Optimization for math functions

BYOND Forums

Announcements · BYOND Help · Bug Reports · Feature Requests · Beta Testers · Beta Bugs · Developer Help · Design Philosophy · Demos & Libraries · Tutorials & Snippets · Art & Sound · Classified Ads · Game Updates · Contests & Events · Linux Talk · On Topic · Off Topic

ID:1305354

Jun 27 2013, 10:20 pm

Tobba

Resolved

Constant math expressions for several functions are now optimized by calculating them at compile-time.

Applies to:

DM Language

Status:

Resolved (499.1197)

This issue has been resolved.

Right now, something like 5*5 gets optimized to 25, but if you do 5*abs(5), it does not, since this optimization does not work for math functions

Support for this would be pretty nice

Jun 28 2013, 10:21 am (Edited on Jun 28 2013, 10:57 am)
Lummox JR	I'm working on the compiler as it happens, and I think this is pretty darn doable. [edit] I've added such optimization for sqrt(), abs(), sin(), cos(), arcsin(), arccos(), log(), and pow().

Jun 28 2013, 11:08 am In response to Lummox JR
Magnum2k	While you're at it, it would be nice if you could add additional trig functions as well, such as tan(), atan2(), atan(), and possibly even ~~floor()~~ and ceil(), though those aren't imperative as the others.

Jun 28 2013, 11:14 am
Nadrew	Can't you make every one of those functions yourself? ceil() especially, which is just round(num,1).

Jun 28 2013, 11:18 am
Magnum2k	Yes, you can. All of the aforementioned trig functions are dependent of one another, in one way or another; however, there's no excuse as to why they aren't built-in.

Jun 28 2013, 11:18 am
Tobba	tan is just sin(a)/cos(a), which the compiler can optimize now anyways Not sure about atan(2) though

Jun 28 2013, 11:24 am
Kaiochao	What does this mean, anyway? The compiler performs arithmetic and simplifies constants while compiling?

Jun 28 2013, 11:27 am
Lummox JR	Nadrew wrote: Can't you make every one of those functions yourself? ceil() especially, which is just round(num,1). ceil(x) is not round(x,1). round(x,1) rounds to the nearest whole number. ceil(x) is -round(-x), since round(x) is floor(x).

Jul 1 2013, 12:21 pm
Lummox JR	Lummox JR resolved issue with message: Constant math expressions for several functions are now optimized by calculating them at compile-time.

Jul 1 2013, 1:56 pm
Super Saiyan X	Lummox - by pow() you mean the internal pow therefore the ** operator, not some secret undocumented procedure, right?

Jul 1 2013, 6:57 pm In response to Super Saiyan X
Albro1	lol searching for potential redbook material.

Jul 2 2013, 8:05 am In response to Super Saiyan X
Lummox JR	Super Saiyan X wrote: Lummox - by pow() you mean the internal pow therefore the operator, not some secret undocumented procedure, right? Aye. The runtime error message still refers to it as pow(), but it is the operator.