comment on

The Bit::Vector code needs new_Dec, not new_Hex. Here are are some additions to your code showing Data::BitStream::XS and two alternatives using Bit::Vector. Neither are improvements over the fast simple methods, but they do show that both are at least reasonable solutions.

$start = time;
my $stream = Data::BitStream::XS->new;
# $stream->write(32, $n) for 1 .. $ITERS; # Below is faster
$stream->put_raw(pack("N",$n) x $ITERS, 32*$ITERS);
$stream->rewind_for_read;
for (1 .. $ITERS ) {
 stuff($stream->read(14), $stream->read(6), $stream->read(6), $stream-
+>read(6));
}
printf "   D::B::XS took: %.12f seconds\n", ( time() - $start ) / $ITE
+RS;

$start = time;
for (1 .. $ITERS ) {
    my $vector = Bit::Vector->new_Dec(32, $n);
    stuff($vector->Chunk_Read(14, 18),
          $vector->Chunk_Read( 6, 12),
          $vector->Chunk_Read( 6, 6),
          $vector->Chunk_Read( 6, 0)
    );
}
printf "Bit::Vector took: %.12f seconds\n", ( time() - $start ) / $ITE
+RS;

$start = time;
my $vector = Bit::Vector->new(32*$ITERS);
$vector->Word_List_Store(($n) x $ITERS);
for (0 .. $ITERS-1 ) {
    stuff($vector->Chunk_Read(14, 32*$_+18),
          $vector->Chunk_Read( 6, 32*$_+12),
          $vector->Chunk_Read( 6, 32*$_+6),
          $vector->Chunk_Read( 6, 32*$_+0)
    );
}
printf "Bit::Vector took: %.12f seconds\n", ( time() - $start ) / $ITE
+RS;
[download]

I get on my computer:

  Shift&and took: 0.000000193217 seconds
    Lookup  took: 0.000000180829 seconds
    Lookup2 took: 0.000000215928 seconds
   (un)pack took: 0.000001851533 seconds
   D::B::XS took: 0.000000643722 seconds
Bit::Vector took: 0.000001739788 seconds
Bit::Vector took: 0.000000957755 seconds
[download]

It really depends on what the usage model is. Both Data::BitStream and Bit::Vector are going to suffer if you insist on making a new object for each word. They will do much better if you let them create a single object and fill it, then pull the data out in arbitrary sized chunks. If the OP had all the values in one chunk or array then this would be more appropriate.

One could also write a new coding role for Data::BitStream to handle retrieving the values using code similar to one of the lookup methods, but there would have to be some real reason to want a stream vs. just doing it using one of the lookup methods.

In reply to Re^4: Efficient bit-twiddling in Perl. by danaj
in thread Efficient bit-twiddling in Perl. by BrowserUk

Are you posting in the right place? Check out Where do I post X? to know for sure.
Posts may use any of the Perl Monks Approved HTML tags. Currently these include the following:
<code> <a> <b> <big> <blockquote> <br /> <dd> <dl> <dt> <em> <font> <h1> <h2> <h3> <h4> <h5> <h6> <hr /> <i> <li> <nbsp> <ol> <p> <small> <strike> <strong> <sub> <sup> <table> <td> <th> <tr> <tt> <u> <ul>
Snippets of code should be wrapped in <code> tags not <pre> tags. In fact, <pre> tags should generally be avoided. If they must be used, extreme care should be taken to ensure that their contents do not have long lines (<70 chars), in order to prevent horizontal scrolling (and possible janitor intervention).
Want more info? How to link or How to display code and escape characters are good places to start.


P is for Practical
	PerlMonks