LONDON – Nvidia Corp., best known as a supplier of graphics processor units (GPUs), has announced that it intends to build high-performance microprocessors for applications ranging from PCs and servers up to workstations and supercomputers. Nvidia (Santa Clara, Calif.) will base its designs on an architecture license from ARM Holdings plc (Cambridge, England) in what is a full-frontal attack on the world's largest chip company, Intel Corp.
Nvidia has been a long-standing licensor of ARM intellectual property but has previously made use of individual core licenses. "Project Denver," as the latest initiative is known, features an Nvidia processor running the ARM instruction set, which will be integrated with an Nvidia GPU array.
This processor stems from a strategic partnership, also announced today, under which Nvidia has obtained rights to develop its own high performance CPU cores based on ARM's future processor architecture. In addition, Nvidia has licensed ARM's current top-of-the-range Cortex-A15 processor for its future-generation Tegra mobile processors.
The news comes as Microsoft has announced that the full Windows 8 operating system will run on the ARM architecture, thereby enabling vendors of non-x86 processors to enter the Windows PC market.
"ARM is the fastest-growing CPU architecture in history," said Jen-Hsun Huang, president and chief executive officer of Nvidia, in a statement. "With Project Denver, we are designing a high-performing ARM CPU core in combination with our massively parallel GPU cores to create a new class of processor," he added.
Warren East, ARM chief executive officer said: "Nvidia is a key partner for ARM and this announcement shows the potential that partnership enables. With this architecture license, Nvidia will be at the forefront of next generation SoC design, enabling the Internet everywhere era to become a reality."
Disclosure: I work for ARM
Regarding Cortex-A15 processor; Our model is to design a processor for multiple markets. The product definition work ensures that ARM profiles the efficiency with which the processor handles multiple workloads. For ARM's applications processors, the first instantiations in actual SoCs are normally focused the mobile arena as evidenced by the 3 lead silicon partners revealed in September 2010 at the Cortex-A15 launch. But the processor's extended addressing, hardware virtualization support and higher performance will, I believe, see it gain adoption over time into a number of additional segments including automotive infotainment, medical and yes, some computing/server applications. One other point. There is nothing in the processor itself that prevents more than 8 or 16 or.. processors to be put down on a piece of silicon. ARM focuses its on-chip interconnect IP toward high volume applications so naturally, this tends to be centered toward smaller processor configurations. Other partners that wish to target higher performance applications are either utilizing their own internally-developed interconnect technology or licensing IP from other 3rd parties to form significantly bigger processor compute subsystems.
Disclosure: I work for ARM..
Been sat reading this thread for the past couple of days trying to decide if/how to chime in here. In this area, I see us as having two beliefs
1) Different applications require different computing requirements. One size does not fit all
2) There is room (indeed there is demand) for technical innovation alongside the processor core to deliver optimized solutions for particular applications. Algorithms that involve bit manipulation are often better handled in hardware than in general purpose processors...
Microsoft's announcement this week delivers an additional element of choice to the ARM ecosystem. It will be massively important for some companies, while this announcement will be a NOP for others.
That is because OS was developed by SW engineers, not HW engineers. There are lots lots of resource wasting in boot process. I hope one day some talent hardware engineers could come up a OS which could have boot time measured in ms not minutes. :-)
Both Intel and AMD produced many different versions of x86 chip with different packages and configuration over years. I don't see any issues for Windows to support that.
If you look Windows 8 news, it is going to support ARM based SoC from Qualcomm/nVidia/TI at the same time. Although all three chips have same ARM core(Qualcomm uses its own core based on ARM architecture), other peripherals devices are vastly different. Open source Linux has been supporting many different flavors of SoC now.
If you look the context of this article, you can find out that we are talking about nVidia being a NOTEBOOK/DESKTOP CPU replacement running Win 8 OS.
For Embedded market what you are saying totally makes sense.
Why people keep bitching about windows boot speed? It is very good to me considering its complexity. Do a quick experiment on your iPhone. Remove battery and put it back on. Time it to see how fast to boot. It is unfair comparison between cold boot speed in PC to wake up speed in your iPhone.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.