Did you modify the test to let it go up to \(2^{15}\) or something? The test currently stops testing Po2 sizes at \(2^{14}\). (And of course \(2^{14} = 16384\).)kalle2990 wrote:This is what I got with my Radeon HD 5870:Result wrote:Framebuffers
Non-Po2 Framebuffers
Non-Po2 Images
Framebuffers up to the size 32768 x 32768
Power of 3 Framebuffers up to the size 19683 x 19683
Survey: Do You Support Framebuffers and/or Non-Po2 [RERUN]
- BlackBulletIV
- Inner party member
- Posts: 1261
- Joined: Wed Dec 29, 2010 8:19 pm
- Location: Queensland, Australia
- Contact:
Re: Survey: Do You Support Framebuffers and/or Non-Po2 [RERU
Thanks for the results guys.
Re: Survey: Do You Support Framebuffers and/or Non-Po2 [RERU
If I read this code correctly, it will multiply the result once more after each successful test, leaving it at \(2^{15}\).BlackBulletIV wrote:Did you modify the test to let it go up to \(2^{15}\) or something? The test currently stops testing Po2 sizes at \(2^{14}\). (And of course \(2^{14} = 16384\).)
Code: Select all
for i = 1, 14 do
status, fb = pcall(love.graphics.newFramebuffer, maxSize, maxSize)
if not status or type(fb) ~= 'userdata' then break end
maxSize = maxSize * 2
end
Code: Select all
if status and type(fb) == 'userdata' then
maxSize = maxSize / 2
end
- BlackBulletIV
- Inner party member
- Posts: 1261
- Joined: Wed Dec 29, 2010 8:19 pm
- Location: Queensland, Australia
- Contact:
Re: Survey: Do You Support Framebuffers and/or Non-Po2 [RERU
Ooooh, I see, that's because I initially start at \(2^1\). Ok everyone, corrections. The maximum possible values are:
Po2: \(2^{15} = 32,768\)
Po3: \(3^{15} = 14,348,907\)
And I don't think anyone's going to get 14m, ha ha!
Po2: \(2^{15} = 32,768\)
Po3: \(3^{15} = 14,348,907\)
And I don't think anyone's going to get 14m, ha ha!
- Jasoco
- Inner party member
- Posts: 3726
- Joined: Mon Jun 22, 2009 9:35 am
- Location: Pennsylvania, USA
- Contact:
Re: Survey: Do You Support Framebuffers and/or Non-Po2 [RERU
Would anyone NEED 14m? Would anyone NEED 32t? lolBlackBulletIV wrote:Ooooh, I see, that's because I initially start at \(2^1\). Ok everyone, corrections. The maximum possible values are:
Po2: \(2^{15} = 32,768\)
Po3: \(3^{15} = 14,348,907\)
And I don't think anyone's going to get 14m, ha ha!
- BlackBulletIV
- Inner party member
- Posts: 1261
- Joined: Wed Dec 29, 2010 8:19 pm
- Location: Queensland, Australia
- Contact:
Re: Survey: Do You Support Framebuffers and/or Non-Po2 [RERU
Nope. But hey, we're just testing. Although I can think of some use for 32K, huuuge composited backgrounds.
- Jasoco
- Inner party member
- Posts: 3726
- Joined: Mon Jun 22, 2009 9:35 am
- Location: Pennsylvania, USA
- Contact:
Re: Survey: Do You Support Framebuffers and/or Non-Po2 [RERU
Do you know how low your framerate would be and how much faster it would be to chop it up into smaller pieces? No one will ever need a 32,000 sized image in a Löve game. Especially since it would only work for the smallest amount of people.BlackBulletIV wrote:Nope. But hey, we're just testing. Although I can think of some use for 32K, huuuge composited backgrounds.
- BlackBulletIV
- Inner party member
- Posts: 1261
- Joined: Wed Dec 29, 2010 8:19 pm
- Location: Queensland, Australia
- Contact:
Re: Survey: Do You Support Framebuffers and/or Non-Po2 [RERU
In fact I didn't know, lol. But no, 32K is overkill (let's not even go into 14M).
Wow, someone voted only framebuffers! That is odd. Well, at least I know that getting no image is possible using my ImageData technique.
Wow, someone voted only framebuffers! That is odd. Well, at least I know that getting no image is possible using my ImageData technique.
Re: Survey: Do You Support Framebuffers and/or Non-Po2 [RERU
Jasoco wrote: Would anyone NEED 14m? Would anyone NEED 32t? lol
Bill Gates wrote:640 k ought to be enough for anyone
Crazy Little Thing Called LÖVE.
Re: Survey: Do You Support Framebuffers and/or Non-Po2 [RERU
My results:
My Development Diary - http://shanegadsby.info
Re: Survey: Do You Support Framebuffers and/or Non-Po2 [RERU
I just saw I made a little mistake when writing the additional code for the fix..
This:
Should be:
This will lead to:
This:
Code: Select all
if status and type(fb) == 'userdata' then
maxSize = maxSize / 2
end
Code: Select all
maxSize = maxSize / 2
Code: Select all
maxSize is 2
Testing 2 x 2...
Success!
Change maxSize to 4
.....
Testing 16384 x 16384...
Success!
Change maxSize to 32768
Test completed, revert: Change maxSize to 16384
or
Testing 16384 x 16384...
Failure!
Test aborted, revert to last working: Change maxSize to 8192
Who is online
Users browsing this forum: No registered users and 0 guests