PCIe 5.0 PHY IP for Storage and High-Bandwidth Connection
Overview
M31 PCIe 5.0 PHY IP provides high-performance, multi-lane capability and low power architecture for high-bandwidth applications. The PCIe 5.0 IP supports a complete range of PCIe 5.0 Base applications and is compliant with the PIPE 5.2 specification. The IP integrates high-speed mixed signal circuits to support 32Gbps PCIe 5.0 traffic and is backward compatible with 16 Gbps PCIe 4.0, 8.0Gbps PCIe 3.1, 5.0Gbps PCIe 2.1 and 2.5Gbps PCIe 1.1 data rates. With the support of TX and RX equalization techniques, the PCIe 5.0 IP is designed to meet the requirements of different channel conditions.
Key Features
- Fully compliant with PCI Express Base 5.0 electrical specification
- Compliant with PIPE5.2 (PCIe) specification
- Supports all power saving modes (P0, P0s, P1, P2) defined in the PIPE5.2 specification
- Supports L1 PM Substates with CLKREQ#
- Supports L1 Clock Power Management (CPM) with CLKREQ#
- Supports Separate Refclk Independent SSC (SRIS) architecture
- Accessible register controls allow user-specific optimization of critical parameters (e.g., TXPLL bandwidth, TX de-emphasis level, CDR bandwidth, and EQ strength)
- Supports robust BIST functions for mass production testing
- Programmable differential reference clock pad that can receive or transmit 100MHz reference clock
- Built-in PLL for receiving 25MHz crystal clock and generating 100MHz reference clock for RC applications
- Built-in LDO for analog core power generation, providing easier IC integration and lower BOM cost for SSD applications
- Macro-based cells that can be combined into bifurcation applications
Block Diagram
Technical Specifications
Foundry, Node
6nm, 7nm, 12nm, 16nm
TSMC
Pre-Silicon:
7nm
Silicon Proven: 12nm , 16nm
Silicon Proven: 12nm , 16nm
Related IPs
- UCIe/BoW BlueLynx™ Dual Mode PHY and subsystem IP for chiplet interconnect
- PHY IP for PCIe 6.0 on TSMC N5
- Fast Access Controller – a plug-and-play IP solution for fast embedded Flash Programming and Memory Testing
- Warping Engine IP block for image transformation, HUDs and fish-eye correction
- GDDR6 PHY IP for 12nm
- LPDDR4x/4 PHY IP for 22nm