这个算力已经远远超过NV的tesla V100的性能，V100 在27TFLOPS的16位浮点运算时，功耗就已经到达300W。而特斯拉今年发布的最强的自动驾驶芯片单颗也只有36TFLOPS的16位浮点的算力。 随着Ascend910（昇腾910）一起发布的还有全场景AI计算框架 MindSpore。
Chinese technology giant Huawei on Friday launched the world's most powerful Artificial Intelligence (AI) processor -- the Ascend 910 -- along with an all-scenario AI computing framework -- MindSpore. The Ascend 910 is a new AI processor that belongs to the company's series of Ascend-Max chipsets. The telecom giant had announced that the
Billed as the single chip with the greatest computing density, Ascend 910 delivers performance of up to 256 teraFLOPS under FP16 and 512 teraOPS under IN8 with declared max power consumption of 310W. In comparison, the GPU Tesla V100 delivers up to 125 teraFLOPS with a max power consumption of 300W, while Google’s TPU 2.0 with four ASICs can

The Atlas 900 PoD (model: 9000) is a basic unit of the AI training cluster based on Huawei Ascend and Kunpeng processors. It features powerful AI computing, optimal AI energy efficiency, and optimal AI scalability. The cluster basic unit is widely used in deep learning model development and training scenarios, and is an ideal option for

The Ascend 910, in conjunction with the MindSpore framework, is used to train AI models. Surpassing even Huawei's own expectations in terms of its performance, the new processor delivers 256
of the AI accelerators delivers the Huawei’s Ascend 910 [16], which reaches 256Tflop/s of raw FP16 performance, leading to 208Gflop/s /mm2 of compute density—nearly an order of magnitude (7.7x) more than IBM Power10 and but still only 55% of the NVIDIA A100’s peak performance. To summarize: modern architectures are moving towards
dxNUpF. 797 689 524 64 337 809 867 538

huawei ascend 910 vs nvidia v100