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

source https://www.youtube.com/watch?v=UoBGrKwEQUI

SiliconCore brings a different take on immersive visualization: a fine-pixel-pitch direct-view LED (dvLED) wall used as a stereoscopic “CAVE-style” display, tuned for high frame-rate scientific and engineering data. Instead of putting each viewer in a headset, the wall renders a shared 3D scene with low latency and high luminance, so teams can review volumetric datasets and simulation outputs together. https://www.silicon-core.com/

—
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.
—

In the demo, the wall runs 3D at 120Hz (effectively 60Hz per eye), with head-tracking so the image updates to your exact viewpoint in real time. SiliconCore says the platform can reach 240Hz, which opens up dual-viewer modes where two tracked users can see the same dataset from different perspectives on the same screen, with less eye/neck strain than typical headset workflows here.

A practical detail is mobility: fold-in “wings” let the display fit through elevators and doors, and a lift plus flight case makes it shippable between labs, floors, or events. Setup is described as largely plug-and-play on the display side, drawing about 500W from a standard outlet, delivering up to ~1,000 nit brightness (with an option for ~1,800 nit using two power feeds) for bright indoor 3D today.

The tracking stack comes from ART (Germany), using IR cameras and reflective markers mounted on active 3D glasses so the renderer can correct perspective continuously as you move. This kind of tracked stereo display is a good fit for HPC centers like HLRS, where collaborative inspection of complex geometry, CFD, digital twins, or medical/scientific volumes benefits from a large shared canvas rather than isolated VR work.

Under the hood, SiliconCore emphasizes vertical control of the LED pipeline (driver ICs, communication, controller/receiver chain), which helps them push high refresh, low power, and deterministic timing that matter for flicker-free stereo and multi-viewer sync. The transcript also hints at real deployments like a NIST CAVE with matched tiles on floor, wall, and ceiling, where identical spare parts and consistent image characteristics simplify maintenance while keeping the 3D workflow usable at close viewing distance edge.

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

source https://www.youtube.com/watch?v=M_-seSm6mzQ

Congatec gives a compact tour of how its computer-on-module portfolio is pushing more compute into smaller footprints without losing the I/O you need for real products. The headline is COM-HPC Mini on Intel 13th Gen Raptor Lake-P, aimed at edge systems that need PCIe bandwidth, modern display links, and deterministic Ethernet rather than “embedded = slow”. https://www.congatec.com/

—
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.
—

COM-HPC Mini lands at 95 x 70 mm and still exposes a serious platform envelope: onboard LPDDR5x, onboard NVMe, PCIe Gen4 plus additional PCIe, SATA, eDP/DP, USB 3.x and USB4, and dual 2.5GbE with TSN capability depending on the variant. The module plugs into an application carrier (shown on a Connect Tech example) where your RAM, connectors, and domain-specific I/O live, and Congatec frames cooling as a first-class design axis with passive, active, and heat-pipe options here.

The booth walk also touches COM Express Type 6 (including the conga-TC700 family on Intel Core Ultra “Meteor Lake” with an integrated NPU-class accelerator) and ruggedized designs like MA-S10 and TC675R that target industrial temperature, vibration, and long-life deployment. On the low-power side, the SMARC lineup spans Arm-based modules such as TI Jacinto parts and the conga-SMX95 on NXP i.MX95, pairing an eIQ Neutron NPU with EdgeLock security for vision, HMI, and control at the rugged edge.

A practical demo ties the stack together: a two-axis balancing table where each motor is driven by its own module and synchronized in real time over Time-Sensitive Networking, while vision inference tracks the moving ball. On the Core Ultra platform, a real-time hypervisor partitions CPU cores into multiple OS instances so motion control, OpenVINO-based object tracking, and UI tasks can coexist with clean timing boundaries, filmed at Embedded World North America 2025 in Anaheim for this demo.

