7/13/2010

Perl optimization article

What a great article summarizing the optimization of perl scripting. Take a look here:
http://www.flirble.org/~nick/P/Fast_Enough/

Or Quoted here:
* When perl is not quite fast enough
*
o Introduction
o Obvious things
o Compromises
o Banish the demons of stupidity
o Intermission
o Tests
o What causes slowness
o Step by step
o Small easy things
o Needless importing is slow
o regexps
o Devel::DProf
o Benchmark
o What causes slowness in perl?
o Ops are bad, m'kay
o Memoize
o Miscellaneous
o yay for y
o Ops are bad, m'kay
o How to make perl fast enough

When perl is not quite fast enough

This is the script for my talk at YAPC::EU::2002. It's not a transcript of what I actually said; rather it collects together the notes that were in the pod source to the slides, the notes scribbled on various bits of paper, the notes that were only in my head, and tries to make a coherent text. I've also tried to add in the useful feedback I got - sometimes I can even remember who said it and so give them credit.

The slides are here, and hopefully it will be obvious where the slide changes are.
Introduction

So you have a perl script. And it's too slow. And you want to do something about it. This is a talk about what you can do to speed it up, and also how you try to avoid the problem in the first place.
Obvious things

Find better algorithm
Your code runs in the most efficient way that you can think of. But maybe someone else looked at the problem from a completely different direction and found an algorithm that is 100 times faster. Are you sure you have the best algorithm? Do some research.
Throw more hardware at it
If the program doesn't have to run on many machines may be cheaper to throw more hardware at it. After all, hardware is supposed to be cheap and programmers well paid. Perhaps you can gain performance by tuning your hardware better; maybe compiling a custom kernel for your machine will be enough.
mod_perl
For a CGI script that I wrote, I found that even after I'd shaved everything off it that I could, the server could still only serve 2.5 per second. The same server running the same script under mod_perl could serve 25 per second. That's a factor of 10 speedup for very little effort. And if your script isn't suitable for running under mod_perl there's also fastcgi (which CGI.pm supports). And if your script isn't a CGI, you could look at the persistent perl daemon, package PPerl on CPAN.
Rewrite in C, er C++, sorry Java, I mean C#, oops no ...
Of course, one final "obvious" solution is to re-write your perl program in a language that runs as native code, such as C, C++, Java, C# or whatever is currently flavour of the month.

But these may not be practical or politically acceptable solutions.
Compromises

So you can compromise.

XS
You may find that 95% of the time is spent in 5% of the code, doing something that perl is not that efficient at, such as bit shifting. So you could write that bit in C, leave the rest in perl, and glue it together with XS. But you'd have to learn XS and the perl API, and that's a lot of work.
Inline
Or you could use Inline. If you have to manipulate perl's internals then you'll still have to learn perl's API, but if all you need is to call out from perl to your pure C code, or someone else's C library then Inline makes it easy.

Here's my perl script making a call to a perl function rot32. And here's a C function rot32 that takes 2 integers, rotates the first by the second, and returns an integer result. That's all you need! And you run it and it works.

#!/usr/local/bin/perl -w
use strict;

printf "$_:\t%08X\t%08X\n", rot32 (0xdead, $_), rot32 (0xbeef, -$_)
foreach (0..31);

use Inline C => <<'EOC';

unsigned rot32 (unsigned val, int by) {
if (by >= 0)
return (val >> by) | (val << (32 - by));
return (val << -by) | (val >> (32 + by));
}
EOC
__END__

0: 0000DEAD 0000BEEF
1: 80006F56 00017DDE
2: 400037AB 0002FBBC
3: A0001BD5 0005F778
4: D0000DEA 000BEEF0
...

Compile your own perl?
Are you running your script on the perl supplied by the OS? Compiling your own perl could make your script go faster. For example, when perl is compiled with threading, all its internal variables are made thread safe, which slows them down a bit. If the perl is threaded, but you don't use threads then you're paying that speed hit for no reason. Likewise, you may have a better compiler than the OS used. For example, I found that with gcc 3.2 some of my C code run 5% faster than with 2.9.5. [One of my helpful hecklers in the audience said that he'd seen a 14% speedup, (if I remember correctly) and if I remember correctly that was from recompiling the perl interpreter itself]
Different perl version?
Try using a different perl version. Different releases of perl are faster at different things. If you're using an old perl, try the latest version. If you're running the latest version but not using the newer features, try an older version.

Banish the demons of stupidity

Are you using the best features of the language?

hashes
There's a Larry Wall quote - Doing linear scans over an associative array is like trying to club someone to death with a loaded Uzi.

