The New Chips on the Block - Part 1
April 09, 2021
Armv9 is here, and although it took nearly 10 years to arrive, it comes right at a major inflection point in the semiconductor industry where all of the major chipmakers have started dabbling in processor architectures outside of their heritage. The implications of this are currently being felt in more advanced use cases like the data center and high-performance computing, but the roadmap decisions of today – such as the massive emphasis on AI and ML workload processing in Armv9 that will surely continue under NVIDIA leadership – will have significant ramifications for the embedded systems of tomorrow. And, oh yeah, what does NVIDIA’s potential stewardship of Arm mean for NVIDIA competitors who also happen to be Arm licensees, like Intel and Xilinix? Brandon and Rich discuss.
Later, Intel’s Sailesh Kottapalli, Senior Architecture Fellow, and Mandy Mock, VP and GM of Product Engineering Systems, join the Insiders to help set the record straight on Moore’s law. To be fair, the number of transistors keeps increasing at a rate that’s fairly consistent with Gordon Moore’s original projections, it’s actually some of the ancillary benefits that we’ve come to expect from that transistor scaling that aren’t keeping pace. So, is it fair to say it’s dead? And, maybe more importantly, should anyone even care?
Finally, in this week’s Tech Market Madness, Perry Cohen tries to make sense of “the edge” with Matt Trifiro of Vapor O and Jacob Smith of Equinix, who co-chaired the LF Edge’s 2021 “State of the Edge” report. All these years later, do we have a definitive answer to the question, “What is the IoT edge?” And, if we can agree on something, how will it reach critical mass? Here’s a hint: Open source.