The broader message is modularity as lifecycle insurance: keep the carrier and software contracts stable, swap compute generations (Raptor Lake-P to Core Ultra), and qualify thermals, security (TPM 2.0), and watchdog/health monitoring once, then iterate. Congatec also notes tighter cooperation with Kontron around the JUMPtec module business, expanding the combined x86/Arm module catalog and manufacturing capacity for OEM roadmaps that need long availability, with room to grow more.

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

source https://www.youtube.com/watch?v=XlMZnL4fhdE

Kingston’s embedded team walks through what “industrial memory” really means when you’re designing single-board computers and edge devices: choosing the right DRAM + NAND mix for capacity, endurance, cost, and how long a product must survive in the field before a service visit even exists. The conversation stays practical around design-in eMMC and DRAM that ships with controlled hardware (stable part numbers and BOM discipline), so integrators can qualify once and manufacture for years with fewer surprises. https://www.kingston.com/en/solutions/embedded-and-industrial

—
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 big part of that promise is qualification work with partners: validating Kingston DRAM and eMMC on real boards (often alongside NXP and other SoCs), then backing it up with signal-integrity testing and protocol analysis in the lab. That’s where the unglamorous details live—high-speed DDR routing margins, timing closure, interface compliance, and catching edge-case behavior before a device ever leaves the bench for a customer site on a board.

On the packaging side, the video maps the “soldered-down menu” you see in embedded designs: eMMC plus discrete DRAM, integrated eMCP (eMMC + LPDDR in one footprint), and ePoP (package-on-package stacked directly on top of a compatible SoC to save PCB area and shorten interconnect). It also explains why Kingston often leans into legacy-friendly roadmaps—low-capacity eMMC like 4GB/8GB and DDR3/DDR4—because long-term availability and revision control can matter more than chasing peak bandwidth in the field.

Edge AI brings the memory hierarchy question into focus: DDR bandwidth and capacity for fast working sets, and NAND (eMMC today, UFS in some next-step designs) for persistent data, model files, and logs. The nuance is endurance math—program/erase limits, write amplification, wear leveling, and using JEDEC health indicators (like eMMC lifetime estimation fields) to plan maintenance instead of guessing. It also touches on why removable media can be risky in industrial settings (vibration, contact reliability, and inconsistent flash behavior), even if it looks cheaper on a spec sheet at trade.

Filmed at Embedded World North America 2025, it’s essentially a design review disguised as a booth chat: how to size memory for video pipelines, sensor logging, and on-device inference, and how to de-risk it with validation workflows rather than late-stage debugging. If you’re building anything from a BeagleBone-class SBC to an industrial HMI or gateway, this is a compact look at the engineering tradeoffs that decide whether a device stays stable out there.

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

source https://www.youtube.com/watch?v=XNRC93iI17M

3M’s electronics materials story is less about any single component and more about making interconnect and protection behave predictably under real constraints: signal integrity, EMI/EMC, heat, vibration, and long service life. In this booth walk-through, Manuel Sen (business manager, electronics for 3M AMIA) connects the dots between cabling, shielding, insulation, and thermal paths—exactly the stuff that decides whether a system passes compliance, stays stable at bandwidth, and survives the factory. https://www.3m.com/3M/en_US/electronics-components-us/

—
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.
—

On the connectivity side, the focus starts with flat cable families built for different environments: high-temperature constructions, halogen-free options, PVC, and twisted-pair variants to reduce noise coupling. Board-to-board robustness and IDC-based sensor connectivity show up as practical answers to rising sensor counts, where discrete wires must reliably reach multiple sensing points without turning assembly into a failure hotspot. This segment is filmed on the SPS show floor in Nuremberg, with a clear emphasis on industrial automation-grade cabling that stays stable over time, not just at first power-up.

Then the conversation shifts to adhesive and tape engineering for electronics, framed in three electrical/thermal roles. Electrically conductive tapes create a controlled grounding path for shielding, helping dump interference to ground instead of letting it pollute sensitive lines. Electrically insulating tapes (including polyester and polyimide constructions) target high temperature plus dielectric performance, while thermally conductive tapes act as a thermal interface to move heat away from dense electronics where airflow and heatsink geometry hit their limit. Converting rolls into die-cut parts for mass production also matters here, as does surface prep—clean, dust-free application is what keeps adhesion stable under vibration and heat.

