Comments on: Solving Project Euler Problems With Ioke http://www.peterkrantz.com/2009/project-euler-in-ioke/ A blog about technology, visualization, music and unmanned vehicle experiments Tue, 20 Jul 2010 05:03:09 +0000 hourly 1 http://wordpress.org/?v=3.0 By: Ola Bini http://www.peterkrantz.com/2009/project-euler-in-ioke/comment-page-1/#comment-2435 Ola Bini Sun, 04 Jan 2009 12:12:12 +0000 http://www.peterkrantz.com/?p=256#comment-2435 Hey, So, I've made one change on trunk so that the cost of creating all those numbers aren't as high. The reason being that originally, every time a number is in code, that will be translated into a internal:createNumber message. This message will run every time a number is encountered, and will as such create all those 11 to 20 numbers every time in the loop. It will also create quite a few copies of 0. This has been changed on trunk. Since numbers and texts are immutable in themselves, the first time it will be cached. The large difference I got was when running with --server, though. For long running processes, that is definitely recommend. Without my number-fixes, I got: 232792560 ruby p5.rb 8,28s user 0,07s system 95% cpu 8,786 total 232792560 ioke p5.ik 3224,51s user 54,80s system 96% cpu 56:46,65 total 232792560 ioke --server p5.ik 1076,38s user 23,17s system 99% cpu 18:20,49 total So as you can see, running with --server cuts the time down to a third. It's still 20 minutes running time, but much better than without it. I'm going to do a new run with the number caching and see what happens. Hey,

So, I’ve made one change on trunk so that the cost of creating all those numbers aren’t as high. The reason being that originally, every time a number is in code, that will be translated into a internal:createNumber message. This message will run every time a number is encountered, and will as such create all those 11 to 20 numbers every time in the loop. It will also create quite a few copies of 0. This has been changed on trunk. Since numbers and texts are immutable in themselves, the first time it will be cached.

The large difference I got was when running with –server, though. For long running processes, that is definitely recommend.

Without my number-fixes, I got:
232792560
ruby p5.rb 8,28s user 0,07s system 95% cpu 8,786 total
232792560
ioke p5.ik 3224,51s user 54,80s system 96% cpu 56:46,65 total
232792560
ioke –server p5.ik 1076,38s user 23,17s system 99% cpu 18:20,49 total

So as you can see, running with –server cuts the time down to a third. It’s still 20 minutes running time, but much better than without it.

I’m going to do a new run with the number caching and see what happens.

]]> By: Pete http://www.peterkrantz.com/2009/project-euler-in-ioke/comment-page-1/#comment-2433 Pete Sat, 03 Jan 2009 19:34:26 +0000 http://www.peterkrantz.com/?p=256#comment-2433 Brute force of #5 in Ruby: <code> n = 20 def is_divisible(n) (n % 20 == 0 and n % 19 == 0 and n % 18 == 0 and n % 17 == 0 and n % 16 == 0 and n % 15 == 0 and n % 14 == 0 and n % 13 == 0 and n % 12 == 0 and n % 11 == 0) end while !is_divisible(n) do n = n + 20 end print n </code> Brute force of #5 in Ruby:

DQpuID0gMjANCg0KZGVmIGlzX2RpdmlzaWJsZShuKQ0KCShuICUgMjAgPT0gMCBhbmQgDQoJbiAlIDE5ID09IDAgYW5kIA0KCW4gJSAxOCA9PSAwIGFuZA0KCW4gJSAxNyA9PSAwIGFuZA0KCW4gJSAxNiA9PSAwIGFuZA0KCW4gJSAxNSA9PSAwIGFuZA0KCW4gJSAxNCA9PSAwIGFuZA0KCW4gJSAxMyA9PSAwIGFuZA0KCW4gJSAxMiA9PSAwIGFuZA0KCW4gJSAxMSA9PSAwKQ0KZW5kDQoNCndoaWxlICFpc19kaXZpc2libGUobikgZG8NCiAgICBuID0gbiArIDIwDQplbmQNCg0KcHJpbnQgbg0K
]]> By: Pete http://www.peterkrantz.com/2009/project-euler-in-ioke/comment-page-1/#comment-2432 Pete Sat, 03 Jan 2009 19:25:39 +0000 http://www.peterkrantz.com/?p=256#comment-2432 The brute force version of #5: <code> n = 20 until(n % 20 == 0 and n % 19 == 0 and n % 18 == 0 and n % 17 == 0 and n % 16 == 0 and n % 15 == 0 and n % 14 == 0 and n % 13 == 0 and n % 12 == 0 and n % 11 == 0, n = n + 20 ) n println </code> The brute force version of #5:

DQpuID0gMjANCg0KdW50aWwobiAlIDIwID09IDAgYW5kIA0KICBuICUgMTkgPT0gMCBhbmQgDQogIG4gJSAxOCA9PSAwIGFuZA0KICBuICUgMTcgPT0gMCBhbmQNCiAgbiAlIDE2ID09IDAgYW5kDQogIG4gJSAxNSA9PSAwIGFuZA0KICBuICUgMTQgPT0gMCBhbmQNCiAgbiAlIDEzID09IDAgYW5kDQogIG4gJSAxMiA9PSAwIGFuZA0KICBuICUgMTEgPT0gMCwNCg0KICBuID0gbiArIDIwDQopDQoNCm4gcHJpbnRsbg0K
]]> By: Ola Bini http://www.peterkrantz.com/2009/project-euler-in-ioke/comment-page-1/#comment-2431 Ola Bini Sat, 03 Jan 2009 18:42:53 +0000 http://www.peterkrantz.com/?p=256#comment-2431 Cool, Could you post the brute force version of 5 too? I would be interested in seeing what takes so much time, since the looping is pretty fast. Although - I don't look too much at performance. =) Cool,

Could you post the brute force version of 5 too? I would be interested in seeing what takes so much time, since the looping is pretty fast. Although – I don’t look too much at performance. =)

]]>