nVidia GT300 Fermi architecture unveiled!
http://www.guru3d.com/news/nvidia-gt...ture-unveiled/
An interesting tidbit of info just surfaced on the web through Bright Side of News, regarding NVIDIA's upcoming GT300 chip (high-end graphics card part. I'll do the blunt copy/paste here:
nVidia GT300 chip is a computational beast like you have never seen before. In fact, we would go as far out and state that this is as closest as GPU can be to a CPU in the whole history of graphics technology. Now, time will tell whatever GT300 was the much needed revolution.
Video DEMOSTRACION; Presentacion
http://www.youtube.com/watch?v=vPeZf8WJDM4 Bugatti Veyron en Raytracing (casi real) vs 3 PS3 para crear un carro cutre con la misma tecnologia http://www.youtube.com/watch?v=oLte5f34ya8
http://www.youtube.com/watch?v=dVsvvcn1AUE&NR=1
http://www.youtube.com/watch?v=79mvSWRocq4
http://www.youtube.com/watch?v=RMtQ62CnBMA
Beside the regular NV70 and GT300 codenames [codename for the GPU], nVidia's insiders called the GPU architecture - Fermi. Enrico Fermi was an Italian physicist who is credited with the invention of nuclear reactor. That brings us to one of codenames we heard for one of the GT300 board itself - "reactor".
When it comes to boards themselves, you can expect to see configurations with 1.5, 3.0 GB and 6GB of GDDR5 memory, but more on that a little bit later.
GPU specifications
This is the meat part you always want to read fist. So, here it how it goes:
•3 Billones de transistores
•40nm, fabricado en TSMC
•348-bit de interfaz de memoria
•512 Shader Processors - Renombrados a CUDA Cores
•32 CUDA Cores por cada Shaders Cluster
•Eso significa 16 Shaders Cluster
•1MB Cache L1 dividida entre 16KB Cache de memoria compartida
•768KB Cache L2 de memoria unificada
•Hasta 6GB de memoria GDDR5
•La Mitad de Velocidad de IEEE754 con precisión doble. Compatible con IEEE 754-2008.
•Arquitectura que soporta nativamente C [CUDA], C++, DirectCompute, DirectX 11, Fortran, OpenCL, OpenGL 3.1 y OpenGL 3.2.
Hace algunos meses atrás, se escuchaba el rumor que NVIDIA estaba preparando una especie de CPU para competir con Intel, pero carecían de licencia X86, y al final terminó siendo Tegra. Ahora sabemos de qué se trataban tales palabras en el fondo, un chip cGPU más que GPGPU, capaz de ejecutar las líneas de código de forma nativa y sin necesidad de un interpretador que lo transforme en lenguaje que la tarjeta gráfica pueda entender, especialmente con C++.
TSMC was in charge of manufacturing the three billion transistor mammoth, but it didn't stop there. Just like the G80 chip, nVidia GT300 packs six 64-bit memory controllers for a grand total of 384-bit, bringing back the odd memory capacity numbers. The memory controller is a GDDR5 native controller, which means it can take advantage of built-in ECC features inside the GDDR5 SDRAM memory and more importantly, GT300 can drive GDDR5 memory in the same manner as AMD can with its really good Radeon HD 5800 series. The additional two memory interfaces will have to wait until 28nm or 22nm full node shrinks, if we get to them with an essentially unchanged architecture. You can expect that the lower-end variants of GT300 architecture will pack less dense memory controller for more cost efficiency, especially on the memory side.
Ferni architecture natively supports C [CUDA], C++, DirectCompute, DirectX 11, Fortran, OpenCL, OpenGL 3.1 and OpenGL 3.2. Now, you've read that correctly - Ferni comes with a support for native execution of C++. For the first time in history, a GPU can run C++ code with no major issues or performance penalties and when you add Fortran or C to that, it is easy to see that GPGPU-wise, nVidia did a huge job.
no te preocupes aqui esta la traduccion p los k tienen problemas ^^!
http://translate.google.com.pe/translate?prev=hp&hl=es&js=y&u=http%3A%2F%2Fwww.gu ru3d.com%2Fnews%2Fnvidia-gt300s-fermi-architecture-unveiled%2F&sl=en&tl=es&history_state0=