For high-speed links, 3M highlights thin, heavily shielded TwinAx-style cabling designed to fold and route through tight mechanical envelopes while keeping signal behavior consistent. That “bend it around corners without losing performance” theme is familiar in hyperscale data centers, but it maps cleanly to industrial systems that now carry higher bandwidth across smaller spaces. It also aligns with modern I/O needs where keeping losses low can reduce dependence on retimers and still target interfaces like PCIe Gen5-class performance over short internal runs.

Fiber and vision cabling round out the picture. Expanded Beam Optical (EBO) shifts the weak point of fiber—contamination at the interface—by using an expanded-beam, dust-tolerant approach suited to rugged or dirty environments, including defense-style deployments. For machine vision, the discussion points to standardized ecosystems like USB3 Vision and CoaXPress 2.0, where cable length, flex life, and torsion resistance decide uptime in moving-camera rigs; 3M’s industrial camera cable assemblies are positioned around repeated-bend durability and stable transmission over longer runs. The through-line stays consistent: material science choices that make cables and tapes behave reliably inside a machine, year after year, with EMI, heat, and motion all accounted for in one stack of tape.

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 )

source https://www.youtube.com/watch?v=lkhtoxG3yOQ

Kontron walks through how its embedded motherboard roadmap is shifting toward on-board AI acceleration, especially with Intel Core Ultra (Meteor Lake-U) platforms that include an integrated NPU (around 11 TOPS) for edge inference. The core message is practical: keep x86 form factors familiar for industrial OEMs, but move compute-heavy vision and signal-processing workloads away from CPU-only execution when latency and power-per-task start to matter. https://www.kontron.com/

—
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 key demo point is the software path: Intel OpenVINO running on the chip’s integrated NPU, showing the “same model, different engine” comparison between CPU and NPU. That framing matters for embedded teams because it ties together model deployment, acceleration backends, and real-world UX: smoother real-time inference, fewer dropped frames, and more predictable thermals on compact systems like Mini-STX boards.

Beyond Meteor Lake, they position a broad catalog that scales from cost-focused Intel Celeron SoC designs up through newer “Lake” generations, plus AMD-based options such as Ryzen Embedded V10000 and R2000 families. In the same conversation, they also point to ARM-oriented work, including customer-specific carrier boards around Raspberry Pi/SoM-style modules, which is often the fastest route to tailor I/O, power domains, and mechanical constraints without rebuilding the whole compute stack.

Filmed at Embedded World North America 2025, the booth tour then jumps to the higher end of the lineup: industrial server motherboards with onboard 10GbE and a Kontron-developed BMC for remote management (KVM-style access, monitoring, lifecycle operations) without relying on extra third-party tooling. They also name-check Intel Xeon Scalable “Granite Rapids” as the kind of platform they’re targeting when embedded requirements start to look like edge-cloud density, but still need industrial longevity and controlled BOM strategy.

Finally, there’s a strong manufacturing and logistics angle: production in Europe, expanded with a second factory in Slovenia featuring multiple SMT lines and in-line inspection capabilities, plus work tied to 5G broadband programs. For North American customers, they emphasize local presence via a San Diego office and warehouse, aiming for quick sample turnaround and short lead times for off-the-shelf boards, which is often what makes an evaluation phase actually move.

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

source https://www.youtube.com/watch?v=11aNgq0AERU

Realtek demos how to bring Matter interoperability to existing Android TV set-top boxes: add Thread via a small USB dongle, run a TV-side controller app, and a legacy STB starts behaving like a smart-home control point rather than only a video endpoint. The pitch is architectural, not just a gadget: Matter is the application layer, while transport can be Wi-Fi, Ethernet, or Thread, so OEMs can upgrade an installed base without redesigning the whole platform. https://www.realtek.com/

—
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.
—

Technically, the key move is the Thread path. Thread is an IPv6 mesh over IEEE 802.15.4, and the dongle effectively supplies that radio plus the border-router role so the STB can speak to low-power endpoints. Realtek’s RTL8762G/RTL877xG family is positioned for multi-protocol designs (BLE 5.3, Thread/Matter, Zigbee, plus interfaces like USB 2.0 HS and Ethernet), which is the kind of plumbing you need when a TV box becomes a controller rather than a passive client device.

