ADAPTIVE FABRIC ALLOCATION FOR LOCAL AND REMOTE EMERGING MEMORIES BASED PREDICTION SCHEMES

Organization Name

Inventor(s)

ADAPTIVE FABRIC ALLOCATION FOR LOCAL AND REMOTE EMERGING MEMORIES BASED PREDICTION SCHEMES - A simplified explanation of the abstract

This abstract first appeared for US patent application 18371513 titled 'ADAPTIVE FABRIC ALLOCATION FOR LOCAL AND REMOTE EMERGING MEMORIES BASED PREDICTION SCHEMES

Simplified Explanation

The patent application describes methods, apparatus, and systems for adaptive fabric allocation for local and remote emerging memories-based prediction schemes. This involves monitoring memory read and write traffic on interconnect segments, making predictions of expected bandwidths, and dynamically reconfiguring upstream and downstream lanes based on these predictions.

Memory read and write traffic is monitored on interconnect segments with reconfigurable lanes.
Predictions of expected read and write bandwidths are made for the interconnect segments.
Upstream and downstream lanes are dynamically reconfigured based on the expected bandwidths.
Interconnect segments include Compute Exchange Link (CXL) flex buses and memory channels for local memory, and fabric links for remote memory.
Management messages or telemetry information are used to provide expected bandwidths for local and remote memory implementations.

Potential Applications

This technology could be applied in data centers, high-performance computing systems, and cloud computing environments to optimize memory transfers and improve overall system performance.

Problems Solved

This technology addresses the challenge of efficiently managing memory read and write traffic on interconnect segments, especially in systems with local and remote memory implementations.

Benefits

The dynamic reconfiguration of upstream and downstream lanes based on predicted bandwidths can lead to improved memory transfer efficiency, reduced latency, and enhanced system performance.

Potential Commercial Applications

"Optimizing Memory Transfers for Data Centers and High-Performance Computing Systems"

Unanswered Questions

=== How does this technology impact power consumption in memory transfer operations? === Are there any limitations to the scalability of this adaptive fabric allocation approach in large-scale systems?

Original Abstract Submitted

Methods, apparatus and systems for adaptive fabric allocation for local and remote emerging memories-based prediction schemes. In conjunction with performing memory transfers between a compute host and memory device connected via one or more interconnect segments, memory read and write traffic is monitored for at least one interconnect segment having reconfigurable upstream lanes and downstream lanes. Predictions of expected read and write bandwidths for the at least one interconnect segment are then made. Based on the expected read and write bandwidths, the upstream lanes and downstream lanes are dynamically reconfigured. The interconnect segments include interconnect links such as Compute Exchange Link (CXL) flex buses and memory channels for local memory implementations, and fabric links for remote memory implementations. For local memory, management messages may be used to provide telemetry information containing the expected read and write bandwidths. For remote memory, telemetry information is provided to a fabric management component that is used to dynamically reconfigure one or more fabric links.

18371513. ADAPTIVE FABRIC ALLOCATION FOR LOCAL AND REMOTE EMERGING MEMORIES BASED PREDICTION SCHEMES simplified abstract (Intel Corporation)

Contents

ADAPTIVE FABRIC ALLOCATION FOR LOCAL AND REMOTE EMERGING MEMORIES BASED PREDICTION SCHEMES

Organization Name

Inventor(s)