Nvidia presenta su nueva arquitectura TURIN

Nvidia presento su nueva tecnología de GPUS llamada Turing, Anunciando que es el salto mas grande en desempeño desde la llegada de Pascal.

También presenta el enfoque a "rendering híbrido", enfocando el salto hacia RAYTRACING (siendo 25% mas rápido que Pascal). Sin olvidar las mejoras también en los núcleos TENSOR demostrados en VOLTA.

Moments ago at NVIDIA’s SIGGRAPH 2018 keynote presentation, company CEO Jensen Huang formally unveiled the company’s much awaited (and much rumored) Turing GPU architecture. The next generation of NVIDIA’s GPU designs, Turing will be incorporating a number of new features and is rolling out this year. While the focus of today’s announcements is on the professional visualization (ProViz) side of matters, we expect to see this used in other upcoming NVIDIA products as well. And by the same token, today’s reveal should not be considered an exhaustive listing of all of Turing’s features.

The big change here is that NVIDIA is going to be including even more ray tracing hardware with Turing in order to offer faster and more efficient hardware ray tracing acceleration. New to the Turing architecture is what NVIDIA is calling an RT core, the underpinnings of which we aren’t fully informed on at this time, but serve as dedicated ray tracing processors. These processor blocks accelerate both ray-triangle intersection checks and bounding volume hierarchy (BVH) manipulation, the latter being a very popular data structure for storing objects for ray tracing.

NVIDIA is stating that the fastest Turing parts can cast 10 Billion (Giga) rays per second, which compared to the unaccelerated Pascal is a 25x improvement in ray tracing performance.

The Turing architecture also carries over the tensor cores from Volta, and indeed these have even been enhanced over Volta. The tensor cores are an important aspect of multiple NVIDIA initiatives. Along with speeding up ray tracing itself, NVIDIA’s other tool in their bag of tricks is to reduce the amount of rays required in a scene by using AI denoising to clean up an image, which is something the tensor cores excel at. Of course that’s not the only feature tensor cores are for – NVIDIA’s entire AI/neural networking empire is all but built on them – so while not a primary focus for the SIGGRAPH crowd, this also confirms that NVIDIA’s most powerful neural networking hardware will be coming to a wider range of GPUs.

New to Turing is support for a wider range of precisions, and as such the potential for significant speedups in workloads that don't require high precisions. On top of Volta's FP16 precision mode, Turing's tensor cores also support INT8 and even INT4 precisions. These are 2x and 4x faster than FP16 respectively, and while NVIDIA's presentation doesn't dive too deep here, I would imagine they're doing something similar to the data packing they use for low-precision operations on the CUDA cores. And without going too deep ourselves here, while reducing the precision of a neural network has diminishing returns – by INT4 we're down to a total of just 16(!) values – there are certain models that really can get away with this very low level of precision. And as a result the lower precision modes, while not always useful, will undoubtedly make some users quite happy at the throughput, especially in inferencing tasks.

Getting back to hybrid rendering in general though, it’s interesting that despite these individual speed-ups, NVIDIA’s overall performance promises aren’t quite as extreme. All told, the company is promising a 6x performance boost versus Pascal, and this doesn’t specify against which parts. Time will tell if even this is a realistic assessment, as even with the RT cores, ray tracing in general is still quite the resource hog.

Mas informacion: www.anandtech.com/show/13214/nvidia-reveals-next-gen-turing-gpu-architecture

www.tomshardware.com/news/nvidia-turing-quadro-rtx-announced,37599.html

Nvidia finalmente ha revelado los precios para Turin.
Los cuales son ESTÚPIDAMENTE ALTOS.



Founder’s Edition Price (USD):

· RTX 2080 Ti - $1,199.00

· RTX 2080 - $799.00

· RTX 2070 - $599.00.

MSRP Price (USD):

· RTX 2080 Ti – Starting at $999.00

· RTX 2080 – Starting at $699.00

· RTX 2070 – Starting $499.00.

Las quejas no se hicieron esperar, ya que según eso.. con RTX.. apenas puedes mantener 60fps (con la versión TI) en 1080p en juegos como TOMB RAIDER.
El desempeño "dado" por nvidia en unos sliders muy sospechosos sin explicación, hacen que la gente especule si la velocidad del 2080 sera 50% mas rápido que la 1080 o si es 100% mas rápido. Dejando "como máximo" un 50% y mínimo de 25%.

Los costos hacen que la gente apunte al tamaño del chip Turin... se dicen de que Turin es simplemente un "Fermi" de demostración de tecnología antes de poder hacerla feasible y eficiente con la tecnología de 7nm que saldrá a finales de este año. Ya que el chip en si es asquerosamente enorme. Casi tan grande como el TITAN V.
Turin incluye todo el chip completo con solo "cortos" en el numero de cuda cores. Teniendo disponibles módulos de RTX (ray tracing, DSSP) y TENSOR cores para hacer el AntiAliasing con inteligencia artificial.



Hay también rumores de que Nvidia tiene bien agarrados a los reviewers, dando NDA para que no haya benchs hasta cierta fecha Y bloqueando los drivers hasta cierta fecha predeterminada.


Lo que han dicho en previews de diferentes grupos:

Overall, Nvidia is claiming the RTX 2080 is about 1.5x faster than the 1080 at 4K resolution, which is a decent amount of "uplift" as they say in marketing slides. If you have DLSS enabled, the delta can be as high as 2X, which is certainly surprising.