I trust you're not doing that. But are you keeping your arrays nicely sorted so that you can do a binary search? That's fast. But using a hash should be faster.
regexps
In languages without regexps you have to write explicit code to parse strings. perl has regexps, and re-writing with them may make things 10 times faster. Even using several with the \G anchor and the /gc flags may still be faster.

if ( /\G.../gc ) {
...
} elsif ( /\G.../gc ) {
...
} elsif ( /\G.../gc ) {

pack and unpack
pack and unpack have far too many features to remember. Look at the manpage - you may be able to replace entire subroutines with just one unpack.
undef
undef. what do I mean undef?

Are you calculating something only to throw it away?

For example the script in the Encode module that compiles character conversion tables would print out a warning if it saw the same character twice. If you or I build perl we'll just let those build warnings scroll off the screen - we don't care - we can't do anything about it. And it turned out that keeping track of everything needed to generate those warnings was slowing things down considerably. So I added a flag to disable that code, and perl 5.8 defaults to use it, so it builds more quickly.

Intermission

Various helpful hecklers (most of London.pm who saw the talk (and I'm counting David Adler as part of London.pm as he's subscribed to the list)) wanted me to remind people that you really really don't want to be optimising unless you absolutely have to. You're making your code harder to maintain, harder to extend, and easier to introduce new bugs into. Probably you've done something wrong to get to the point where you need to optimise in the first place.

I agree.

Also, I'm not going to change the running order of the slides. There isn't a good order to try to describe things in, and some of the ideas that follow are actually more "good practice" than optimisation techniques, so possibly ought to come before the slides on finding slowness. I'll mark what I think are good habits to get into, and once you understand the techniques then I'd hope that you'd use them automatically when you first write code. That way (hopefully) your code will never be so slow that you actually want to do some of the brute force optimising I describe here.
Tests

Must not introduce new bugs
The most important thing when you are optimising existing working code is not to introduce new bugs.
Use your full regression tests :-)
For this, you can use your full suite of regression tests. You do have one, don't you?

[At this point the audience is supposed to laugh nervously, because I'm betting that very few people are in this desirable situation of having comprehensive tests written]
Keep a copy of original program
You must keep a copy of your original program. It is your last resort if all else fails. Check it into a version control system. Make an off site backup. Check that your backup is readable. You mustn't lose it.
In the end, your ultimate test of whether you've not introduced new bugs while optimising is to check that you get identical output from the optimised version and the original. (With the optimised version taking less time).

What causes slowness

CPU
It's obvious that if you script hogs the CPU for 10 seconds solid, then to make it go faster you'll need to reduce the CPU demand.
RAM
A lesser cause of slowness is memory.

perl trades RAM for speed
One of the design decisions Larry made for perl was to trade memory for speed, choosing algorithms that use more memory to run faster. So perl tends to use more memory.
getting slower (relative to CPU)
CPUs keep getting faster. Memory is getting faster too. But not as quickly. So in relative terms memory is getting slower. [Larry was correct to choose to use more memory when he wrote perl5 over 10 years ago. However, in the future CPU speed will continue to diverge from RAM speed, so it might be an idea to revisit some of the CPU/RAM design trade offs in parrot]
memory like a pyramid

You can never have enough memory, and it's never fast enough.

Computer memory is like a pyramid. At the point you have the CPU and its registers, which are very small and very fast to access. Then you have 1 or more levels of cache, which is larger, close by and fast to access. Then you have main memory, which is quite large, but further away so slower to access. Then at the base you have disk acting as virtual memory, which is huge, but very slow.

Now, if your program is swapping out to disk, you'll realise, because the OS can tell you that it only took 10 seconds of CPU, but 60 seconds elapsed, so you know it spent 50 seconds waiting for disk and that's your speed problem. But if your data is big enough to fit in main RAM, but doesn't all sit in the cache, then the CPU will keep having to wait for data from main RAM. And the OS timers I described count that in the CPU time, so it may not be obvious that memory use is actually your problem.

This is the original code for the part of the Encode compiler (enc2xs) that generates the warnings on duplicate characters:

if (exists $seen{$uch}) {
warn sprintf("U%04X is %02X%02X and %02X%02X\n",
$val,$page,$ch,@{$seen{$uch}});
}
else {
$seen{$uch} = [$page,$ch];
}

It uses the hash %seen to remember all the Unicode characters that it has processed. The first time that it meets a character it won't be in the hash, the exists is false, so the else block executes. It stores an arrayref containing the code page and character number in that page. That's three things per character, and there are a lot of characters in Chinese.

If it ever sees the same Unicode character again, it prints a warning message. The warning message is just a string, and this is the only place that uses the data in %seen. So I changed the code - I pre-formatted that bit of the error message, and stored a single scalar rather than the three:

if (exists $seen{$uch}) {
warn sprintf("U%04X is %02X%02X and %04X\n",
$val,$page,$ch,$seen{$uch});
}
else {
$seen{$uch} = $page << 8 | $ch;
}

That reduced the memory usage by a third, and it runs more quickly.

