GitHub - syedsayir/mapReduceText

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 4 Commits
RawText		RawText
bak		bak
textReplicated		textReplicated
farooqs_aalawneh.zip		farooqs_aalawneh.zip
ff		ff
mapreduce.c		mapreduce.c
mapreduce.h		mapreduce.h
mpi.sub		mpi.sub
omp.sub		omp.sub
readme.tx		readme.tx
ser.tx		ser.tx

Repository files navigation

1) Forking multiple threads for Readers not beneficial.
2) Using parallel sections for readers and mappers not beneficial.
3) can't really have reducer in parallel with anything else. Besides, reduction is a very tiny percentage of the total time.

4) cache competition on L3 might be causing the 16 threads to be missing more and values to messup.