PROGRESS REPORT

■ THIS WEEK  (7/15~7/20)

□ Phoenix 2.0

about degradation after 32 threads:  

first, I conjectured that the reason was skewed distribution of words.

so I randomly generated input file using a-z and white space.

but there still exists slight degradation.

now, I divided the start_worker function into 7 parts and find the degradation part.

here I attached my work. 

phoenix_timeval.pdf

□ Booksim 2.0

the simulation script is still running.

the result file using uniform traffic is attached.

I will analysis the result when other data come out.

Uniform_result.pdf

■ NEXT WEEK  (7/21~7/27)

□ Phoenix 

Test more with huge data set and re-plot the speedup graph.

Find way to determine working set size and L2.

□ Simics & Gems

a little later..

□ Booksim 2.0

Plot all graph using other traffics and analysis the result.