It struck us as curious they wouldn't be showcasing their new GPU at 4K, since one would assume that is what it was designed to do. However, since the Tomb Raider demo had ray tracing enabled, we assumed perhaps the performance hit was such that you simply can't run the game at 4K with ray tracing. I asked Nvidia about this and they said it can, and Tomb Raider ran sub-60fps simply because it's pre-alpha with beta drivers. When I asked why there were no 4K demo stations setup they just said they wanted to focus on ray tracing this time around, not 4k gaming. I'm not sure why they couldn't do both. One Nvidia rep assured me, of course, that 4K gamers would have no complaints with an RTX GPU.

IGN

Lo interesante que casi todos los demos se han hecho en 1080p y no en 4k. Lo cual es el "target" de 2080 y 2080TI.

The company's numbers suggest that the new card delivers circa 2x the performance of the last-gen GTX 1080 on certain games. Much of this remarkable gen-on-gen leap is derived via new Nvidia technology called DLSS - deep learning super-sampling. The same benchmarks running without the technique in play, along with titles not supporting DLSS, show performance increases more along the lines of 30 to 40 per cent.

It's important to stress that DLSS does not work on all games but for those that do support it, the key question concerns image quality. In its demo, Nvidia compares DLSS to standard UE4 temporal anti-aliasing. Based on the subjective eyeballing of three DF staff members, we'd say that based on the Infiltrator demo at least, Nvidia gets pretty close in achieving comparable quality here. Close-up head-to-heads revealed a minuscule amount of shimmer on the DLSS side but overall, the demo was rather impressive

Benchmark numbers for the RTX 2070 and the top-tier RTX 2080 Ti have not been revealed, but the RTX 2080 numbers - assuming they stack up, of course - do highlight that we are looking at a very different, new and improved GPU architecture that can't be assessed on core count alone. There has been some concern about standard 3D performance as we should expect an RTX 2080 to outperform the older generation GTX 1080 Ti, yet the new card has significantly fewer CUDA cores - 2944 vs 3584. However, the benchmark numbers Nvidia has shared suggest that the RTX 2080 should still have a performance advantage over GTX 1080 Ti despite this, with a further substantial leap delivered by the top-tier RTX 2080 Ti.

Digital Foundry

Overall, the idea is that even in traditionally rasterized games without DLSS, the GeForce RTX 2080 brings around 50% higher performance than the GeForce GTX 1080 under 4K HDR 60Hz conditions. Because this excludes real-time raytracing or DLSS, this would be tantamount to ‘out of the box’ performance. Though there were no graphics settings or driver details to go with these disclosed framerates, so I'm not sure I'd suggest reading into these numbers and bar charts one way or another.

Anandtech


Sobre precio y desempeño:

Before getting too far here, it’s important to point out that NVIDIA has offered little in the way of information on the cards’ performance besides their formal specifications. Essentially the entirety of the NVIDIA Gamescom presentation – and even most of the SIGGRAPH presentation – was focused on ray tracing/hybrid rendering and the Turing architecture’s unique hardware capabilities to support those features.

At any rate, with NVIDIA having changed the SM for Turing as much as they have versus Pascal, I don’t believe FLOPS alone is an accurate proxy for performance in current games. It’s almost certain that NVIDIA has been able to improve their SM efficiency, especially judging from what we’ve seen thus far with the Titan V. So in that respect this launch is similar to the Maxwell launch in that the raw specifications can be deceiving, and that it’s possible to lose FLOPS and still gain performance

Meanwhile as far as pricing is concerned, while NVIDIA does have their $999/$699/$499 baseline MSRPs, the reality is that you won’t be able to find cards at these prices. And likely won’t be able to for weeks or months. The launch-day sell-out phenomena means that board partners have prioritized higher-end card designs, with pricing to match.

So there’s certainly merit to any arguments that at least for the launch, prices are closer to $1150 and $750 respectively

ADORETV Traduce el Bullshit del marketing the Nvidia.
Turing en promedio es 47% mas rápido según las gráficas no "copeteadas", no el 2X que comparan con SLI de 1080GTX.

Notando también que el 2080RTX es ahora el equivalente en precio del TI, y el 2080TI es el equivalente del TITAN.

Con la novedad que uno de los editores de TomsHardware esta siendo quemado literalmente por todos los fans y expertos de hardware por una nota editorial en la que parece que esta 100% dentro de los bolsillos de Nvidia.

www.tomshardware.com/news/nvidia-rtx-gpus-worth-the-money,37689.html

Con cosas super chistosas y a su vez que es obvio que fue mas comprado que un presidente mexicano..

However, what these price-panicked pundits don't understand is that there's value in being an early adopter. And there's a cost to either delaying your purchase or getting an older-generation product so you can save money. Unless and until final benchmark results show otherwise, the new features of the Turing cards make them worth buying, even at their current, sky-high prices.

Y Gemas como..

When your whole life flashes before your eyes, how much of it do you want to not have ray tracing?

Respuestas de expertos fueron desde burla hasta ataques completos..
Por lo ridículo que han sido las "razones" que el editor en jefe de Tomshardware se puso a inventar para que compraran las tarjetas después de que ningún experto apoyara el "el preorder" de Turin.

Los comentarios tampoco se hicieron esperar..
Ahora la pregunta es si este "jefe editor" mantendrá su trabajo.

que quemada!! tsss TH ya no es lo que era

Buenísimo post Tamalero, ah como me reí con el video de Gamers Nexus.

Diaboliquin wrote: Buenísimo post Tamalero, ah como me reí con el video de Gamers Nexus.

date otra vuelta en youtube.. porque otros grupos de reviewers que ya dieron opiniones similares.