Ignore:
Timestamp:
May 15, 2015, 8:33:45 AM (4 years ago)
Author:
cameron
Message:

Wrap up

File:
1 edited

Legend:

Unmodified
Added
Removed
  • docs/Working/icGrep/evaluation.tex

    r4560 r4561  
    279279on a newer Intel i7-4700MQ machine, considering three SIMD ISA alternatives
    280280and both single-threaded and multi-threaded versions.
    281 All speedups are relative to the base single-threaded SSE2 performance on the i7-2600 machine.
     281All speedups are relative to the base single-threaded SSE2 performance on this machine,
     282which is given in seconds per GB in the first column.
    282283%
    283284The SSE2 results are again using the generic binaries compiled for compatibility
     
    338339\end{table}
    339340
     341In each case, the use of three-operand form with AVX1 confers a slight
     342speedup.  The change to use 256 bits with AVX2 gives a further overall improvement,
     343but some mixed results due to the limitations of 256 bit addition.   Combining
     344the AVX2 ISA with multithreading gives and average overall 61\% speedup compared to base.
    340345
    341346% Interestingly, the SSE2 column of Table \ref{table:relperf} shows that by simply using a newer hardware and compiler
Note: See TracChangeset for help on using the changeset viewer.