Tuesday, March 11th 2014
Samsung Mass Producing Industry's Most Advanced 4 Gb DDR3, Using 20 nm Process
Samsung Electronics Co., Ltd., the world leader in memory technology, today announced that it is mass producing the most advanced DDR3 memory, based on a new 20 nanometer process technology, for use in a wide range of computing applications. Samsung has pushed the envelope of DRAM scaling, while utilizing currently available immersion ArF lithography, in its roll-out of the industry's most advanced 20-nanometer (nm) 4-gigabit (Gb) DDR3 DRAM.
With DRAM memory, where each cell consists of a capacitor and a transistor linked to one another, scaling is more difficult than with NAND Flash memory in which a cell only needs a transistor. To continue scaling for more advanced DRAM, Samsung refined its design and manufacturing technologies and came up with a modified double patterning and atomic layer deposition.
Samsung's modified double patterning technology marks a new milestone, by enabling 20nm DDR3 production using current photolithography equipment and establishing the core technology for the next generation of 10nm-class DRAM production. Samsung also successfully created ultrathin dielectric layers of cell capacitors with an unprecedented uniformity, which has resulted in higher cell performance.
With the new 20nm DDR3 DRAM applying these technologies, Samsung also has improved manufacturing productivity, which is over 30 percent higher than that of the preceding 25 nanometer DDR3, and more than twice that of 30nm-class* DDR3.
In addition, the new 20nm 4Gb DDR3- based modules can save up to 25 percent of the energy consumed by equivalent modules fabricated using the previous 25 nanometer process technology. This improvement provides the basis for delivering the industry's most advanced green IT solutions to global companies.
"Samsung's new energy-efficient 20-nanometer DDR3 DRAM will rapidly expand its market base throughout the IT industry including the PC and mobile markets, quickly moving to mainstream status," said Young-Hyun Jun, executive vice president, memory sales and marketing, Samsung Electronics. "Samsung will continue to deliver next-generation DRAM and green memory solutions ahead of the competition, while contributing to the growth of the global IT market in close cooperation with our major customers."
According to market research data from Gartner, the global DRAM market will grow from $35.6 billion US dollars in 2013 to $37.9 billion US dollars in 2014.
With DRAM memory, where each cell consists of a capacitor and a transistor linked to one another, scaling is more difficult than with NAND Flash memory in which a cell only needs a transistor. To continue scaling for more advanced DRAM, Samsung refined its design and manufacturing technologies and came up with a modified double patterning and atomic layer deposition.
With the new 20nm DDR3 DRAM applying these technologies, Samsung also has improved manufacturing productivity, which is over 30 percent higher than that of the preceding 25 nanometer DDR3, and more than twice that of 30nm-class* DDR3.
In addition, the new 20nm 4Gb DDR3- based modules can save up to 25 percent of the energy consumed by equivalent modules fabricated using the previous 25 nanometer process technology. This improvement provides the basis for delivering the industry's most advanced green IT solutions to global companies.
"Samsung's new energy-efficient 20-nanometer DDR3 DRAM will rapidly expand its market base throughout the IT industry including the PC and mobile markets, quickly moving to mainstream status," said Young-Hyun Jun, executive vice president, memory sales and marketing, Samsung Electronics. "Samsung will continue to deliver next-generation DRAM and green memory solutions ahead of the competition, while contributing to the growth of the global IT market in close cooperation with our major customers."
According to market research data from Gartner, the global DRAM market will grow from $35.6 billion US dollars in 2013 to $37.9 billion US dollars in 2014.
30 Comments on Samsung Mass Producing Industry's Most Advanced 4 Gb DDR3, Using 20 nm Process
First of, GDDR3, GDDR4, and GDDR5 are all based on DDR3. GDDRx is DDR optimized for bandwidth over latency as opposed to standard DDR3 which tends to prefer lower latency over bandwidth. GPUs needs to read and write large chunks of cohesive memory, which is where the bandwidth optimizations of GDDRx come in handy. CPUs on the other hand tend to work with memory that you might not even know where it is until several instructions before the memory access, which is completely ass-backwards from what GPUs are trying to do.
All in all, in cases like the new Playstation, the APU's iGPU can take advantage of that memory. While the CPU might still run well with it, it might not run optimally with it. In the case of APUs, it's a balancing act.We need lower latency memory, weather or not that is achieve but boosting frequency is anyone's guess. I would rather see timings drop for the same speeds we have now. The bigger bus mostlikely isn't going to help a CPU unless your hammering it with a multi-threaded workload, and even still it needs to be memory-bandwidth or memory-latency bound for any speed improves to memory to really make CPUs any faster.
Most benchmarks have shown that faster memory doesn't yield much improvement for CPUs. For APUs it's a different story, but I'm pretty sure that is only because of the iGPU for the same reasons I mentioned earlier in this post.
My comment is the same as what you are saying, we have reached the limit of what 32GB of high speed memory can give us with todays CPU architecture, it takes too long with the higher latency to page through the whole of anything that takes up that much space to utilize the performance of the cores, the number of wait cycles to process that much data is overwhelmingly inefficient, the next logical step then is to move to something that can provide these pages to the CPU faster and unlike our current ideas of serial processing, a move towards more parallel processing where independent workloads are assigned and resources dynamically allocated by the HSA. There is a reason why a little APU can drive 1080 on the PS4, not that I am a console peasant, but I want to know what the real performance of the system as a whole is compared to its closest PC cousin. I bet we would be suprized what it adds to the system.
Secondly, I am not claiming to be right. I was merely challenging his claims. Surely that is not prohibited by these forums, their purpose being, after all, discussion.
Actually, I only have superficial knowledge when it comes to computer hardware, so I can hardly make any claims about how it works.Woah, there. Hold you horses, sir! I am but a jester who does not accept each and every fact I am presented with, without so much as an explanation. I do not think there is anything wrong with that.
-- -- --
Steevo and Aquinus, those last comments were interesting to read. Thanks for shedding some light on the inner workings of memory, guys. :)
Steevo, I hope you won't hold grudge for my poking fun at the fact that you enjoy giving sarcasm, yet do not seem to take very well to receiving it. I didn't mean to upset you or anything, man. ;)