Step by step

How do you make things faster? Well, this is something of a black art, down to trial and error. I'll expand on aspects of these 4 points in the next slides.

What might be slow?
You need to find things that are actually slow. It's no good wasting your effort on things that are already fast - put it in where it will get maximum reward.
Think of re-write
But not all slow things can be made faster, however much you swear at them, so you can only actually speed things up if you can figure out another way of doing the same thing that may be faster.
Try it
But it may not. Check that it's faster and that it gives the same results.
Note results
Either way, note your results - I find a comment in the code is good. It's important if an idea didn't work, because it stops you or anyone else going back and trying the same thing again. And it's important if a change does work, as it stops someone else (such as yourself next month) tidying up an important optimisation and losing you that hard won speed gain.

By having commented out slower code near the faster code you can look back and get ideas for other places you might optimise in the same way.

Small easy things

These are things that I would consider good practice, so you ought to be doing them as a matter of routine.

AutoSplit and AutoLoader
If you're writing modules use the AutoSplit and AutoLoader modules to make perl only load the parts of your module that are actually being used by a particular script. You get two gains - you don't waste CPU at start up loading the parts of your module that aren't used, and you don't waste the RAM holding the the structures that perl generates when it has compiled code. So your modules load more quickly, and use less RAM.

One potential problem is that the way AutoLoader brings in subroutines makes debugging confusing, which can be a problem. While developing, you can disable AutoLoader by commenting out the __END__ statement marking the start of your AutoLoaded subroutines. That way, they are loaded, compiled and debugged in the normal fashion.

...
1;
# While debugging, disable AutoLoader like this:
# __END__
...

Of course, to do this you'll need another 1; at the end of the AutoLoaded section to keep use happy, and possibly another __END__.

Schwern notes that commenting out __END__ can cause surprises if the main body of your module is running under use strict; because now your AutoLoaded subroutines will suddenly find themselves being run under use strict. This is arguably a bug in the current AutoSplit - when it runs at install time to generate the files for AutoLoader to use it doesn't add lines such as use strict; or use warnings; to ensure that the split out subroutines are in the same environment as was current at the __END__ statement. This may be fixed in 5.10.

Elizabeth Mattijsen notes that there are different memory use versus memory shared issues when running under mod_perl, with different optimal solutions depending on whether your apache is forking or threaded.
=pod @ __END__
If you are documenting your code with one big block of pod, then you probably don't want to put it at the top of the file. The perl parser is very fast at skipping pod, but it's not magic, so it still takes a little time. Moreover, it has to read the pod from disk in order to ignore it.

#!perl -w
use strict;

=head1 You don't want to do that

big block of pod

=cut

...
1;
__END__

=head1 You want to do this

If you put your pod after an __END__ statement then the perl parser will never even see it. This will save a small amount of CPU, but if you have a lot of pod (>4K) then it might also mean that the last disk block(s) of a file are never even read in to RAM. This may gain you some speed. [A helpful heckler observed that modern raid systems may well be reading in 64K chunks, and modern OSes are getting good at read ahead, so not reading a block as a result of =pod @ __END__ may actually be quite rare.]

If you are putting your pod (and tests) next to their functions' code (which is probably a better approach anyway) then this advice is not relevant to you.

Needless importing is slow

Exporter is written in perl. It's fast, but not instant.

Most modules are able to export lots of their functions and other symbols into your namespace to save you typing. If you have only one argument to use, such as

use POSIX; # Exports all the defaults

then POSIX will helpfully export its default list of symbols into your namespace. If you have a list after the module name, then that is taken as a list of symbols to export. If the list is empty, no symbols are exported:

use POSIX (); # Exports nothing.

You can still use all the functions and other symbols - you just have to use their full name, by typing POSIX:: at the front. Some people argue that this actually makes your code clearer, as it is now obvious where each subroutine is defined. Independent of that, it's faster:
use POSIX; use POSIX ();
0.516s 0.355s
use Socket; use Socket ();
0.270s 0.231s

POSIX exports a lot of symbols by default. If you tell it to export none, it starts in 30% less time. Socket starts in 15% less time.
regexps

avoid $&
The $& variable returns the last text successfully matched in any regular expression. It's not lexically scoped, so unlike the match variables $1 etc it isn't reset when you leave a block. This means that to be correct perl has to keep track of it from any match, as perl has no idea when it might be needed. As it involves taking a copy of the matched string, it's expensive for perl to keep track of. If you never mention $&, then perl knows it can cheat and never store it. But if you (or any module) mentions $& anywhere then perl has to keep track of it throughout the script, which slows things down. So it's a good idea to capture the whole match explicitly if that's what you need.

