ATI's stream enhances Folding@Home performance by over 3000%.

mascarilha

mascarilha

I folded Myself
ATI's stream enhances Folding@Home performance by over 3000%.

A while ago, NVIDIA released Gelato, which was its entry into the idea of stream computing. Stream computing is a way to run CPU-heavy applications on a GPU. This theoretically would be great, considering the multiple shader and vertex processors on a GPU. While NVIDIA's Gelato was an early proponent of the concept, ATI seems to have nearly perfected it. Their version of stream computing can increase a programs performance by between 10 to 40 times it's orignal performance, effectively making a graphics card's performance equal to the performance of a rack of servers. Of course, this is only if a program is coded to utilize the extra processing power. If it isn't, then the program will spend too much time communicating between CPU and GPU, and very little work will get done, causing a performance loss.




Pictured from From left to right: Chas Boyd of Microsoft, Jeff Yates of Havok, Vijay Pande of Stanford University, Michael Mullaney of Peakstream, and Dave Orton of ATI. Stanford University, known for creating Folding@home, reports that their program is fully compatible with ATI's stream, and the performance benefits are very real. Packstream makes middleware, and their programs fully support stream. And because they make middleware, programmers can use Packstreams API's to make their programs, which will hence be fully compatible with stream computing. Microsoft says that some components of Vista, like Aero, already use some elements of stream technology. And Havok was there to confirm physics on a GPU.


http://www.theinq.com/default.aspx?article=34783
Clicar para expandir...
 
Última edição:
Pronto, é a loucura!

E de repente boards com varios Full size PCIe e Crossfire, passam a fazer parte dos sonhos molhados do pessoal do Folding. :D
 
First was Vijay Pande of Stanford University, one of the people behind Folding@Home. Running Folding@Home on a GPU, specifically an X1900 class card. It is seeing between a 20-40 times speedup depending - not least how fast a CPU can feed the card. This would mean a GPU can do what most of a rack of servers can.

DX10 and Vista would use many of these stream techniques. Vista goes a lot farther than most of its predecessors by using a GPU to accelerate the desktop, at least in Aero Glass.
Clicar para expandir...


esta tecnologia parece mesmo aliciante , estou deveras curioso :D




Early examples promise speedups of tens of times over the fastest CPU out there at a fraction of the cost
Clicar para expandir...



:kfold: Faster


cumps
 
Última edição:
onde é que eu posso fazer o download? hum hum? isso é compativel com que familias de placas graficas da ATi?

EU QUERO! :D

Aparicio disse:
E para o BOINC não? :sad1:
Clicar para expandir...
claro que sim. se leres o texto da noticia, ela diz que pode aumentar o desempenho de programas em 10 a 40x, desde que o programa esteja preparado para multi processamento, como é o caso do boinc (mas atenção que não é o caso de todas as science app)
 
Estou a ver que o mercado do Folding não é desprezável. :D

Nem quero imaginar o que será uma PS3 com o Cell e o GPU a remarem para o mesmo lado. 100x mais rápido?
 
blastarr disse:
Sempre é "3000% mais rápido" (30x) ?
Clicar para expandir...
Esperemos que sim.! :)

Mas pelo menos da nossa parte não o poderemos dizer, pois não sabemos a quantidade de informação gerada naquele período de tempo.

Eu "acredito" que sim, com toda a certeza.
 
Inicie sessão ou registe-se para poder participar.
Back
Topo