AMD HAS TEASED more information about its forthcoming Vega-based graphics cards, revealing that they will come with either 4GB or 8GB memory and hinting that a launch is imminent.
Both versions will use the second-generation high-bandwidth memory (HBM2), which in addition to offering improved memory bandwidth will also provide greater power efficiency - something AMD cards have been lacking as of late.
Scott Herkelman, vice president and general manager of Radeon gaming at AMD, suggested that 4GB of HBM2 Vega graphics card memory could potentially have similar memory bandwidth to Nvidia's 11GB GTX 1080 Ti, although we'll believe that when we see the independent benchmarks.
More plausibly, he added that Radeon RX Vega graphics cards ought to be smaller than some of the AMD graphics cards that have been shipped of late, particularly the £200 stop-gap Polaris 10-based RX 480s, which appeared in the middle of last year.
This was produced on GlobalFoundries' 14nm FinFET process and was very much a 'warm up' in advance of the slew of Vega-based GPUs that AMD had promised for the second quarter of 2017.
With the second quarter just weeks away, and AMD in the habit of launching stuff at the beginning of quarters, the first of AMD's long-awaited new Vegas ought to be seeing the light of day very shortly.
The news was revealed at AMD's Tech Summit in China over the weekend.
The company also revealed plans to push into gaming laptops, given the more compact nature of the GPUs that the company is planning, although heat dissipation will (of course) also be a factor there.
What wasn't revealed were core configurations and clock speeds, and whether the company will offer essentially the same device at different levels of performance and form-factors (and prices, natch).
The Vega is AMD's response to Nvidia's 10-series, a new GPU architecture designed to take advantage of the latest 14nm and 16nm chip manufacturing processes. It's very much the 'second foot to fall' in AMD's 2017 comeback tour, which appears to be going splendidly so far. µ
Max-Q GPUs will launch in a 'record' 80 devices
Plus the way the app works means you may as well not have a password anyway
And you can pick up tickets for the unveiling
All hail (a) Tesla!