$text =~ /.* rules/;
$line = $&; # Now every match will copy $& - slow

$text =~ /(.* rules)/;
$line = $1; # Didn't mention $& - fast

avoid use English;
use English gives helpful long names to all the punctuation variables. Unfortunately that includes aliasing $& to $MATCH which makes perl think that it needs to copy every match into $&, even if you script never actually uses it. In perl 5.8 you can say use English '-no_match_vars'; to avoid mentioning the naughty "word", but this isn't available in earlier versions of perl.
avoid needless captures
Are you using parentheses for capturing, or just for grouping? Capturing involves perl copying the matched string into $1 etc, so it all you need is grouping use a the non-capturing (?:...) instead of the capturing (...).
/.../o;
If you define scalars with building blocks for your regexps, and then make your final regexp by interpolating them, then your final regexp isn't going to change. However, perl doesn't realise this, because it sees that there are interpolated scalars each time it meets your regexp, and has no idea that their contents are the same as before. If your regexp doesn't change, then use the /o flag to tell perl, and it will never waste time checking or recompiling it.
but don't blow it
You can use the qr// operator to pre-compile your regexps. It often is the easiest way to write regexp components to build up more complex regexps. Using it to build your regexps once is a good idea. But don't screw up (like parrot's assemble.pl did) by telling perl to recompile the same regexp every time you enter a subroutine:

sub foo {
my $reg1 = qr/.../;
my $reg2 = qr/... $reg1 .../;

You should pull those two regexp definitions out of the subroutine into package variables, or file scoped lexicals.

Devel::DProf

You find what is slow by using a profiler. People often guess where they think their program is slow, and get it hopelessly wrong. Use a profiler.

Devel::DProf is in the perl core from version 5.6. If you're using an earlier perl you can get it from CPAN.

You run your program with -d:DProf

perl5.8.0 -d:DProf enc2xs.orig -Q -O -o /dev/null ...

which times things and stores the data in a file named tmon.out. Then you run dprofpp to process the tmon.out file, and produce meaningful summary information. This excerpt is the default length and format, but you can use options to change things - see the man page. It also seems to show up a minor bug in dprofpp, because it manages to total things up to get 106%. While that's not right, it doesn't affect the explanation.

Total Elapsed Time = 66.85123 Seconds
User+System Time = 62.35543 Seconds
Exclusive Times
%Time ExclSec CumulS #Calls sec/call Csec/c Name
106. 66.70 102.59 218881 0.0003 0.0005 main::enter
49.5 30.86 91.767 6 5.1443 15.294 main::compile_ucm
19.2 12.01 8.333 45242 0.0003 0.0002 main::encode_U
4.74 2.953 1.078 45242 0.0001 0.0000 utf8::unicode_to_native
4.16 2.595 0.718 45242 0.0001 0.0000 utf8::encode
0.09 0.055 0.054 5 0.0109 0.0108 main::BEGIN
0.01 0.008 0.008 1 0.0078 0.0078 Getopt::Std::getopts
0.00 0.000 -0.000 1 0.0000 - Exporter::import
0.00 0.000 -0.000 3 0.0000 - strict::bits
0.00 0.000 -0.000 1 0.0000 - strict::import
0.00 0.000 -0.000 2 0.0000 - strict::unimport

At the top of the list, the subroutine enter takes about half the total CPU time, with 200,000 calls, each very fast. That makes it a good candidate to optimise, because all you have to do is make a slight change that gives a small speedup, and that gain will be magnified 200,000 times. [It turned out that enter was tail recursive, and part of the speed gain I got was by making it loop instead]

Third on the list is encode_U, which with 45,000 calls is similar, and worth looking at. [Actually, it was trivial code and in the real enc2xs I inlined it]

utf8::unicode_to_native and utf8::encode are built-ins, so you won't be able to change that.

Don't bother below there, as you've accounted for 90% of total program time, so even if you did a perfect job on everything else, you could only make the program run 10% faster.

compile_ucm is trickier - it's only called 6 times, so it's not obvious where to look for what's slow. Maybe there's a loop with many iterations. But now you're guessing, which isn't good.

One trick is to break it into several subroutines, just for benchmarking, so that DProf gives you times for different bits. That way you can see where the juicy bits to optimise are.

Devel::SmallProf should do line by line profiling, but every time I use it it seems to crash.
Benchmark

Now you've identified the slow spots, you need to try alternative code to see if you can find something faster. The Benchmark module makes this easy. A particularly good subroutine is cmpthese, which takes code snippets and plots a chart. cmpthese was added to Benchmark with perl 5.6.

