User:Devon McCormick

submited by
Style Pass
2024-06-11 14:00:04

I curate the SIGAPL (Special Interest Group on APL) page under the auspices of ACM's SIGPLAN (Special Interest Group on Programming LANguages.)

o Array-thinking: an explanation of the fundamental difference between array-thinking and scalar-thinking and how the former helps us think about algorithms.

o An early (unsuccessful) attempt to use the J database "JDB" on the "Netflix Challenge" data. It appears to have been unsuccessful because this dataset was too large for JDB to handle.

o BigFiles - work with files larger than 2 GB: Scripts/Working with Big Files. This is now largely obsolete with the advent of 64-bit versions of J that have extended the scope of the file primitives to handle larger files.

o Detailed example of how to simply parallelize code in J. This assumes the (common) sort of problem where coordination of the various parallel processes can be handled at the point where a "master" process spins them off and that the problem domain lends itself to "coarse" parallelism. I currently run two multi-core J routines almost every day using this technique and it's worked well for years now with substantial speed-ups and little difficulty.

o Complete examples of parallel J code presented at the APL2010 conference in Berlin, Germany, in September, 2010. These are the two examples of parallel code tasks, mentioned above, as well as an implementation of mutex (mutual exclusion), all written in J (version 6.02).

Leave a Comment
Related Posts