Posted by Marc Greenberg on November 19, 2013
Mobile DRAM has certainly come a long way in the last few years, but I didn’t expect to see this prediction quite so soon, that “…the shipment of mobile DRAM is likely to officially surpass that of PC DRAM in 2014“.
That quote, from the DRAMeXchange research division of TrendForce, would cap off an amazing run of popularity for mobile memory if it proves to be true.
The history of mobile DDR DRAM memory is proof that great things can come from humble beginnings.
There were manufacturer specifications for mobile DDR DRAM long before there was a JEDEC spec for it. The first LPDDR specifications (sometimes called LPDDR1, mDDR) which were very similar to the DDR standard (“DDR1”) that was being used in the PCs of the day, the main differences being reduced standby power, the lack of a DLL in the LPDDR parts, and (for me) the annoying property that the CKE pin needed to be driven oppositely for LPDDR parts compared to DDR parts at bootup. LPDDR was several years behind DDR in specification maturity and LPDDR didn’t even get its own JEDEC specification number (JESD209) until 2007, before that it was considered by JEDEC as a variant of the PC DRAM and given a JESD79 standard number like the PC DRAMs.
LPDDR2 was an amazing and difficult effort. A small group of people were regular attendees of the JEDEC JC42.6 subcommittee meetings – I attended more than a dozen in-person meetings and countless calls – and we struggled to get meeting space, time on the agenda, and tried to avoid getting sucked into the mainstream PC DRAM proceedings while we explained why LPDDR2 had to be different from DDR2. We had to invent an IO, and let this be my public apology for the name “HSUL12” (High Speed Unterminated Logic) which I proposed for the LPDDR2 IO, but in my defense I will say it was the best option, and none of us predicted that HSUL12 would eventually become a terminated IO at the higher speeds of LPDDR3.
Suddenly there was this pressure to make LPDDR2 faster, and we didn’t believe it could be done, until several smart people on the committee figured out that if you were to terminate the IOs, reduce the IO capacitance, and provide a means of training the interface, that you could make an LPDDR3 that was as fast as DDR3.
Before the JEDEC committee was done with LPDDR3, there was the pressure for LPDDR4, and although DDR4 is hitting the market first, it appears that LPDDR4 will be faster than DDR4, which is also quite a turn of events. And people are already looking at what comes next…
It would be pretty amazing to me to see if mobile DRAM does indeed exceed PC DRAM shipments. I had always hoped for it, but never predicted it as soon as 2014.
Graham Allan is the Sr. Product Marketing Manager for DDR PHYs at Synopsys. Graham graduated from Carleton University's Electrical Engineering program with a passion for electronics that landed him in the field of DRAM design at Mosaid in Ottawa, Canada. Beginning at the 64Kb capacity, Graham worked on DRAM designs through to the 256Mb generation. Starting in 1992, Graham was a key contributor to the JEDEC standards for SDRAM, DDR SDRAM and DDR3 SDRAM. Graham holds over 20 patents in the field of DRAM and memory design.