News Article
A new phase of memory research
Samsung announced this week that it has completed the first working prototype of a Phase-change Random Access Memory (PRAM) that Samsung believes will replace high density NOR flash within the next decade.
The company unveiled the 512Megabit (Mb) device this week. PRAM features the fast processing speed of RAM for its operating functions combined with the non-volatile features of flash memory for storage. Another key advantage of PRAM is its extremely fast performance. Because PRAM can rewrite data without having to first erase data previously accumulated, it is effectively 30-times faster than conventional flash memory. Incredibly durable, PRAM is also expected to have at least 10-times the life span of flash memory. Samsung expects that PRAM will be available beginning sometime in 2008.
The current technologies of DRAM, SRAM, and NOR were all developed in support of box-level computing applications that attached to the wall for both their power and their data connection. DRAM was supposed to be the cheapest cost per bit, SRAM was supposed to be the fastest, and NOR was the nonvolatile memory that held the basic BIOS for the PC.
We have shifted away from this particular application for the development of future memory technologies and now the focus is on mobile applications. However as we change the end application, we also change the value proposition for the new memory technologies. And this is where the new memory technologies fit into the picture.
Moore's Law is still the driving force for logic design, and Moore's Law is pushing us toward an even more mobile platform for highly personal computing and digital entertainment. But what is still missing to trigger that application is an embeddable high-performance non-volatile memory that can go onto the same die as processing logic.
Freescale very recently announced their MRAM technology with a small footnote in the press release that the technology was eventually intended as an embeddable NVM for automotive applications. Within the following week, both Renesas and NEC also announced that logic-based products with embedded MRAM would be available within the next two years. Fujitsu also already has several products with FeRAM, another one of the new forms of nonvolatile memory technology, and Fujitsu expects to make that technology generally available as part of their SOC program within the next two years. TI has also been focused on this same kind of technology.
Intel licensed a phase-change nonvolatile memory technology several years ago in order to utilize that technology as a potential technology for cell phones, although the future of that program is not clear.
Samsung licensed that same phase-change technology last year, and it now appears that they have made great progress in developing that technology. What is significant about the Samsung announcement is that Samsung is already one of the top manufacturers of cell phones. Samsung is also very focused on developing a cell phone platform as the platform of choice for highly personal computing applications.
If you look at Samsung's current offering for 3G cell phone memory, you can see a multi-chip package with two DRAM die, two PSRAM die, two NAND die, and two NOR die. It is obvious that the memory technologies developed for box-level computing applications are just not the most efficient memory technologies for these highly mobile applications. It would be great if a single nonvolatile memory technology could combine the density of DRAMs and enough performance to at least eliminate a few of the 8 memory die required for 3G performance! And the performance range of some of the developing new memory technologies--and specifically the performance range of phase-change memories--does neatly fit into that 3G application and can ideally become eventually embedded into the processor die itself.
Phase-change nonvolatile memory is in the early wave of the new memory technologies that are being developed for the next level of computing that is clearly predicted as a direct extension of Moore's Law. We expect that this kind of technology will mark the leading edge of memory technologies for the next generation of memory development.
