v0.2.1-beta

SkePi

Data parallel and stream parallel skeletons implemented in erlang.

DONE:

• farm stream parallel skeleton
• pipe stream parallel skeleon
• ability to compose stream skeletons
• map data parallel skeleton
• reduce data parallel skeleton (parallel fold version)
• mapreduce data parallel skeletons (smart version, naive version, dictionaries version)
• usage examples
• logging, profiling and benchmarking utilities
• utility functions for purposes such as chunking, file indexing and more

FIXED:

• mapreduce sequential version (used in example/test_mapred_google.erl)

TODO:

• mapreduce data parallel skeleton (hadoop/buckets version)
• fault-tolerance
• documentation
• better error handling

v0.2.0-beta

SkePi

Data parallel and stream parallel skeletons implemented in erlang.

DONE:

• farm stream parallel skeleton
• pipe stream parallel skeleon
• ability to compose stream skeletons
• map data parallel skeleton
• reduce data parallel skeleton (parallel fold version)
• mapreduce data parallel skeletons (smart version, naive version, dictionaries version)
• usage examples
• logging, profiling and benchmarking utilities
• utility functions for purposes such as chunking, file indexing and more

FIXED:

• mapreduce sequential version (used in example/test_mapred_google.erl)

TODO:

• mapreduce data parallel skeleton (hadoop/buckets version)
• fault-tolerance
• documentation
• better error handling

ske-pi

SkePi

Data parallel and stream parallel skeletons implemented in erlang.

DONE:

• farm stream parallel skeleton
• pipe stream parallel skeleon
• ability to compose stream skeletons
• map data parallel skeleton
• mapreduce data parallel skeletons (naive version, dictionaries version)
• usage examples
• logging, profiling and benchmarking utilities
• utility functions for purposes such as chunking, file indexing and more

TOFIX:

• mapreduce sequential version (used in example/test_mapred_google.erl)

TODO:

• reduce data parallel skeleton (parallel fold version)
• mapreduce data parallel skeletons (smart version)
• mapreduce data parallel skeleton (hadoop/buckets version)
• fault-tolerance
• documentation
• better error handling