This page was machine-translated and may differ from the original. View original
SK Hynix Begins Mass Production of Next-Generation Memory "HBM2E"
SK Hynix Mass Production of HBM2E Faster Than Samsung Electronics
Processing 460GB of data per second with 1,024 I/Os
SK Hynix announced on the 2nd that it has entered full-scale mass production of HBM2E. This comes 10 months after developing HBM2E last August, which is faster than Samsung Electronics, which declared mass production within this year.
.jpg)
SK Hynix's HBM2E is a product capable of processing data at 3.6 gigabits per second (Gbps), and can process 460 gigabytes (GB) of data per second through 1,024 I/Os. This is a throughput capable of delivering 124 Full-HD movies (3.7 GB) in one second.
The capacity was also increased by more than double compared to the previous generation to 16GB by vertically connecting eight 16-gigabit (Gb) DRAM chips using TSV (Through Silicon Via) technology.
HBM2E is a memory solution that is considered suitable for next-generation AI systems, such as deep learning accelerators and high-performance computing, which require high computational power.
In addition, it is expected to be adopted in exascale supercomputers (high-performance computing systems capable of performing 100 quadrillion calculations per second) that will lead next-generation basic and applied science research, such as weather change, biomedical science, and space exploration.
Processing 460GB of data per second with 1,024 I/Os
SK Hynix announced on the 2nd that it has entered full-scale mass production of HBM2E. This comes 10 months after developing HBM2E last August, which is faster than Samsung Electronics, which declared mass production within this year.
▲ SK Hynix begins mass production of HBM2E [Photo = SK Hynix]
SK Hynix's HBM2E is a product capable of processing data at 3.6 gigabits per second (Gbps), and can process 460 gigabytes (GB) of data per second through 1,024 I/Os. This is a throughput capable of delivering 124 Full-HD movies (3.7 GB) in one second.
The capacity was also increased by more than double compared to the previous generation to 16GB by vertically connecting eight 16-gigabit (Gb) DRAM chips using TSV (Through Silicon Via) technology.
HBM2E is a memory solution that is considered suitable for next-generation AI systems, such as deep learning accelerators and high-performance computing, which require high computational power.
In addition, it is expected to be adopted in exascale supercomputers (high-performance computing systems capable of performing 100 quadrillion calculations per second) that will lead next-generation basic and applied science research, such as weather change, biomedical science, and space exploration.
본 기사에 대한 정정·반론·추후보도 청구는 보도 청구 안내를, 그간 게재된 보도문은 정정·반론보도 모아보기를 참고해 주세요.














