develooper Front page | perl.perl5.porters | Postings from January 2018

Re: regrepeat()

Thread Previous | Thread Next
From:
Karl Williamson
Date:
January 30, 2018 18:41
Subject:
Re: regrepeat()
Message ID:
615d7152-4bf1-c4db-c2e5-ff29c1d0250f@khwilliamson.com
On 01/02/2018 05:37 AM, Dave Mitchell wrote:
> On Mon, Jan 01, 2018 at 07:08:10PM +0100, demerphq wrote:
>> On 31 December 2017 at 22:36, Karl Williamson <public@khwilliamson.com> wrote:
>>> This function is called during regular expression pattern matching for
>>> things like
>>>
>>>   (foo)+
>>>
>>> to  match as many 'foo's as there are.  There is special code to handle the
>>> case where foo is a single byte, such as in
>>>
>>>   a+
>>>
>>> It turns out that these cases can be sped up dramatically if what we are
>>> matching is a long string of 'a's in a row.  We simply load a word with 4 or
>>> 8 a's and look at the string a word-at-a-time, which uses 1/4 or 1/8 the
>>> number of instructions.  By using a mask, this can be extended to work for
>>>
>>>   [aA]+
>>>
>>> as well.  The code in each case is just over 20 lines of C.
>>>
>>> My question is, does this happen often enough in real life to justify the
>>> extra code?
>>>
>>> Leon pointed out that in DNA, there may be longish strings of 'A's.
>>
>> I think this is highly likely to be worth it.
> 
> +1
> 

Now done by
ab1efbdc1f74b2f4db076efa0b4d54f387d74efe
070e8b2ef4f827a7e0d3199f7b37883a09545802

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About