Changeset 1370 for docs


Ignore:
Timestamp:
Aug 24, 2011, 5:19:43 PM (8 years ago)
Author:
lindanl
Message:

modifications

Location:
docs/HPCA2012
Files:
3 edited

Legend:

Unmodified
Added
Removed
  • docs/HPCA2012/01-intro.tex

    r1363 r1370  
    1313can be productively used in scaling our
    1414chips~\cite{esmaeilzadeh-isca-2011}. This is because exploiting
    15 parallelism across multiple cores tends to require heavweight threads
     15parallelism across multiple cores tends to require heavy weight threads
    1616that are difficult to manage and synchronize.
    1717
  • docs/HPCA2012/05-corei3.tex

    r1366 r1370  
    1818parsers as they experience other overheads related to branch
    1919mispredictions. Parabix's data reorganization significantly improves
    20 the overall cache miss rate. We experience ?$\times$ less misses at
    21 the L1 and ?$\times$ less misses at the L2 level. The improved cache
     20the overall cache miss rate. We experience 7$\times$ less misses than
     21Expat and 25 $\times$ less misses than Xerces at the L1 and 104$\times$ less misses than
     22Expat and 15 $\times$ less misses than Xerces at the L2 level. The improved cache
    2223utilization keeps the SIMD units busy and prevent memory related
    2324stalls. Note that cache misses also cause increased application energy
  • docs/HPCA2012/06-scalability.tex

    r1339 r1370  
    3737\includegraphics[width=0.40\textwidth]{plots/Expat_scalability.pdf}
    3838}
    39 \caption{Average Performance Parabix vs. Expat (y-axis: CPU Cycles per kB)}
     39\caption{Average Performance Parabix vs. Expat (y-axis: ns per kB)}
    4040\label{Scalability}
    4141\end{figure}
Note: See TracChangeset for help on using the changeset viewer.