TRASTARO escribió:Interesante y revelador;
de ser cierto cosa que dudo, esto que dice Robert Hallock, jefe de mercadotecnia de AMD. Segun este mercadologo, en configuraciones de varias GPUs en paralelo [CrossFireX o SLI], con la API mantle y direct3D_12 mientras el juego no usen la configuracion AFR [la mas usada estos dias] para repartir el trabajo, el juego vera al conglomerado de GPUs como una sola GPU y la VRAM la vera como la suma de la VRAM en las tarjetas.
Considerando dos tarjetas de video en paralelo:
AVR [Alternate Frame Rendering] se refiere a que una GPU renderea toda la imagen del frame 1, 3, 5, etc. [impares], mientras el otro GPU renderea toda la imagen del frame 2, 4, 6, etc. [pares]
SFR [Split Frame Rendering] se refiere a que cada GPU se encarga de renderea la mitad de la pantalla
Mas del render en configuraciones CFX y SLI:
http://moddingmx.com/foro/index.php?/to ... rossfirex/Esto quiere decir, que si usando dos tarjetas de video con 4GB de VRAM y repartiendo el trabajo bajo el concepto SFR, donde cada tarjeta rendere la mitad de la pantalla, para el juego tendriamos una sola tarjeta de video con 8GB de VRAM.
Esto de sumar la VRAM, siendo sinceros, no lo creo viniendo de la gente del departamento de mercadotecnia, porque recordemos lo que paso cuando la previa salida de los procesadores AMD FX, donde las declaraciones de un aumento exponencial del rendimiento era anunciado precisamente por este departamento.
Lo que si creo, y ya estaba contemplado desde los inicios de mantle y finalmente seguro direct3D_12 lo tendra, es que esta API usaria tecnicas de SFR mas avanzadas, donde cada GPU trabajaria sobre la totalidad de la escena a mostrar, pero repartiendose el trabajo, asi una se encargaria de geometrias y texturas y otros calculos, mientras la otra GPU se encargaria de sombreado, efectos de luz/sombra y otros efectos visuales y se encargaria ya de la salida final de la imagen al monitor/televisor/proyector.hilo_otros-titulos-de-juegos-que-usaran-mantle_2014932_s10De mi Post del 2 de octubre 2014.
..
The way DirectX11 handles multiple GPUs is “AFR” or Alternate Frame Rendering, which as the name suggests means if you have two comparably powered GPUs they simply take turns rendering frames. This is in many respects the easiest approach to take – and is a great way of making your game CPU bound!]So possibly our Mantle version could show some big improvements when using this method.
However, with the independent control over the GPUs Mantle gives us, we could approach the problem very differently - for example one GPU could be rendering the basic geometry in the scene, while another handles lighting and shadows for the same frame, with the final image composited at the end. This may also provide a route for when GPUs aren’t of a comparable power level – for example an integrated APU motherboard coupled with a desktop GPU. It’s the potential for completely new approaches like this which excites me the most about Mantle and the APIs which will follow it.
http://www.tweaktown.com/news/43347/gef ... index.htmlAnthony Garreffa, tweaktown.com escribió:DirectX 12 and Mantle could pave the way for combining VRAM on your GeForce and Radeon GPUs! Finally!
Sure, your flashy new GeForce GTX 980 has 4GB of VRAM, and so does the one next to it in SLI. But while you have a total of 8GB of VRAM between two cards, only one set of VRAM is being used, it's not being combined. That is, for now. Things could change according to a recent tweet from AMD's Robert Hallock.
TweakTown image news/4/3/43347_23_geforce-radeon-gpus-soon-combine-vram-thanks-dx12-mantle.jpg
Hallock teased that with the upcoming APIs in Mantle and DirectX 12, two GPUs in SLI or Crossfire could possibly act as 'one big' GPU. Hallock said: "Mantle is the first graphics API to transcend this behavior and allow that much-needed explicit control. For example, you could do split-frame rendering with each GPU ad its respective framebuffer handling 1/2 of the screen. In this way, the GPUs have extremely minimal information, allowing both GPUs to effectively behave as a single large/faster GPU with a correspondingly large pool of memory".
TweakTown image news/4/3/43347_18_geforce-radeon-gpus-soon-combine-vram-thanks-dx12-mantle.jpg
"Ultimately the point is that gamers believe that two 4GB cards can't possibly give you the 8GB of useful memory", he continued. He added: "That may have been true for the last 25 years of PC gaming, but thats not true with Mantle and its not true with the low overhead APIs that follow in Mantle's footsteps". This isn't confirmation that memory stacking is going to happen, but it's a much better direction to be heading in, that's for sure. Especially with 4K and beyond gaming, VRAM is more important than ever and wasting 4-8GB of VRAM on SLI/Crossfire setups is just silly.
