IBM unveils its first seven-nanometre processor-powered Power10 CPU to be made by Samsung for next-generation hybrid cloud computing
IBM today unveiled its latest tenth-generation Power10 CPUs. The Power10 CPUs will be made with a 7-nanometre chip will power the next-generation hybrid cloud computing and will be available for sale in the second half of 2021.
IBM’s lead architect is calling the new Power10 processor “a holy grail of the industry:” It comes with the ability to pool petabytes of memory between multiple servers, potentially enabling significant cost savings for data centers. ‘Instead of having to put a lot of memory into every box, I could actually plug in a more reasonable amount of memory in each box, saving lots and lots of infrastructure costs,’ Bill Starke, IBM’s lead chip architect says.
IBM said that the 7-nm Power10 CPUs will be manufactured by Samsung Electronics. In the announcement, Big Blue says that the new 7nm processor-powered Power10 will be half the size of its 14nm Power9 predecessor. It added that Power10 will handle three times more data compared to Power9. Beyond its new “Memory Inception” feature for multi-petabyte memory clusters, the upcoming chip also comes with new memory encryption and artificial intelligence capabilities.
If IBM manages to stay on schedule and release the Power10-based servers by second half of 2021, it would put the company ahead of market leader, Intel which has promised its 7-nm chips in later 2021, early 2022. AMD, on the other hand, has had 7nm server processors on the market for roughly a year.
IBM said its Memory Inception feature allows multiple Power10-based systems to share multiple petabytes of memory, which can improve cloud capacity and economics for memory-intensive applications from vendors like SAP and SAS Institute, as well as inferencing for large-scale AI models.
Starke stated that the Power10’s Memory Inception essentially allows the Power10 processor to “trick a system into thinking that memory in another system belongs to this system,” and it has significantly lower latency than traditional methods like remote direct memory access over Mellanox InfiniBand interconnects.
“This is an extremely low latency way to basically use just the speed of electricity through a cable and nothing more in your way to getting to other systems’ memories,” he said. “And what this really means that this has kind of been a holy grail of the industry.”
IBM’s Power9 is already used in one of the world’s fastest supercomputers, Summit, at the US Department of Energy’s Oak Ridge National Laboratory, alongside Nvidia Tesla GPUs.