CONTACT MLE
We are glad that you preferred to contact us. Please fill our short form and one of our friendly team members will contact you back.


    NPAP-10G Remote Eval.NPAP-25G Remote Eval.


    X
    CONTACT MLE

    MLE NVMe FPGA Full System Stack for AMD Versal AI Edge

    MLE NVMe FPGA Full System Stack for AMD Versal AI Edge TB20241108 Introduction When it comes to high-speed data recording, NVMe Streamer or similar NVMe IP cores are typically required to enable efficient data streaming from FPGAs to NVMe SSDs, as the onboard processors of FPGA boards are often not powerful enough. However, implementing NVMe IP cores presents challenges, particularly in handling file systems, where data read/write paths must be defined and data management gets complicated if multiple storage patterns are required. To address these needs, MLE has developed the NVMe Full System Stack for AMD Versal AI Edge FPGAs. The NVMe Full System Stack enables seamless data streaming between the ARM cores on AMD Versal Edge FPGAs and NVMe SSDs via the PS PCIe Root Port.  Key benefits of the AMD Versal AI Edge NVMe FPGA Full System Stack include: Support for data rates from 1 Gbps to 3.8 Gbps Support dynamic, complex file system Support complex storage patterns, enabling multiple applications to access data through a block device  MLE NVMe FPGA Full System Stack MLE’s NVMe FPGA Full System Stack facilitates connectivity between the ARM Processing System on AMD Versal AI Edge FPGAs with NVMe PCIe SSDs via the Opsero FPGA Drive FMC Gen4 card. MLE ported and validated this

    Recording with NVMe SSDs

    Sustained, High-Speed Data Recording with NVMe SSDs

    MLE has been providing NVMe Streamer, an FPGA-based technology which enables users to directly stream onto NVM Express (NVMe) SSDs data to and from Programmable Logic (PL). The objective behind NVMe Streamer was to provide a solution for high speed data recording (and re-play) without any CPUs involved, either because your FPGA does not have an embedded CPU or because you are looking for a solution with deterministically high read/write bandwidth and performance scalability.