Binary Modular Dataflow Machine
≈ 3 min read
Names in other languages
No names in other languages are held for this entry yet.
Article bodies are English and are not machine-translated. A name in another language is not a translation of the article.
Binary Modular Dataflow Machine (BMDFM) is a software package that enables running an application in parallel on shared memory symmetric multiprocessing (SMP) computers using the multiple processors to speed up the execution of single applications. BMDFM automatically identifies and exploits parallelism due to the static and mainly dynamic scheduling of the dataflow instruction sequences derived from the formerly sequential program. The BMDFM dynamic scheduling subsystem performs a symmetric multiprocessing (SMP) emulation of a tagged-token dataflow machine to provide the transparent dataflow semantics for the applications. No directives for parallel execution are needed.
Binary Modular Dataflow Machine (BMDFM) is a software package that enables running an application in parallel on shared memory symmetric multiprocessing (SMP) computers using the multiple processors to speed up the execution of single applications. BMDFM automatically identifies and exploits parallelism due to the static and mainly dynamic scheduling of the dataflow instruction sequences derived from the formerly sequential program. The BMDFM dynamic scheduling subsystem performs a symmetric multiprocessing (SMP) emulation of a tagged-token dataflow machine to provide the transparent dataflow semantics for the applications. No directives for parallel execution are needed. (setq a (foo0 i)) # a = foo0(i); (setq b (foo1 (+ i 1))) # b = foo1(i+1); (setq b (++ b)) # b++; (outf "a = %d\n" a) # printf("a = %d\n", a); (outf "b = %d\n" b) # printf("b = %d\n", b); The two first statements are independent, so a dataflow engine of BMDFM can run them on different processors or processor's cores. The two last statements can also run in parallel but only after "a" and "b" are computed. The dataflow engine recognizes dependencies automatically because of its ability to build a dataflow graph dynamically at runtime. Additionally, the dataflow engine correctly orders the output stream to output the results sequentially. Thus even after the out-of-order processing the results will appear in a natural way. Suppose that above code fragment now is nested in a loop: (for i 1 1 N (progn # for (i = 1; i The dataflow engine of BMDFM will keep variables "a" and "b" under unique contexts for each iteration. Actually, these are different copies of the variables. A context variable exists until it is referenced by instruction consumers. Later non-referenced contexts will be garbage collected at runtime. Therefore, the dataflow engine can exploit both local parallelism within the iteration and global parallelism as well running multiple iterations simultaneously. BMDFM GitHub Repository BMDFM download page on the BMDFM Official Site BMDFM download page on download3k BMDFM download page on sourceforge
Text derived from https://en.wikipedia.org/w/index.php?oldid=1368271067 under CC BY-SA 4.0; modifications have been made.