This site may earn affiliate commissions from the links on this page. Terms of employ.

With Turing fast approaching for consumer cards, Nvidia is bringing new GPUs to market place for information center and the HPC universe besides. Last calendar week, the visitor appear its new T4 GPU family unit, specifically intended for AI and inference workloads and taking over for the Tesla P4 in this role.

Nvidia claims the new GPU is upwardly to 12x more power-efficient than its Pascal predecessor. The company has released a suite of benchmark tests showing the T4 blasting past its contest, though as always, such vendor results should be treated with a grain of common salt. We've seen Intel release exam results claiming its ain Xeon processors are excellent at inference, for case. The degree to which this is or is not true is likely the result of optimization flags and specific examination configurations or scenarios.

2018-09-12_1521

Specs on the new T4 are impressive. 16GB of GDDR6 feeds a cluster of 2560 CUDA cores and 320 Turing Tensor cores, all within a svelte 75W power profile. THG reports that the Tesla T4 has an INT4 and even an experimental INT1 mode, with up to 65TFLOPS of FP16, 130 TFLOPS of INT8, and 260 TFLOPS of INT4 performance on-tap. The older P4, in contrast, offers 5.5TFLOPS of FP16 and 22 TFLOPS of INT8. Nvidia says there are optimizations for AI video applications likewise and a buffed-up decoder that can handle up to 38 HD video streams simultaneously.

Aslope the new T4, Nvidia is also launching new tools for evolution, including a refresh of its TensorRT software package and a new, Turing-optimized version of CUDA (CUDA ten) that includes libraries and models that take been optimized for Turing. Spotter for the battle over inferencing to heat up in 2022. Intel is pushing into that infinite with Xeon, AMD wants a piece of the pie for its Radeon Instinct line-upward of motorcar accelerators, Nvidia's Turing is going to play in this space, so, come 2022, Intel will accept GPUSEEAMAZON_ET_135 See Amazon ET commerce architectures of its own to compete with. And that'south before you consider the custom accelerators that everyone from Fujitsu to Google has been building and deploying.

Information technology's non always clear how these technologies will impact consumers; Nvidia'south push to innovate ray tracing and DLSS are the nigh prominent example we have and so far of a company working to take the designs information technology'due south built for HPC and bringing them over to the consumer space. We don't yet know if information technology'll work. Only in that location's clearly a multi-manner fight brewing betwixt the largest titans of the industry — and Nvidia wants to take an early leadership position with its new line of GPUs.

Now Read: Nvidia AI Erases Noise From Images, Nvidia Unveils Turing GPU Architecture, and Tesla Dumps Nvidia, Goes it Solitary