glob2-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [glob2-devel] new gradient system


From: MOA333
Subject: Re: [glob2-devel] new gradient system
Date: Thu, 19 Aug 2004 22:52:46 +0000
User-agent: Mozilla Thunderbird 0.7.3 (X11/20040812)

Nuage wrote:

Summary: improved gradient speed by 97%, which improves overall speed by 5%.


i did some new tests:

- first of all you should realy do somethink with the checksum. i mean somethink to make the checksum at least 30 times faster. I suggest 100 times faster. Why not making the checksum once every 30 secconds, i dont know what this does exacty but i dont think that playing 10 secconds without checksum is bad, i can play for minutes witout checksum and the game is still working well.

- seccond, if i disable the checksum, i still have the gradian think that take more than 80% of resources and 110% of CPU on a 512*512 map; of course is better than 1000% of cpu and 99% of resources like before, but if you could just make it 10 or 20 times faster, maybe more, EVEN if the workers will stay more time before finding the right way;

- third, suppose that this two first problems get solved, do we realy nead to print so many frames on the screen? what about printing once on the screen every 3 cycles. It will be the same for the eye, but it will speed up the game, because when i disable both checksums and gradient the glx is taking the most of the time, well, still resonable;

- 4th, on huge maps printing the map is taking almost as much as printing the screen with glx. please, could you do somethink and print the map less often, like making a refresh of the map every 10 secconds or some other idea.

I am sure that the point 3 will make the game faster for little computers on maps up to 128*128;
 The first and seccond point could make the 512*512 game realy fast;
All 4 of this ill make the 512*512 and even maybe 1024*1024 maps fast even on less then 1 Ghz computers.

Only to find a way to desactivate the checksums and to make the gradient 10 or 20 times faster


 Is there a hope?






reply via email to

[Prev in Thread] Current Thread [Next in Thread]