Qualcomm, Nvidia vie for top spot in AI chip efficiency test

In test data released Wednesday by MLCommons, an engineering consortium that maintains benchmarks widely used by the AI ​​chip industry, Qualcomm’s AI beat Nvidia’s flagship H chips classify images based on how much data each chip can perform per watt Central server query. Qualcomm’s chip is in .4 Server queries per watt with 03.Nvidia 4 queries per watt.

Qualcomm also beat Nvidia in object detection with a score of 3.8 queries per watt, This compares to 2.4 queries per watt for Nvidia. Object detection can be used in applications such as analyzing footage of a retail store to see where shoppers go most often.

However, Nvidia topped the charts in both absolute performance terms and power efficiency terms in the natural language processing test, which means It is the most widely used AI technique in systems such as chatbots. Nvidia hit .8 queries per watt, while Qualcomm came in second with 8.9 queries per watt. 400