So to compare two code snippets orig and new by running each for 10000 times you'd do this:

use Benchmark ':all';

sub orig {
...
}

sub new {
...
}

cmpthese (10000, { orig => \&orig, new => \&new } );

Benchmark runs both, times them, and then prints out a helpful comparison chart:

Benchmark: timing 10000 iterations of new, orig...
new: 1 wallclock secs ( 0.70 usr + 0.00 sys = 0.70 CPU) @ 14222.22/s (n=10000)
orig: 4 wallclock secs ( 3.94 usr + 0.00 sys = 3.94 CPU) @ 2539.68/s (n=10000)
Rate orig new
orig 2540/s -- -82%
new 14222/s 460% --

and it's plain to see that my new code is over 4 times as fast as my original code.
What causes slowness in perl?

Actually, I didn't tell the whole truth earlier about what causes slowness in perl. [And astute hecklers such as Philip Newton had already told me this]

When perl compilers your program it breaks it down into a sequence of operations it must perform, which are usually referred to as ops. So when you ask perl to compute $a = $b + $c it actually breaks it down into these ops:

* Fetch $b onto the stack
* Fetch $c onto the stack
* Add the top two things on the stack together; write the result to the stack
* Fetch the address of $a
* Place the thing on the top of stack into that address

Computers are fast at simple things like addition. But there is quite a lot of overhead involved in keeping track of "which op am I currently performing" and "where is the next op", and this book-keeping often swamps the time taken to actually run the ops. So often in perl it's the number of ops your program takes to perform its task that is more important than the CPU they use or the RAM it needs. The hit list is

1. Ops
2. CPU
3. RAM

So what were my example code snippets that I Benchmarked?

It was code to split a line of hex (54726164696e67207374796c652f6d61) into groups of 4 digits (5472 6164 696e ...) , and convert each to a number

sub orig {
map {hex $_} $line =~ /(....)/g;
}

sub new {
unpack "n*", pack "H*", $line;
}

The two produce the same results:
orig new
21618, 24932, 26990, 26400, 29556, 31084, 25903, 28001, 26990, 29793, 26990, 24930, 26988, 26996, 31008, 26223, 29216, 29552, 25957, 25646 21618, 24932, 26990, 26400, 29556, 31084, 25903, 28001, 26990, 29793, 26990, 24930, 26988, 26996, 31008, 26223, 29216, 29552, 25957, 25646

but the first one is much slower. Why? Following the data path from right to left, it starts well with a global regexp, which is only one op and therefore a fast way to generate a list of the 4 digit groups. But that map block is actually an implicit loop, so for each 4 digit block it iterates round and repeatedly calls hex. Thats at least one op for every list item.

Whereas the second one has no loops in it, implicit or explicit. It uses one pack to convert the hex temporarily into a binary string, and then one unpack to convert that string into a list of numbers. n is big endian 16 bit quantities. I didn't know that - I had to look it up. But when the profiler told me that this part of the original code was a performance bottleneck, the first think that I did was to look at the the pack docs to see if I could use some sort of pack/unpack as a speedier replacement.
Ops are bad, m'kay

You can ask perl to tell you the ops that it generates for particular code with the Terse backend to the compiler. For example, here's a 1 liner to show the ops in the original code:

$ perl -MO=Terse -e'map {hex $_} $line =~ /(....)/g;'

LISTOP (0x16d9c8) leave [1]
OP (0x16d9f0) enter
COP (0x16d988) nextstate
LOGOP (0x16d940) mapwhile [2]
LISTOP (0x16d8f8) mapstart
OP (0x16d920) pushmark
UNOP (0x16d968) null
UNOP (0x16d7e0) null
LISTOP (0x115370) scope
OP (0x16bb40) null [174]
UNOP (0x16d6e0) hex [1]
UNOP (0x16d6c0) null [15]
SVOP (0x10e6b8) gvsv GV (0xf4224) *_
PMOP (0x114b28) match /(....)/
UNOP (0x16d7b0) null [15]
SVOP (0x16d700) gvsv GV (0x111f10) *line

At the bottom you can see how the match /(....)/ is just one op. But the next diagonal line of ops from mapwhile down to the match are all the ops that make up the map. Lots of them. And they get run each time round map's loop. [Note also that the {}s mean that map enters scope each time round the loop. That not a trivially cheap op either]

Whereas my replacement code looks like this:

$ perl -MO=Terse -e'unpack "n*", pack "H*", $line;'

LISTOP (0x16d818) leave [1]
OP (0x16d840) enter
COP (0x16bb40) nextstate
LISTOP (0x16d7d0) unpack
OP (0x16d7f8) null [3]
SVOP (0x10e6b8) const PV (0x111f94) "n*"
LISTOP (0x115370) pack [1]
OP (0x16d7b0) pushmark
SVOP (0x16d6c0) const PV (0x111f10) "H*"
UNOP (0x16d790) null [15]
SVOP (0x16d6e0) gvsv GV (0x111f34) *line

There are less ops in total. And no loops, so all the ops you see execute only once. :-)

[My helpful hecklers pointed out that it's hard to work out what an op is. Good call. There's roughly one op per symbol (function, operator, variable name, and any other bit of perl syntax). So if you golf down the number of functions and operators your program runs, then you'll be reducing the number of ops.]

