Search

GUC Announces GLink-3D Die-on-Die Interface IP Using TSMC N5, N6 Process - HPCwire

kembaliui.blogspot.com

HSINCHU, Taiwan, May 24, 2021 — Global Unichip Corp. (GUC), an Advanced ASIC Leader, announces GLink-3D die-on-die interface IP using TSMC’s N5 and N6 processes and 3DFabric advanced packaging technology for AI, HPC, and Networking applications.

AI/HPC/Networking memory demand is growing quickly and the SRAM to Logic ratio is also increasing. Logic gains higher density and performance when scaled to N5/N3 process nodes but SRAM scaling from N7 to N5/N3 is moderate. SRAM/Logic disintegration allows the implementation of separate SRAM and Logic at the most efficient process nodes. Layers of CPU and SRAM (Last Level Cache, packet buffers) dies can be assembled over and under interconnect/IO dies using TSMC 3DFabric packaging technology. Such expandable SRAM and modular computing applications are enabled by GUC GLink-3D high bandwidth, low latency, low power, and point-to-multipoint interface between 3D stacked dies. CPUs, SRAMs, Interconnects, and I/Os (SerDes, HBM, DDR) can be implemented in the most efficient process nodes. Different die combinations can be assembled to address different market segments. At boot time, assembled SRAM and CPU dies are identified, unique die IDs are distributed, available memory space and computing resources are defined and a point-to-multipoint GLink-3D interface to the stacked dies is enabled.

TSMC’s 3DFabric SoIC platform technology allows much more efficient connectivity. GLink-3D achieved six times higher bandwidth/area density, six times lower latency and twice lower power consumption than best-in-class 2.5D interface GLink-2.0 (it was taped out in Dec 2020). Several 3D die stacks can be assembled using CoWoS and InFO_oS, interconnected using GLink-2.5D links and combined with HBM memories.

“3D die stacking technology will start a revolution in the way we design HPC, AI and Network Processors. Die-to-die interface is not limited any more to the die boundary, it can be located exactly where processors need to connect to SRAM and additional CPUs. 3DFabric and GLink-3D pave way to the processors of the future, combining huge and scalable processing power with vast, high bandwidth and low latency memory, when every component is implemented using the most efficient process node.” said Igor Elkanovich, CTO of GUC.

GLink-3D key features:

  • Supports TSMC-SoIC stacking of combination of N5 and N6 process nodes
  • Point-to-multipoint interface allows the main die to interface with several stacked dies simultaneously
  • Robust, full duplex 9Tbps traffic per mm2
  • Speed: 5.0 Gbps per lane
  • Extremely low end-to-end latency (< 2ns) with low power design (< 0.2pJ/bit)
  • Single supply voltage 0.75V ± 10%

About GUC

GLOBAL UNICHIP CORP. (GUC) is an Advanced ASIC Leader who provides the semiconductor industry with leading IC implementation and SoC manufacturing services, using advanced process and packaging technology. Based in Hsin-Chu, Taiwan, GUC has developed a global reputation with a presence in China, Europe, Japan, Korea, and North America. GUC is publicly traded on the Taiwan Stock Exchange under the symbol 3443.


Source: GUC

Adblock test (Why?)



"interface" - Google News
May 24, 2021 at 09:44PM
https://ift.tt/3wys0OK

GUC Announces GLink-3D Die-on-Die Interface IP Using TSMC N5, N6 Process - HPCwire
"interface" - Google News
https://ift.tt/2z6joXy
https://ift.tt/2KUD1V2

Bagikan Berita Ini

0 Response to "GUC Announces GLink-3D Die-on-Die Interface IP Using TSMC N5, N6 Process - HPCwire"

Post a Comment

Powered by Blogger.