[DiscordArchive] Actually yeah, if you don't mind the lack of percision is there really anything wrong with storing 8
[DiscordArchive] Actually yeah, if you don't mind the lack of percision is there really anything wrong with storing 8
Original source
Actually yeah, if you don't mind the lack of percision is there really anything wrong with storing 8bit numbers for random loot and squaring them?
rektbyfaith
06-20-2022, 06:53 PM #1Archived author: whatwere • Posted: 2022-06-20T18:53:57.319000+00:00
Original source
Actually yeah, if you don't mind the lack of percision is there really anything wrong with storing 8bit numbers for random loot and squaring them?
Original source
no, but you should measure before you do optimizations like that
rektbyfaith
06-20-2022, 06:54 PM #2Archived author: <o> • Posted: 2022-06-20T18:54:18.193000+00:00
Original source
no, but you should measure before you do optimizations like that
Original source
if you don't measure, you could spend 90% of your optimization energy on things that don't matter at all
rektbyfaith
06-20-2022, 06:54 PM #3Archived author: <o> • Posted: 2022-06-20T18:54:39.760000+00:00
Original source
if you don't measure, you could spend 90% of your optimization energy on things that don't matter at all
Original source
255^2 is equalish to 65k, highest 16 bit number so you're not saving that much I guess.
rektbyfaith
06-20-2022, 06:56 PM #4Archived author: whatwere • Posted: 2022-06-20T18:56:08.150000+00:00
Original source
255^2 is equalish to 65k, highest 16 bit number so you're not saving that much I guess.
Original source
generally, start out sketching out what you actually want to be able to do, then try to figure out the best way you can do that without over-engineering, take some limit checks, study similar solutions, write something sane to the best of your ability and measure
rektbyfaith
06-20-2022, 06:59 PM #5Archived author: <o> • Posted: 2022-06-20T18:59:46.883000+00:00
Original source
generally, start out sketching out what you actually want to be able to do, then try to figure out the best way you can do that without over-engineering, take some limit checks, study similar solutions, write something sane to the best of your ability and measure
Original source
you'll never be able to consistently write systems that don't need to be refactored later anyways, so it's futile to try and predict everything from the go
rektbyfaith
06-20-2022, 07:01 PM #6Archived author: <o> • Posted: 2022-06-20T19:01:23.070000+00:00
Original source
you'll never be able to consistently write systems that don't need to be refactored later anyways, so it's futile to try and predict everything from the go
1 Guest(s)
1 Guest(s)