[These were supposed to be the bonus slides. I talked to fast (quelle surprise) and so manage to actually get through the lot with time for questions]
Memoize

Caches function results
MJD's Memoize follows the grand perl tradition by trading memory for speed. You tell Memoize the name(s) of functions you'd like to speed up, and it does symbol table games to transparently intercept calls to them. It looks at the parameters the function was called with, and uses them to decide what to do next. If it hasn't seen a particular set of parameters before, it calls the original function with the parameters. However, before returning the result, it stores it in a hash for that function, keyed by the function's parameters. If it has seen the parameters before, then it just returns the result direct from the hash, without even bothering to call the function.
For functions that only calculate
This is useful for functions that calculate things with no side effects, slow functions that you often call repeatedly with the same parameters. It's not useful for functions that do things external to the program (such as generating output), nor is it good for very small, fast functions.
Can tie cache to a disk file
The hash Memoize uses is a regular perl hash. This means that you can tie the hash to a disk file. This allows Memoize to remember things across runs of your program. That way, you could use Memoize in a CGI to cache static content that you only generate on demand (but remember you'll need file locking). The first person who requests something has to wait for the generation routine, but everyone else gets it straight from the cache. You can also arrange for another program to periodically expire results from the cache.

As of 5.8 Memoize module has been assimilated into the core. Users of earlier perl can get it from CPAN.
Miscellaneous

These are quite general ideas for optimisation that aren't particularly perl specific.

Pull things out of loops
perl's hash lookups are fast. But they aren't as fast as a lexical variable. enc2xs was calling a function each time round a loop based on a hash lookup using $type as the key. The value of $type didn't change, so I pulled the lookup out above the loop into a lexical variable:

my $type_func = $encode_types{$type};

and doing it only once was faster.
Experiment with number of arguments
Something else I found was that enc2xs was calling a function which took several arguments from a small number of places. The function contained code to set defaults if some of the arguments were not supplied. I found that the way the program ran, most of the calls passed in all the values and didn't need the defaults. Changing the function to not set defaults, and writing those defaults out explicitly where needed bought me a speed up.
Tail recursion
Tail recursion is where the last thing a function does it call itself again with slightly different arguments. It's a common idiom, and some languages can automatically optimise it away. Perl is not one of those languages. So every time a function tail recurses you have another subroutine call [not cheap - Arthur Bergman notes that it is 10 pages of C source, and will blow the instruction cache on a CPU] and re-entering that subroutine again causes more memory to be allocated to store a new set of lexical variables [also not cheap].

perl can't spot that it could just throw away the old lexicals and re-use their space, but you can, so you can save CPU and RAM by re-writing your tail recursive subroutines with loops. In general, trying to reduce recursion by replacing it with iterative algorithms should speed things up.

yay for y

y, or tr, is the transliteration operator. It's not as powerful as the general purpose regular expression engine, but for the things it can do it is often faster.

tr/!// # fastest way to count chars
tr doesn't delete characters unless you use the /d flag. If you don't even have any replacement characters then it treats its target as read only. In scalar context it returns the number of characters that matched. It's the fastest way to count the number of occurrences of single characters and character ranges. (ie it's faster than counting the elements returned by m/.../g in list context. But if you just want to see whether one or more of a character is present use m/.../, because it will stop at the u first, whereas tr/// has to go to the end)
tr/q/Q/ faster than s/q/Q/g
tr is also faster than the regexp engine for doing character-for-character substitutions.
tr/a-z//d faster than s/[a-z]//g
tr is faster than the regexp engines for doing character range deletions. [When writing the slide I assumed that it would be faster for single character deletions, but I Benchmarked things and found that s///g was faster for them. So never guess timings; always test things. You'll be surprised, but that's better than being wrong]

Ops are bad, m'kay

Another example lifted straight from enc2xs of something that I managed to accelerate quite a bit by reducing the number of ops run. The code takes a scalar, and prints out each byte as \x followed by 2 digits of hex, as it's generating C source code:

#foreach my $c (split(//,$out_bytes)) {
# $s .= sprintf "\\x%02X",ord($c);
#}
# 9.5% faster changing that loop to this:
$s .= sprintf +("\\x%02X" x length $out_bytes), unpack "C*", $out_bytes;

The original makes a temporary list with split [not bad in itself - ops are more important than CPU or RAM] and then loops over it. Each time round the loop it executes several ops, including using ord to convert the byte to its numeric value, and then using sprintf with the format "\\x%02X" to convert that number to the C source.

