Dolphin’s hardened DDR4/3/2 SDRAM PHY and LPDDR4/3/2 SDRAM PHY IP is a silicon-proven, Combo PHY supporting speeds up to 4266 Mbps. It is fully compliant with the DFI 4.0 specification, and features include slew rate control, per-bit de-skew, gate training, read and write leveling and built-in self test (BIST).
In addition, our PHY IP is optimized to provide a complete solution when combined with Dolphin’s DDRx and LPDDRx SDRAM Memory Controller IP.
Download Product Overview
TSMC
Technology Levels
The hardened PHY macros are currently available in the following technology levels:
Dolphin Technology offers high performance DDR4/3/2 SDRAM and LPDDR4/3/2 SDRAM Memory Controller IP across a broad range of process technologies.
These Memory Controllers are fully compliant with the DFI 4.0 specification, support speeds up to 4266 Mpbs, and are optimized to provide a comprehensive solution when combined with Dolphin’s DDR PHY IP.
Download Product Overview
Features |
Design Status |
- Supports JEDEC Standard DDR4/3/2 and LPDDR4/3/2 SDRAM
- DFI 4.0 Compliant Interface with 1:1 (Matching), 1:2 and 1:4 Frequency Ratios
- Supports speeds of up to 4266 Mbps (2133 MHz) LPDDR4
- Optimized to provide a complete solution along with the Dolphin Technology DDR PHY solution
- Built-in Gate Training, Read/Write Leveling, and VREF Training
- Multi-Port Configurable AXI4 Interface with QoS Signaling
- Single AXI4-Lite Programming Interface
- Multi-port arbitration engine with programmable dynamic priority algorithm ensure high performance and low-latency.
- Pipeline Option for Frequency versus Latency Tradeoff
- Fully configurable for various performances and requirements, ensuring maximum performance for different system environments.
- FPGA portable. Compatible with Xilinx PHY and Altera PHY.
- Available with BFM verification suite.
- Support for AXI4 Dynamic QoS Signaling for Non-Blocking Communications
- Support for Low-Latency Bypass Ports/Channels
- Advanced Dynamic QoS Support Based on the Queuing Theory and Traffic Hysteresis
- Forward Priority Propagation for Queued Transactions
- Traffic-Configurable Address Mapping Scheme with Two Column Segments
- Run-time Configurable Timing Parameters and Memory Settings
- Intelligent Bank Management and Auto-Precharge Scheduling
- Intelligent Traffic Direction Aggregation and Switching
- Programmable and Dynamic Auto-refresh Scheduling
- Proprietary Non-Blocking Priority-Based DDRx/LPDDRx SDRAM Control and Data Buses Access Algorithm
- Real-Time QoS Capability
- Multi-Channel Configuration (Up to 4)
- Multi-Rank Configuration (Up to 4)
- Support for Out-of-Order Memory Access
|
Data Sheets are available under NDA
REQUEST NDA
|