I believe this problem is already fixed upstream. At least my test result in pmagic_2013_05_01_i586 is fine.
In case someone is interested:
I tested this on my laptop, Fujitsu Lifebook AH531
CPU Model: Intel(R) Core(TM) i3-2310M CPU @ 2.10GHz
(and burnt 2 CDs just for this...
Test 1: pmagic_2013_05_01_i686
Booting in Run-from-RAM mode, Firefox starts in less than 3 seconds.
Test 2: pmagic_2013_05_01_i686
Booting in Live mode, Firefox starts in about 15 seconds. CPU usage is less than 8%.
Test 3: pmagic_2013_05_01_i586
Booting in Live mode, Firefox starts in about 20 seconds. CPU usage is about 30%. (Given that my CPU has 2 physical cores and 4 threads, 30% for one processor should be reasonable.)
I also tested pmagic_2013_01_29_i586
included in UBCD, and Firefox indeed starts very slowly (and 100% CPU usage) in there (however I didn't count the time it starts).
So, it seems the problem is fixed already. But the story isn't over yet. Pmagic_2013_05_01_i586
has a regression that it cannot run from RAM
, so wait until Mr. Verner (PMagic maintainer) releases the next version of PMagic.