I suspect that your cpu usage problem is related to this failure when running the Thread::Pool::Simple test suite:

c:\Perl\packages\Thread-Pool-Simple-0.23>nmake test

Microsoft (R) Program Maintenance Utility Version 7.00.9466
Copyright (C) Microsoft Corporation.  All rights reserved.

        C:\Perl\bin\perl.exe "-MExtUtils::Command::MM" "-e" "test_harn
+ess(0, 'blib\lib', 'blib\arch')" t\1.t t\2.t t\3.t t\4.t
t\1....ok 1/0Can't call method "down" on unblessed reference at C:\Per
+l\packages\Thread-Pool-Simple-0.23\blib\lib/Thread/Pool/Simple.pm lin
+e 39.
# Looks like your test died just after 1.
t\1....dubious
        Test returned status 255 (wstat 65280, 0xff00)
        after all the subtests completed successfully
t\2....ok 1/0Can't call method "down" on unblessed reference at C:\Per
+l\packages\Thread-Pool-Simple-0.23\blib\lib/Thread/Pool/Simple.pm lin
+e 39.
# Looks like your test died just after 1.
t\2....dubious
        Test returned status 255 (wstat 65280, 0xff00)
        after all the subtests completed successfully
t\3....ok 1/0Can't call method "down" on unblessed reference at C:\Per
+l\packages\Thread-Pool-Simple-0.23\blib\lib/Thread/Pool/Simple.pm lin
+e 39.
# Looks like your test died just after 1.
t\3....dubious
        Test returned status 255 (wstat 65280, 0xff00)
        after all the subtests completed successfully
t\4....ok 1/0Can't call method "down" on unblessed reference at C:\Per
+l\packages\Thread-Pool-Simple-0.23\blib\lib/Thread/Pool/Simple.pm lin
+e 39.
# Looks like your test died just after 1.
t\4....dubious
        Test returned status 255 (wstat 65280, 0xff00)
        after all the subtests completed successfully
Failed Test Stat Wstat Total Fail  Failed  List of Failed
----------------------------------------------------------------------
+---------
t\1.t        255 65280     1    0   0.00%  ??
t\2.t        255 65280     1    0   0.00%  ??
t\3.t        255 65280     1    0   0.00%  ??
t\4.t        255 65280     1    0   0.00%  ??
Failed 4/4 test scripts, 0.00% okay. 0/4 subtests failed, 100.00% okay
+.
NMAKE : fatal error U1077: 'C:\Perl\bin\perl.exe' : return code '0x2'
Stop.
[download]

This failure can be traced to this line from the T::P::S code:

    my $self = &share({});
[download]

Which make the T:::P::S object a shared hash. The problem is that when a few lines later, it creates an instance of Thread::Semaphore and stores the returned handle into that hash:

    $self->{shutdown_lock} = Thread::Semaphore->new();
[download]

Because the hash is shared, the blessedness of the Thread::Semaphore object is lost. Ie. It stops being a blessed scaler and becomes a plain ordinary scalar. And when it attempts to call a methd on that scalar:

    $self->{shutdown_lock}->down(); ## line 39
[download]

it fails (Note:The line number is different because I added some debug:

???SCALAR(0x19a690c)
Can't call method "down" on unblessed reference at 
C:\Perl\packages\Thread-Pool-Simple-0.23\blib\lib/Thread/Pool/Simple.p
+m 
line 43.
[download]

This is a newly introduced bug that didn't exist when I tried T::P::Simple a few few weeks ago. Now, you can "fix" this test failure by upgrading to the latest CPAN versions of threads and threads::shared.

Unfortunately, with that upgrade, as with so many of the releases ever since they were dual lived, comes a different set of problems. The T::P::Simple now completes all its tests, but leaks memory like a sieve:

c:\Perl\packages\Thread-Pool-Simple-0.23>nmake test

Microsoft (R) Program Maintenance Utility Version 7.00.9466
Copyright (C) Microsoft Corporation.  All rights reserved.

        C:\Perl\bin\perl.exe "-MExtUtils::Command::MM" "-e" "test_harn
+ess(0, 'blib\lib', 'blib\arch')" t\1.t t\2.t t\3.t t\4.
t\1....ok 1/0Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
t\1....ok
t\2....ok 901/0Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
t\2....ok
t\3....ok 901/0Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
Scalars leaked: 1
t\3....ok
t\4....ok 2/0Scalars leaked: 1
t\4....ok
All tests successful.
Files=4, Tests=1805, 20 wallclock secs ( 0.00 cusr +  0.00 csys =  0.0
+0 CPU)
[download]

Which wouldn't be so bad if T::P::Simple was the only module affected, but with the latest versions of threads & threads::shared, programs that ran completely clean with 5.8.6 and its distributed version of threads/threads::shared, now leak like a sieve also..