The Perl Compiler
rurban - Reini Urban
cPanel, formerly Graz, Austria
See also a screencast of the YAPC talk at http://vimeo.com/14058377
rurban maintains cygwin perl since 5.8.8 and 3-4 modules, guts, B::* => 5.10, types, Ctypes.
Mostly doing Perl, LISP, C, bash and PHP.
Compiler was started 1995 by Malcom Beattie, abandoned 2007 by p5p, revived 2008 by me.
Very dynamic language: magic; tie; eval "require $foo;" -> which packages to import? Easy to compile, hard to make faster.
In the Perl Compiler suite B::C are three seperate compilers:
perl toke.c/op.c - B::C - perl op walker run.c
Eliminate the whole parsing and dynamic allocation time.
After compilation walk the "op tree" - run.c
int1. The op tree is not a "tree", it is reduced to a simple linked list of ops.
Every "op" (a pp_<opname> function) returns the next op.
2. PERL_ASYNC_CHECK was called after every single op, until 5.14.
Normal Perl functions start at INIT, after BEGIN and CHECK.
The O modules start at CHECK, and skip INIT.
The B::C compiler, invoked via O, freezes the state in CHECK, and invokes then the walker.
$ perl -MO=C,-omyprog.c -e'print $a;'
$ cc_harness -o myprog myprog.c
$ ./myprog
5.6.2, 5.8.9, 5.14 non-threaded B::C are quite usable and have the least known bugs, but the others became also pretty stable. 5.15 still has some XSLoader problems.
Best are in the following order: 5.14, 5.6.2, 5.8.9, 5.10, 5.12 non-threaded.
See testsuite and STATUS
Which software is compiler critical?
Which software is compiler critical?
Execution time is the same (sans B::CC)
Which software is compiler critical?
Execution time is the same (sans B::CC)
Startup time is radical faster
Which software is compiler critical?
Execution time is the same (sans B::CC)
Startup time is radical faster.
Web Apps with fast response times -
1 sec more or less => good or bad software
Which software is compiler critical?
Execution time is the same (sans B::CC)
Startup time is radical faster.
Web Apps with fast response times -
Optimise static initialization - strings and arrays
Optimise static initialization - strings and arrays
non-threaded ! +10-20% performance
ltrace reveils Gthr_key_ptr, gv_fetchpv, savepvn,
av_extend and safesysmalloc as major culprits, the later
three at startup-time.
Optimise static initialization - strings and arrays
non-threaded ! +10-20% performance
common constant strings with gcc -Os => automatically optimised
Optimise static initialization - strings and arrays
non-threaded ! +10-20% performance
common constant strings with gcc -Os => automatically optimised
av_extend - run-time malloc => static arrays ?
av_extend - run-time malloc => static arrays ?
static arrays are impossible if not Readonly
can not be extended at run-time, need to be realloc'ed into the heap.
But certain arrays can: -fro-inc (Readonly @INC), and compad names and symbols.
av_extend - run-time malloc => static arrays ?
pre-allocate faster with -fav-init or -O3 with independent_comalloc()
Same for hashes and strings (nyi).
cPanel has used B::C compiled 5.6 for a decade, and will switch to 5.14.
cPanel offers web hosting automation software that manages provider data, domains, emails, webspace. A typical large webapp. Perl startup time can be too slow for many AJAX calls which need fast initial response times.
mod_perl or pre-loaded perl-apps would help for startup-time, but not on this case, and not for reduzed memory size.
Larger code base => more significant startup improvements. Basically O(1)
Web Service Daemon
Resident Size (perl) 9756
Resident Size (perlcc) 9072
DNS Settings Client
Startup Time (perl) 0.074
Startup Time (perlcc) 0.021
HTML Template Processor
Startup Time (perl) 0.695
Startup Time (perlcc) 0.037
2011: Find and fix all remaining B::C bugs for 5.14.
2012: CC type and sub optimisations. use types.
run-time ops vs compile-time
BEGIN blocks only compile-time side-effects.
BEGIN {
use Package; # okay
chdir "dir"; # not okay.
# only done at compile-time, not at the user
print "stuff"; # okay, only at compile-time
eval "what"; # hmm; depends
} Move eval "require Package;" to BEGIN
run-time ops vs compile-time +
dynamic range 1..$foo
goto/next/last $label
Undetected modules behind eval "require":
use -uModule to enforce scanning these
user make test (via cpan):
45x (bytecode + c -O0 - O4 + cc -O0 - O2)
=> 8 min
author make test:
45x bytecode + c -O0 - O4 + cc -O0 - O2 (8 min)
modules.t top100 (16 min)
+ testcore.t (16 min)
=> ~40 min
author make test 40 min
for 5-10 perls (5.6, 5.8, 5.10, 5.12, 5.14 / threaded + non-threaded) 5*2=10
on 5 platforms (cygwin, debian, centos, solaris, freebsd)
=> 33 h (10*5*40 = 2000min) = 1-2 days, similar to the gcc testsuite.
top100 modules. See webpage or svn repo for results for all tested perls / modules
With 5.8 non-threaded 3 fails File::Temp B::Hooks::EndOfScope YAML
With blead debugging + threaded 27 fails
log.modules-5.010001:pass MooseX::Types #TODO generally
log.modules-5.012001-nt:fail MooseX::Types #TODO generally
log.modules-5.013003-nt:pass MooseX::Types #TODO generally
log.modules-5.013003d:fail MooseX::Types #TODO generally What can we statically leave out per pp_?
Now: arguments passing, return values for 50% ops
Planned: more + direct xsub calls.
Now: Unroll for known static types pp_opname completely into simple arithmetic.
Known static types at compile-time? User declarations or Devel::TypeCheck
Currently:
my $EnameE_i; IV integer
my $EnameE_ir; IV integer in a pseudo register
my $EnameE_d; NV double Future ideas are type qualifiers such as
my (int $foo, double $foo_d);
and attributes such as
my ($foo:Cint, $foo:Creg_int, $foo:Cdouble);