timon botez wrote:
Do you mean that it wont be possible with current library? Whats the reason for not getting OpenGL acceleration when doing this? Would it be possible for my developer to write such addition based on the hardware spec? I appreciate your response.
the problem is not openGL-specifique. it is rather a problem with the hardware.
in those early days, when the first dual-head gamer's graphics-cards with hw-openGL-acceleration where sold, you would often get hw-acceleration only on the first head. so you couldn't really do dual-head gem (at least not with a reasonable performance). nowadays, dual-head cards normally (?) support hw-acceleration on both heads.
anyhow, i have no idea how it is with "many-head" cards (like 8 or 16) but i guess these cards will not have openGL-hw-acceleration at all (or, if they do, all heads will be supported, since i don't think that anybody would buy such a card, to only play doom on the first screen and watch the stock exchange indices on all the other monitors...)
but it all depends on the specs of the card.
probably it would be simpler to connect a bunch of computers (via netsend/netreceive) rather than have everything done on one machine.
and i guess, if you could get 4 or more multi-head geforce (radeon ?) cards (1 agp and 3 pci) to work in one computer (which *would* be possible, i guess; but you might experience irq-problems soon), you will get hw-acceleration everywhere.
mfg.a.sdr IOhannes
Nvidia cards are the only ones I know of on the market that do have HW on both heads at the same time. ATI cards will only do accelerated overlay on one head or the other, not both.
I've never seen a card that does more than two heads both/all accelerated either, but apparently the nividia chipsets allow this.
Also seems to me two slower machines with fast-cheap cards would be cheaper than one supper machine with one fast dual-head card. If you are reading a lot of textures off disk two machines becomes even more efficient.
Ben
IOhannes m zmoelnig wrote:
timon botez wrote:
Do you mean that it wont be possible with current library? Whats the reason for not getting OpenGL acceleration when doing this? Would it be possible for my developer to write such addition based on the hardware spec? I appreciate your response.
the problem is not openGL-specifique. it is rather a problem with the hardware.
in those early days, when the first dual-head gamer's graphics-cards with hw-openGL-acceleration where sold, you would often get hw-acceleration only on the first head. so you couldn't really do dual-head gem (at least not with a reasonable performance). nowadays, dual-head cards normally (?) support hw-acceleration on both heads.
anyhow, i have no idea how it is with "many-head" cards (like 8 or 16) but i guess these cards will not have openGL-hw-acceleration at all (or, if they do, all heads will be supported, since i don't think that anybody would buy such a card, to only play doom on the first screen and watch the stock exchange indices on all the other monitors...)
but it all depends on the specs of the card.
probably it would be simpler to connect a bunch of computers (via netsend/netreceive) rather than have everything done on one machine.
and i guess, if you could get 4 or more multi-head geforce (radeon ?) cards (1 agp and 3 pci) to work in one computer (which *would* be possible, i guess; but you might experience irq-problems soon), you will get hw-acceleration everywhere.
mfg.a.sdr IOhannes
GEM-dev mailing list GEM-dev@iem.at http://iem.at/cgi-bin/mailman/listinfo/gem-dev
-
B. Bogart -
IOhannes m zmoelnig