The curtain's been dropped on much of Nvidia's upcoming Fermi-based graphics cards, and the five-hundred-and-twelve-core GF100 looks aforementioned a behemoth, indeed. A completely overhauled architecture is every about threesome things: scalability, parallelism, and geometry. Oh, and ripping your eyeballs out.
Tom's Hardware, Anandtech, HotHardware and others go pretty unfathomable on the new architecture, which is today eminently scalable. Here's the overall structure of the GF100, which should give you an idea of the scalability—the GF100 is made up of quaternary graphics processing clusters (GPC), themselves composed of quaternary moving multiprocessors (which are made up of 32 CUDA cores and texture units) and a raster engine:
To go deeper on architecture, you're meliorate soured reading the 10-page reports from some of the sites linked above, but bottom line, Tom's Hardware is predicting something aforementioned double the action of Nvidia's current GTX 285. Anandtech also points out that Nvidia's geometry action exclusive went 3x between the NV30 engine in the ancient GeForce FX 5800 and current GT200 in the GTX 280, but the Fermi-based GF100 has 8x the geometry action of the GT200. The endgame being that " it allows them to verify the aforementioned assets from the aforementioned games as AMD and create something that module look better. With more geometry power, NVIDIA can ingest tessellation and displacement mapping to create more Byzantine characters, objects, and scenery than AMD can at the aforementioned take of performance."
There is a cost. Even though it's at the 40nm process, those 3 1000000000 transistors are feat to run hot, and the GF100 maybe the hottest single-card GPU ever. It's also not feat to be cheap. At all. [Tom's Hardware, Anandtech]
Generated by best funky, gaul abizz.
Powered by Mas BAsy.
Technorati
Tidak ada komentar:
Posting Komentar