Intel has issued a mountain of patches for the Intel Linux graphics driver, precursor to discrete GPU support on the open source Linux kernel. The 42 patches and 4,000 lines of code implement local memory support, a crucial step towards offering full Linux driver support for Intel Xe arriving in 2020.
Last year Intel announced its ambitious goal to not only build a discrete graphics product but turn the whole project around by 2020. Since hiring some of the best boffins in the biz – such as ex-AMD and Apple Raja Koduri – the company has been slowly teasing its upcoming product. And so far it’s all looking rather positive for the open source community, with Intel announcing extensive Linux driver support and VESA Adaptive Sync (the basis for AMD FreeSync) support.
“Local memory implementation is the first of many steps toward robust Linux support for our future discrete graphics solutions,” a tweet from Intel Graphics says. Until this point, Intel’s integrated graphics silicon, regardless of its various on-chip configurations, has not required this functionality.
These recently unearthed local memory patches allow for utilisation of different memory regions by the Linux kernel – a necessity for discrete graphics cards packed with their own pools of speedy VRAM.
While that’s promising for the open source community, there’s still plenty of work to get on with for complete Linux driver support. Current patches only imitate the necessary infrastructure for local memory regions, and, according to the Linux gurus over at Phoronix, it is often roughly a year between initial Linux kernel patches roll out and physical hardware debuts on the market.
Our journey toward a new visual computing experience is underway, and that includes a commitment to the #OpenSource community. Local memory implementation is the first of many steps toward robust Linux support for our future discrete graphics solutions. https://t.co/AUFTA1zo1V pic.twitter.com/G3POZXYW6H
— Intel Graphics (@IntelGraphics) February 15, 2019
Aside from a few promo reels Intel has largely kept the details of its upcoming dGPU under wraps. We do know, however, that the company intends on reshaping its current graphics technology for use in both client and data centre dGPU applications.
Over at CES 2019 Gregory Bryant, VP of client computing, confirmed that the company’s upcoming GPU will be built upon the foundational building blocks laid out by Gen11 graphics – the graphics silicon soon to be baked into 10nm Ice Lake CPUs.
And all this comes at the same time as Intel acquires Ineda Systems, and some 100 engineers, for their graphics technology expertise.