high bandwidth memory (HBM2/2E) H-SERIES PHY
The PHY IP Cores of EVERYTHING, in smallest size ever
Exceptional Performance
The new H-series delivers performance in AI, ML, Graphics and High Performance Computing where it counts. The numbers below are estimation and are compared to our GDDR6 memory solution(PHY) offerings.
1
%
Less Area
1
%
Less power
1
GB/s
Total Bandwidth
HBM H-series PHY
HBM2
Optimized for area with low latency while balancing area and power
- Test chip to validate size, performance and power
- Available as a Hard Macro delivered as GDS II
- 2.5D Interposer level
- Data rates upto 2400 MB/sec
- pseudo-channel to support up to 16 channels for max. bandwidth
- 8 independent memory channels
- Upto 8H stack HBM2 supported
- 1:1, 1:2, 1:4 PHY SDRAM clock ratios
- DFI 5.0 Interface
- Best for AI, ML Graphics and Networking
- contact@memtech.ai
- 16nm and below
HBM2E
Best in-class performance with a focus on low power and low latency
- Test chip to validate size, performance and power
- Available as a Hard Macro delivered as GDS II
- 2.5D Interposer level
- Data rates upto 3200 MB/sec
- Pseudo-channel to support upto 16 channels for max.bandwidth
- 8 Independent memory channels
- Upto 8H stack HBM2E supported
- 1:2, 1:2, 1:4, 1:8 PHY SDRAM clock ratios
- DFI 5.0 Interface
- Best for High Performance Computing, AI/ML, Graphics & Networking
- contact@memtech.ai
- 16nm & below