The new code effectively merges the split and looped ord into one op, using unpack's C format to generate the list of numeric values directly. The more interesting (arguably sick) part is the format to sprintf, which is inside +(...). You can see from the .= in the original that the code is just concatenating the converted form of each byte together. So instead of making sprintf convert each value in turn, only for perl ops to stick them together, I use x to replicate the per-byte format string once for each byte I'm about to convert. There's now one "\\x%02X" for each of the numbers in the list passed from unpack to sprintf, so sprintf just does what it's told. And sprintf is faster than perl ops.
How to make perl fast enough

use the language's fast features
You have enormous power at your disposal with regexps, pack, unpack and sprintf. So why not use them?

All the pack and unpack code is implemented in pure C, so doesn't have any of the book-keeping overhead of perl ops. sprintf too is pure C, so it's fast. The regexp engine uses its own private bytecode, but it's specially tuned for regexps, so it runs much faster than general perl code. And the implementation of tr has less to do than the regexp engine, so it's faster.

For maximum power, remember that you can generate regexps and the formats for pack, unpack and sprintf at run time, based on your data.
give the interpreter hints
Make it obvious to the interpreter what you're up to. Avoid $&, use (?:...) when you don't need capturing, and put the /o flag on constant regexps.
less OPs
Try to accomplish your tasks using less operations. If you find you have to optimise an existing program then this is where to start - golf is good, but remember it's run time strokes not source code strokes.
less CPU
Usually you want to find ways of using less CPU.
less RAM
but don't forget to think about how your data structures work to see if you can make them use less RAM.

© 2002 Nicholas Clark

11 comments:

Anonymous said...

Кстати, не зря blogonews называется местом сбора всего самого интересного. [URL=http://blogonews.net]Интересно почитать[/URL] - безусловно правильное название для этого блога.

Anonymous said...

singular the briny guaranteed to insist upon you be luminary of measureless! Our all tangible commingle of herbs and aminos is Dr. formulated and proven to urge vacation, base mentally maltreatment gudgeon and unchanging boost your wisecracks!


[url=http://minichill.com]Energy Drink[/url]
[url=http://minichill.com]Energy Drinks[/url]

[url=http://minichill.com/]Relaxzen[/url]
[url=http://minichill.com/]Relaxzen drink[/url]
[url=http://minichill.com/]iChill review[/url]

[url=http://minichill.com/home/index.html]facebook[/url]
[url=https://www.getreversemortgagehelp.com/]american reverse mortgage[/url]


[url=http://minichill.com/5%20hour%20energy.html]5 Hour Energy drink[/url]


[url=http://minichill.com/5%20hour%20energy.html]best energy drink[/url]
[url=http://minichill.com/lab/relaxation/index.html]Valerian Root[/url]
[url=http://minichill.com/lab/happiness/index.html]Valerian Root[/url]
[url=http://minichill.com/lab/focus/index.html]L-Theanine[/url]
[url=http://minichill.com/lab/anti-anxiety/index.htmll]GABA[/url]
[url=http://minichill.com/lab/anti-anxiety/index.htmll]Gamma Aminobutyric Acid [/url]


titanic Julian with Nutriment and D you can unexploded a well-behaved life. I proposition unqualifiedly, it's awesome so be brought up on and ry it, do it in this prime!
Mini Chill? contains a sound mingling of herbs and amino acids called Relarian?, that has been proven, in published clinical trials not solely to expendable to impart duel inflection and the hots, but to in actuality recondition your heavens and harden in shift request pinpoint! Mini Aloofness doesn?t fabricator drowsiness, so whether you?re in the bull's-eye of a stressful date at in the planning stages unemployed or enjoying a lifetime free with your friends, Mini Chill? is guaranteed to adjust your day.



[url=http://minichill.com/ChillRecipes.html]alcohol graphics[/url]
[url=http://routeworldbrokers.com/]NY route opportunities[/url]
[url=http://routeworldbrokers.com/]NY route opportunities[/url]
[url=http://minichill.com/ChillRecipes.html]alcohol abuse treatment[/url]


[url=http://www.finmedsys.com/]medical billing company[/url]
[url=http://www.finmedsys.com/]medical billing outsourcing[/url]
[url=http://www.finmedsys.com/]medical billing services[/url]

[url=http://minichill.com/lab/relaxation/index.html]relax gift[/url]

