From the Dell booth at Supercomputing SC25 in St. Louis, this walkthrough focuses on a compact AI platform labeled “XC9785” on-site (likely the PowerEdge XE9785L naming), pairing dual-socket AMD EPYC CPUs with eight AMD Instinct MI355X accelerators for both training and inference. MI355X targets dense, memory-forward AI/HPC, with up to 288GB HBM3E per GPU, roughly 8TB/s memory bandwidth, and expanded low-precision formats (MXFP6/MXFP4) aimed at better throughput-per-watt when models are memory-bound rather than ALU-bound. https://www.dell.com/en-us/shop/ipovw/poweredge-xe9785l
—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—
A useful detail is the bare GPU board shown without cold plates or heat sinks, which makes the packaging constraints obvious: power delivery, signal integrity, and coolant routing all sit centimeters apart. Dell frames the XE9785L as a 3U direct-liquid-cooled design intended for OV3 (OCP Open Rack v3) deployments in IR7000/IR9000 racks, where shared manifolds and blind-mate interfaces enable higher rack density while keeping service steps predictable in a packed rack.
Next to the accelerator chassis, they point to storage-heavy compute nodes built around AMD’s latest CPUs: up to 24 direct-attached U.2 NVMe drives, with an on-the-spot claim of roughly 300Gbps of storage throughput, plus PCIe Gen5 full-height expansion for high-speed networking. In practice, that maps to the less glamorous but critical parts of AI: fast local NVMe for checkpoints, data staging, and feature/vector stores, then high-bandwidth I/O for scale-out fabrics and NVMe-oF style data paths.
The clearest spec callout is power: MI355X can run up to about 1.4kW per GPU in a liquid-cooled configuration, which is why rack-level cooling architecture, monitoring, and leak-risk mitigation start to matter as much as raw accelerator count. The takeaway is a rack-ready building block for converged HPC and AI workflows—dense GPU compute, PCIe Gen5 I/O, and NVMe-rich nodes designed to keep data moving as aggressively as matrix math in deployment.
I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga
This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK
Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁
Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY
