from the free-vega dept.
AMD's Vega GPUs will be launching during the first half of 2017 (in 3-5 months). While AMD's previous Polaris chips were aimed at mainstream segments, AMD will aim for higher performance enthusiast users with Vega. Vega's next-generation compute unit will allow more lower precision instructions per clock cycle, such as FP16 and INT8 operations. The GPUs will also include High Bandwidth Memory 2.0. HBM2 allows up to 8 GB of DRAM per stack rather than the 1 GB limit imposed by HBM1 (used in AMD's Fury GPUs).
AMD has also announced FreeSync 2 for variable refresh displays, which apparently has more to do with high dynamic range (HDR) color than screen tearing and frame latency:
Trying to explain FreeSync 2 can get a bit tricky. Unlike the original FreeSync that it takes its name from, it's a multi-faceted technology: it's not just variable refresh, instead it's HDR as well. But it's also a business/platform play in a different way than FreeSync was. And while strictly speaking it's a superset of FreeSync, it is not meant to replace FreeSync wholesale. Perhaps the best way to think of FreeSync 2 is that it's a second, parallel initiative that is focused on what AMD, its monitor partners, and its game development partners can do to improve the state of high-end monitors and gaming.
In terms of features then, what is easily the cornerstone feature of [FreeSync 2] – and really its reason to be – is improving support for HDR gaming under Windows. As our own Brandon Chester has discussed more than once, the state of support for next-generation display technologies under Windows is mixed at best. HiDPI doesn't work quite as well as anyone would like it to, and there isn't a comprehensive & consistent color management solution to support monitors that offer HDR and/or color spaces wider than sRGB. The Windows 10 Anniversary Update has improved on the latter, but AMD is still not satisfied with the status quo on Windows 10 (never mind all the gamers still on Windows 7/8). As a result FreeSync 2 is, in part, their effort to upend the whole system and do better.
FreeSync 2 also at Tom's Hardware.
In response to increased demand, Samsung is increasing production of the densest HBM2 DRAM available:
Samsung on Tuesday announced that it is increasing production volumes of its 8 GB, 8-Hi HBM2 DRAM stacks due to growing demand. In the coming months the company's 8 GB HBM2 chips will be used for several applications, including those for consumers, professionals, AI, as well as for parallel computing. Meanwhile, AMD's Radeon Vega graphics cards for professionals and gamers will likely be the largest consumers of HBM2 in terms of volume. And while AMD is traditionally a SK Hynix customer, the timing of this announcement with AMD's launches certainly suggests that AMD is likely a Samsung customer this round as well.
Samsung's 8 GB HBM Gen 2 memory KGSDs (known good stacked die) are based on eight 8-Gb DRAM devices in an 8-Hi stack configuration. The memory components are interconnected using TSVs and feature over 5,000 TSV interconnects each. Every KGSD has a 1024-bit bus and offers up to 2 Gbps data rate per pin, thus providing up to 256 GB/s of memory bandwidth per single 8-Hi stack. The company did not disclose power consumption and heat dissipation of its HBM memory components, but we have reached out [to] Samsung for additional details.