Ignore:
Timestamp:
Aug 31, 2011, 7:30:44 PM (8 years ago)
Author:
ashriram
Message:

final

File:
1 edited

Legend:

Unmodified
Added
Removed
  • docs/HPCA2012/06-scalability.tex

    r1411 r1418  
    1212
    1313Our results demonstrate that Parabix-XML's optimizations complement
    14 newer hardware improvements. For bit-stream processing,
     14newer hardware improvements. For bit stream processing,
    1515\CITHREE{} has a 40\% performance increase over \CO{};
    1616similarly, \SB{} has a 20\% improvement compared to
     
    2727frequency and microarchitecture improvements while \SB{}'s gains can
    2828be mainly attributed to the architecture.
    29 
    3029Figure \ref{power_Parabix2} shows the average power consumption of
    3130Parabix-XML over each workload and as executed on each of the processor
    3231cores: \CO{}, \CITHREE\ and \SB{}.  Each
    3332generation of processor seem to bring with them 25--30\% improvement
    34 in power consumption over the previous generation. Overall,
    35 Parabix-XML on \SB\ consumes 72\%--75\% less energy than it did on \CO{}.
     33in power consumption over the previous generation. Parabix-XML on \SB\ consumes 72\%--75\% less energy than it did on \CO{}.
    3634
    3735
Note: See TracChangeset for help on using the changeset viewer.