[url=http://minichill.com/lab/relaxation/index.html]can't relax[/url]
[url=http://minichill.com/ChillRecipes.html]marijuana brownies[/url]
[url=http://minichill.com/ChillRecipes.html]alcohol rehabilitation[/url]

Anonymous said...

well guys! coincide the latest unbind [url=http://www.casinolasvegass.com]casino[/url] games like roulette and slots !corroborate into public notice like a sparkle the all uncharted unburden [url=http://www.casinolasvegass.com]online casino[/url] games at the all stylish www.casinolasvegass.com, the most trusted [url=http://www.casinolasvegass.com]online casinos[/url] on the cobweb! profit from our [url=http://www.casinolasvegass.com/download.html]free casino software download[/url] and finish first in money.
you can also brake other [url=http://sites.google.com/site/onlinecasinogames2010/]online casinos bonus[/url] . you should also check this [url=http://www.realcazinoz.com/fr]Casino en ligne[/url], [url=http://www.realcazinoz.com/it]Casino Online[/url] and [url=http://www.realcazinoz.com/es]casino en linea[/url] games. join the the largest [url=http://www.texasholdem-online-poker.com/]online poker[/url] room. check this new [url=http://www.realcazinoz.com/paypalcasino.htm]paypal casino[/url]. [url=http://www.ttittancasino.com]Online Casino Spiele[/url] , buy [url=http://www.web-house.co.il/acai-berry.htm]acai berry[/url] . [url=http://www.avi.vg/search2.php?a=sex4sexx&ser_key=bondage+]bondage[/url] [url=http://www.thecasino.co.il/ilcasino.htm]casino[/url] . [url=http://en.gravatar.com/willinger18]online casino games[/url] , [url=http://www.web-house.co.il/buy-k2.htm]Buy k2[/url] and new [url=http://casino-online.wikispaces.com/Online+Casino+Games]online casino[/url]

Anonymous said...

Sweet web site, I had not noticed bsjpark.blogspot.com earlier during my searches!
Carry on the fantastic work!

Anonymous said...

In the current finance wording, your cash advance loans happen to be becoming a lot more famous. Which is principally as all these [url=http://myfastloans.co.uk]short term loans 5000[/url] mortgages assistance visitors to handle unanticipated economic emergencies. When you even need to deal with finances troubles, it's fundamental for you to really think no matter whether any such financial loan symbolizes the best option for you [url=http://cashloans-247.co.uk]same day loans self employed[/url] or simply never. That allows you to bring the most appropriate selection, you will need to understand a couple of very important specifics that always depend on the actual cash advance loans over the internet.

That cash loans through 60 minutes ordinarily send out a small degree, that fails to [url=http://myfastloans.co.uk]fast loans[/url] extend past £1, 500. Hence, should you need well over £1, 500, it is best to certainly contemplate a different loan solution. Or, your skill can be to fully grasp this personal loan and likewise receive money via relatives or even associates. One more standard dilemma might be the belief that any cash loans happen to be given to get a targeted stage, which specifically ceases if you're going to get your salary.

Likewise, it is rather vital for know unless you use a continuous job, you may not find any of the online payday loan other options. The main reason for this is exactly due to the fact not often covered show a warranty, that in this case need to be a restricted salary. Subsequently, here, the lending company will not take on the job particularly given that she should not be confident of the fact that you may refund the actual personal loan. Still, in case you have a gentle position and also the mortgage company approves people program, you might have [url=http://instantcashloans4u.co.uk]instantcashloans4u.co.uk[/url] the cash transferred in your bank account in pertaining to 60 minutes.

Another matters you must bear in mind are classified as the interest rates and extra premiums, ones own credit ranking and also the words and phrases from the lending product. By contemplating each one of [url=http://smallloans247.co.uk]http://smallloans247.co.uk[/url] components, you can actually choose your house offered payday advances on the internet are generally created for anyone or in no way.

Anonymous said...

payday loans online http://2applyforcash.com/ alcosycle Payday Loans Online boype [url=http://2applyforcash.com]Payday Loans Online[/url] Online Payday Loan People string is a social networking site methods are paid you''.

Anonymous said...

disclosure on the network two successful involved groups behind manipulation and concise with each other to do [url=http://www.ddtshanghaiescort.com/shanghai-escort.html]shanghai massage[/url] more it

Anonymous said...

Hello. And Bye. Thank you very much.

Whelen Troy said...

Cool! im blog walking, please blog walk my site: www.projectmanager.com. thanks!

Anonymous said...

na6gj9 wartrol customer reviews iw5hz3 does wartrol work ns3xd8 wartrol ws2ke6 buy wartrol qt9oa9 buy wartrol

Anonymous said...

Hеllo eveгуone, it's my first pay a visit at this web site, and post is really fruitful designed for me, keep up posting such articles or reviews.

Look into my web-site: payday loans