SAN FRANCISCO—South Korea's Samsung Electronics Co. Ltd. said Tuesday (Aug. 2) it acquired spin-transfer torque random access memory (STT-RAM) vendor Grandis Inc. Financial terms of the deal were not disclosed.
Grandis will be merged into Samsun R&D operations focused on developing next-generation memory , where new semiconductor materials and structures are reviewed for their long-term commercial value, Samsung said. The company said Grandis (Milpitas, Calif.) would contribute to its development of memory technology and become a key part of the company's global R&D network.
Grandis was established in 2002 to develop SST-RAM, which is said to combine the cost benefits of DRAM, the fast read and write performance of SRAM, and the non-volatility of flash. STT-RAM is also said to solve the key drawbacks of first-generation, field-switched magnetic-RAM (MRAM). The company received total funding of $15 million from several investors, including Applied Ventures, Sevin Rosen Funds, Matrix Partners, Incubic and Concept Ventures.
In June 2010, Grandis updated its product roadmap with the ambitious goals of replacing DRAM, and eventually, NAND, with its next-generation MRAM.
STT-RAM is a second-generation MRAM technology that is said to solve some of the problems posed by conventional MRAM structures. Most MRAMs that are now being developed write data by applying the magnetic field generated by a current running through a wire near a tunneling magnetoresistive element to change the magnetization. That enables fast operation, but, according to Grandis, it also gobbles power.
For years, developers of MRAM and other next-generation memory technologies have claimed that they would become the ultimate universal memory and replace today’s memories. But, to date, none of the technologies have lived up to the hype, while current memory technologies such as DRAM and NAND have continued to scale.
It should be known by now Samsung, Hynix, Micron, Toshiba all got their hands in the alternative memories like MRAM and RRAM. When DRAM and Flash are respectively replaced by these technologies, those who failed to pay for at least some research in these areas will be wiped out. Long-term consolidation catalyzed by investment in new technology.
Yes, it is GAME OVER for phase-change memory (PCM). Samsung realized last summer that its PCM has horrible power consumption in handsets, in addition to being slow in write, costly to make, with poor density and impossible to scale. Now the scamsters at IBM and Numonyx (now part of Micron) look real stupid!
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.