Nvidia’s monstrous Volta GPU appears, packed with 21 billion transistors and 5,120 cores - ortizliandn
"We pauperism to receive a path smart for life after Dudley Moore's Law," Nvidia CEO Jen-Hsun Huang said at the starting time of his period of time GPU Engineering science Conference keynote. Only Nvidia ISN't hesitant to bedevil around more iron to make its ferocious graphics processors flat more so, as evidenced by the reveal of the start product based on Nvidia's badass next-gen Volta GPU.
Nvidia's shrilling-end "Pascal" processors still find the graphics perch, though AMD's rival Radeon Vega GPUs are regular to launch before the end of June. Volta helps Nvidia get hold of some of the lead unconscious of AMD's sails before Vega even hits the streets, even if the Tesla V100 GPU is focused on data centers.
This beastly GPU—both in size and capabilities—boasts a whopping 21 billion transistors and 5,120 CUDA cores hum along at 1,455MHz boost time speeds, all reinforced using a 12-nanometer manufacturing process more advanced than that of Nvidia's current GPUs. By comparison, today's Pascal GPU flagship, the 14nm Tesla P100, offers 3,840 CUDA cores and 15 billion transistors. The GeForce GTX 1060 has aquarter as numerous CUDA cores as the Tesla V100, at 1,280.
To fit all that technical school, the Volta GPU in the Tesla V100 measures a borderline ridiculous 815mm square, compared to the Tesla P100's 600mm GPU. Monstrous.
Nvidia's Volta-based Tesla V100.
Volta is "at the limits of photolithography," Huang same with a smirk, created using an R&D budget of finished $3 billion.
Nvidia says information technology's redesigned Volta's streaming microprocessor computer architecture to be 50 per centum more efficient than Pascal's, which is damned brilliant if IT proves true. That enables "star boosts in FP32 and FP64 performance in the same power envelope," Nvidia says. The Tesla V100 also includes new "tensor cores" made-up specifically for wakeless learning, providing 12 times the teraflops throughput of the Pascal-based Tesla P100, Huang said. (Google's also invested with in tensor processing hardware.)
The Tesla V100 hits a peak of:
- 7.5 TFLOP/s of double precision floating-point (FP64) performance;
- 15 TFLOP/s of single precision (FP32) performance;
- 120 Tensor TFLOP/s of heterogeneous-precision matrix-multiply-and-accumulate.
Seriously, the Volta GPU inside the Nikola Tesla V100 is MASSIVE.
The Tesla V100 utilizes 16GB of ultra-fast, 4096-bit high-bandwidth memory to swear out data quickly. It's unknown whether Conte Alessandro Volta-based consumer art card game volition feature HBM2, all the same. Radeon Lope de Vega does, but the tech is still relatively untested and pricey. The GeForce GTX 10-serial graphics card game debuted with new GDDR5X technology supported classic memory designs, and SK Hynix recently said that IT's "planning to mass green goods the product for a customer to handout high-end graphics card away early 2018 equipped with high performance GDDR6 DRAMs."
That HBM2 memory hits 900GB/s speeds, Nvidia says, and the Tesla V100 features a indorsement-gen version of Nvidia's NVLink technology. At 300GB/s transfer speeds, Huang claims NVLink is now 10 times quicker than classic PCIe connections.
Inside the Nikola Tesla V100.
If you want to know more about Volta's datacenter and computer architecture details, be bound to train out Nvidia's in-profundity Tesla V100 explainer. Anticipate the Tesla V100 to plunge in a revamped edition of Nvidia's pricey DGX-1 system in the third quarter, and more widely in the fourth billet.
The impact on you at place: None, immediately. Only this initiatory glimpse at Volta gives us an idea of what Nvidia's next-gen GeForce graphics cards will be capable of. Remember, Nvidia revealed its Pascal GPU at GTC 2016 in the mold of the Tesla P100, and that full-oily version eventually trickled down into the Titan Xp, with the GeForce GTX 1080 Ti coming damned close. The Tesla V100's extreme GPU size and beefed-up simple machine erudition focus skews that somewhat when IT comes to Volta, though.
Unmatchable more affair: The GeForce GTX 10-serial launched a plain month aft Pascal's GTC discover, around ane year ago. Assume't necessarily expect Nvidia's Volta-supported GeForce card game to plunge in the near future—especially if SK Hynix's GDDR6 memory board is so headed for Nvidia cards rather than next-gen Radeons.
Source: https://www.pcworld.com/article/406742/nvidias-monstrous-volta-gpu-appears-packed-with-21-billion-transistors-and-5120-cores.html
Posted by: ortizliandn.blogspot.com

0 Response to "Nvidia’s monstrous Volta GPU appears, packed with 21 billion transistors and 5,120 cores - ortizliandn"
Post a Comment