AMD is definitely on a roll in the United States for pre-exascale and future exascale systems, having won the deals at Lawrence Berkeley National Lab with the “Perlmutter” system, at Oak Ridge National laboratory with the “Frontier” system, and as the presumed front runner at Lawrence Livermore National Laboratory with the “El Capitan” system. …
Arm Holdings has announced that the next revision of its ArmV8-A architecture will include support for bfloat16, a floating point format that is increasingly being used to accelerate machine learning applications. …
If you thought the gang at Intel were Moore’s Law biggest devotees, you probably haven’t heard Philip Wong expound on the subject. …
It had to happen sooner or later, and really, the wonder is that it hasn’t already happened, given the nature of the infrastructure buildouts at hyperscalers and cloud builders, the skittishness of enterprises due to the trade war between the United States and China, the uncertainty in Europe over the British exit, and general concern with macroeconomic issues because it has now been almost twelve years since the last recession. …
It would be convenient for everyone – chip makers and those who are running machine learning workloads – if training and inference could be done on the same device. …
In July, VMware acquired Bitfusion, a company whose technology virtualizes compute accelerators with the goal of enabling modern workloads like artificial intelligence and data analytics to take full advantage of systems with GPUs or with FPGAs. …
If you think about it for a minute, it is amazing that any of the old-time IT suppliers, like IBM and Hewlett Packard, and to a certain extent now Microsoft and Dell, have persisted in the datacenter for decades or, in the case of Big Blue, for more than a century. …
Carey Kloss has been intimately involved with the rise of AI hardware over the last several years, most notably with his work building the first Nervana compute engine, which Intel captured and is rolling into two separate products: one chip for training, another for inference. …
It has been a long time coming, and it might have been better if this had been done a decade ago. …
The rapid changes underway in modern datacenters and HPC environment are demanding more compute power from a tech industry that is running into significant barriers to supplying that capacity. …
All Content Copyright The Next Platform