i wouldnt call fermi a broken architecture personally... imo the strategy behind it is questionable, ie focus on gpgpu and massive hw performance instead of efficiency and especially power efficiency...Originally Posted by Picao84
and no, B wont solve that... but what B can do is solve the yield and heat issues, or at least improve them... that would mean higher clocks and better performance, and less heat, which would mean a dual fermi card is feasible and nvidia can get the perf crown back...
why are they hiding all the infos in gpuz? sounds like its from the same guys who spread the previous fake gpuz shot of fermi...
it does...I'm not really sure that GPU-Z 0.3.9 supports Fermi's models though ... W1zzard, a comment, maybe ?
edit : Thanks Chickenfeed
oh wow, its real then!Yes, i wonder if he put thisI hope he responds.
but low mem clocks...
only 800 ie 3200 effective?
when a new highend and mainstream card is launched, its always nvidia ref design and partners only put on a sticker, flash the bios, and then bundle different cables and games in a shiny box.
after a while, nvidias favorite partners are allowed to make their own cards... but they have to follow some rules to make sure the cards wont blow up etc...
low end cards are usually all designed by partners to cut as many corners as possible or to give the partners room to diversify, ie make a beefier highend card and a very cheap lowend card...
in this case partners usually buy sets from nvidia. its uncommon to buy gpus only these days, its usually sets of memory and gpus, and they are sold in bundles pretty much.
afaik ati does it the same way... so, those scores of a 480 are actually real from what i can tell, at least the gpuz shot is real... so what do those numbers look like?
Bookmarks