On the device side, they show a compact endpoint that pairs Matter support with a touch LCD UI on the RTL8772GWP, so an accessory can be both a physical control surface and a networked node. In the booth it just flips LEDs, but the same pattern maps to sensors, locks, and small appliances where you want local state, local interaction, and predictable control even when “cloud first” is not an option.

The tour then pivots into latency-sensitive peripherals: ultra-low-latency wireless audio for TWS earphones, and input devices (mouse, keyboard, joystick) targeting up to an 8K report rate, where 8,000 updates per second implies ~0.125 ms sampling granularity in the best case. Realtek’s ecosystem here is about end-to-end latency budgets across radio, USB scheduling, firmware, and host drivers, and the fast booth-walkthrough style echoes the interview pacing you also see around Web Summit Lisbon 2025.

There’s also a “doorbell camera to smartwatch” concept: a live stream moves over 2.4 GHz Wi-Fi to a server (cloud in production, local PC in the demo) and then to a wearable client, illustrating how Realtek platforms can span capture, transport, and playback. Across the demos, the consistent theme is providing silicon plus SDKs and reference stacks so partners can ship connected products without rebuilding every layer, on a busy show floor.

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

source https://www.youtube.com/watch?v=cGLRkLZ5bEs

RIKEN’s FugakuNEXT program is framed as a 2030 deployment that builds on what Fugaku (deployed in 2020) did well for large-scale science: strong FP64 throughput for conventional simulation, and a software ecosystem tuned for production HPC. The pitch here is that the next decade’s workload mix is shifting toward hybrid AI + simulation, so the system architecture needs to treat AI performance as a first-class goal rather than an add-on. https://www.fugaku.riken.jp/

—
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.
—

On the CPU side, the discussion centers on Fujitsu’s ARM-based roadmap (the transcript references the current-generation CPU and its move to leading-edge process nodes) and the idea of a richer SoC that could also integrate an NPU alongside general-purpose cores. The key technical tension is familiar in exascale planning: keep strong FP64 for numerics, while also improving throughput per watt for AI-relevant kernels, memory bandwidth pressure, and node-level balance, without locking the platform into a single workload style.

For accelerators, they explicitly assume GPUs will be in the loop, citing NVIDIA’s recent high-end generations (Blackwell) and the following roadmap (Rubin), with the expectation that FugakuNEXT will land after that point. In practical terms, that implies heavy reliance on mixed precision (FP16/BF16/FP8), tensor-style matrix engines, and high-bandwidth memory, while still keeping a path for double-precision science and deterministic numerics when the application demands it at scale.

A lot hinges on the CPU–GPU connection: PCI Express is named as an option, alongside an “NVLink-class” approach, but the bigger requirement is cache-coherent behavior between CPU memory and GPU memory so programmers can treat the machine less like loosely coupled devices and more like a coherent heterogeneous node. They also point to the program mechanics behind that goal—testbeds, coding and code-design strategy, evaluation loops, and a timeline that aligns facility build-out with system integration and software readiness to make it.

What comes through is an HPC roadmap that assumes AI isn’t separate from simulation anymore: think surrogate models inside solvers, learned preconditioners, data assimilation, and large-scale training/inference running next to classic MPI-heavy workloads. The project is presented as a coordinated effort with leadership and support teams, plus an expanded datacenter footprint to host the next platform, reflecting how much of “supercomputer design” is now co-design across silicon, interconnect, cooling, compilers, and production operations for reliability at scale today.

