AMD 7nm Vega Radeon Instinct GPU AI Accelerators Enter Lab Testing [hothardware.com]
AMD's current generation Vega graphics architecture – which powers its Radeon RX Vega family of graphics cards -- is based on a 14nm manufacturing process, but the chip company is already moving along with next generation process technology. During the company's conference call with analysts following its Q1 2018 earnings report (which it knocked out of the park [hothardware.com], by the way), AMD CEO Dr. Lisa Su made some comments regarding its upcoming 7nm GPUs.
"I'm also happy to report that our next-generation 7-nanometer Radeon Instinct product, optimized for machine learning workloads, is running in our labs," said Dr. Su [seekingalpha.com]. "We remain on track to provide samples to customers later this year."
If you recall, Radeon Instinct is AMD's product line for machine intelligences and deep learning accelerators [hothardware.com]. The current lineup features a mixture of Polaris- and Vega-based GPUs and could be considered competitors for NVIDIA's Tesla family of products. [...] According to commentary from AMD at this year's CES, 7nm Vega products for mobile along with the 7nm Radeon Instinct accelerators will ship during the latter half of 2018.
From The Next Platform [nextplatform.com], "The Slow But Sure Return Of AMD In The Datacenter":
"Server revenue increased in double digit percentage sequentially, across mega datacenter, OEM, and channel customers," Su said on the call. "Epyc processor unit shipments nearly doubled from the previous quarter. We continued to grow our datacenter momentum with dozens of new design wins across key workloads, including HPC, storage, virtualization, and cloud applications."
We have reported on many of these design wins, including those from Hewlett Packard Enterprise [nextplatform.com], Dell EMC [nextplatform.com], and Cray [nextplatform.com]. There are more than 40 server designs based on Epyc, and Su added that AMD is on track to have middle single digit server market share across units by the end of 2018. That was the number we were projecting back in March 2017 [nextplatform.com], when we said that Epyc would give AMD a second chance in servers, and we also said that this would just be share and profits taken directly from Intel. Part of the reason that we think Intel has charged a pretty hefty premium for beefier versions of the "Skylake" Xeon SP processors [nextplatform.com] is that it knew market share losses were inevitable in 2018 and it needed to make up that lost revenue and profit somehow. So far, that gambit seems to have worked.