvados.di.unipi.it/fastflow|calvados.di.unipi.it]]
+  * download code (via SVN!!) at [[https://sourceforge.net/projects/mc-fastflow/|SourceForge]]
+{{ :magistraleinformaticanetworking:spm:spm28nova.mp4 |H1}}
+{{ :magistraleinformaticanetworking:spm:spm28novb.mp4 |H2}}
+{{ :magistraleinformaticanetworking:spm:spm28-17.pdf |PDF}}
+**Assignments**
+  * try to write a FF pipeline (with core patterns) that implements the Sieve of Eratostene: one stage emits the stream of integers and stage i filters out all the numbers multiple of i. The last stage prints the results. Try generating 20 integers and using k<20 filter stages.
+=== Afternoon lesson ===
+by David Del Rio Astorga: 1h
+  * Introduction to GrPPI (slides sent as PDF to the course mailing list).
+==== December 12 ====
+Morning:  2h
+  * methodology to develop parallel code with structured programming
+  * final project presentation (see the [[http://didawiki.di.unipi.it/doku.php/magistraleinformaticanetworking/spm/esamirec17#program|exams]] page)
+{{ :magistraleinformaticanetworking:spm:spm12dic17a1.mp4 |Video1}}{{ :magistraleinformaticanetworking:spm:spm12dic17a2.mp4 |Video2}}
+{{ :magistraleinformaticanetworking:spm:spm12dic17a1.pdf |H1}}{{ :magistraleinformaticanetworking:spm:spm12dic17a2.pdf |H2}}
+Afternoon: 2h (By Massimo Torquati)
+  * high level usage of core FastFlow patterns
+  * {{ :magistraleinformaticanetworking:spm:fastflow_overview_spm12dic.pdf |Slides}}
+  * {{ :magistraleinformaticanetworking:spm:farmsquare.tgz |Additional code files}}
+==== End of lessons ====