In the world of random access memory (RAM) there really hasn’t been the kind of rapid change that we’ve seen in many other areas of computer hardware. DRAM (dynamic random-access memory) has ruled the roost for almost 40 years, but that may be about to change.
Conventional DRAM temporarily stores and retrieves data as needed to run programs by charging and discharging a series of capacitors representing ones and zeros. If the computer is turned off, the data stored in DRAM is lost, causing the computer to “forget” what you were doing. This can be pretty annoying.
PCRAM, however, has no such limitations. By switching between a crystalline and gas-like state, it can store data, and otherwise function normally, without a power supply. Indeed, with PCRAM it would be possible to pick right back up where you left off, and resume whatever program you were using before the computer was shut off. This is probably a good thing in most cases, though admittedly sometimes it’s important to be able to do a cold boot. I assume if the technology makes it to market this issue will be addressed? I assume.
Anyway, PCRAM has actually been around for several years. The problem is, until very recently it’s offered unreliable speeds, making it considerably slower than DRAM. But that’s a thing of the past, possibly. A research team out of Shanghai recently published an article in Science describing a new phase change material that in trials at least produces a much faster and robust memory system. The new material can switch between the gas and crystalline phases in less than one nanosecond, making it about 10x as fast as DRAM.
But don’t expect PCRAM to replace DRAM any time soon. If this kind of memory technology sounds familiar to you it’s probably because in addition to PCRAM there’s also magnetoresistive RAM (MRAM) and resistive RAM (RRAM), all vying to dethrone DRAM. None of these technologies, however, are quite at the point where they can compete effectively with DRAM in terms of both reliability and cost. So, while promising, my guess is that it’ll be a few years yet before we see PCRAM, or some other option, offer a realistic alternative.