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

source https://www.youtube.com/watch?v=C_lyh2gMTV8

Allison Maddox from Dell Technologies walks through a preview of the upcoming IR7000 integrated rack: a rack-scale platform that bundles compute sleds, a direct-liquid-cooled networking switch, an integrated rack controller, a power shelf, and a 4U rack cooling distribution unit (RCDU). The emphasis is on OV3/ORv3-style disaggregated power plus shared plumbing, so the rack becomes the repeatable building block for AI and HPC pods, rather than a one-off integration project. https://www.dell.com/en-us/shop/storage-servers-and-networking-for-business/sf/integrated-rack-scalable-systems

—
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 key engineering point is risk management for direct liquid cooling (DLC). Dell’s integrated rack controller is shown doing rack-level leak detection using a sensing rope routed through strategic zones, aiming to detect coolant escapes outside a node before they reach expensive GPUs, switches, or the power shelf. In practice this turns DLC from “pipes and hope” into instrumented infrastructure, with rack-scale telemetry, alarms, and clearer fault isolation for risk.

The OV3 mechanics are designed for serviceability at density: pulling a compute system out disconnects it from both the busbar and the coolant manifold, and pushing it back in blind-mate reconnects power and liquid without manual fittings. That matters when racks are delivered as tall, high-density assemblies (the full IR7000 racks on the booth floor are in the 44U–50U class), because maintenance time, connector wear, and human error quickly dominate uptime math at dock.

The cooling argument is straightforward thermodynamics: high-TDP CPUs and accelerator GPUs create heat flux that air cooling struggles to move efficiently, so cold plates and circulating fluid carry heat to a CDU/RCDU that pumps, conditions, and stabilizes the loop. The intended buyers—AI/HPC customers, cloud service providers, finance, and large enterprises—are looking for standardized rack patterns, predictable cooling envelopes, and operational controls that scale, which is why liquid cooling is so visible in this Supercomputing SC25 St. Louis walkthrough there.

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

source https://www.youtube.com/watch?v=RUIRzsf9oyc

Diamond Technologies walks through a compact portfolio of OEM data-collection building blocks: embedded 1D/2D barcode scan engines and RFID modules aimed at instrument makers who need reliable decode, simple mechanical integration, and predictable interfaces. The demo highlights a small embedded barcode reader built around a VGA imager, with options that move up to 1.3-megapixel wide-angle imaging when you need a larger field of view without growing the optical path. https://www.diamondt.com/

—
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.
—

On the RFID side, the module shown uses a dual-stack antenna approach so a single footprint can cover both LF and HF bands, paired with an Ethernet interface for straightforward network attachment. In practice that combination maps well to asset tracking, consumables authentication, and patient/sample ID workflows where you may need both proximity cards and tag formats, plus deterministic connectivity at the edge.

The conversation then widens to Beacon EmbeddedWorks (the parent brand after the acquisition) and its system-on-module strategy: Open Standard Module (OSM) form factors that let teams swap compute while keeping a consistent carrier design. Examples called out include NXP i.MX8M Plus and i.MX9 options, plus an STM32MP2 alternative, all positioned to reduce bring-up time, accelerate BSP work, and keep peripheral routing sane for production hardware shown on the booth floor.

A notable highlight is a tiny wearable-class Qualcomm platform module (the W5 family) that packages CPU plus GPU/NPU/VPU capabilities, multimedia encode/decode, eMMC, and LPDDR into a power-bounded footprint. The live demo pairs this with rail-level power telemetry (Beacon’s Watson monitoring) to show how toggling radios like Bluetooth/BLE impacts total draw, which is exactly the kind of feedback loop you want when designing always-on, battery-limited devices.

Finally, the tour touches real edge AI integration: a PCIe vision-inspection add-in running color classification and defect detection (the “Skittles” example) under SOM control, illustrating how inference workloads can be split between accelerator hardware and an embedded controller. The broader theme is long lifecycle support for regulated markets—defense, aerospace, medical, and FDA-locked designs—where stable SOM availability, software maintenance, and reproducible manufacturing matter as much as raw compute for the product team.

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

source https://www.youtube.com/watch?v=61fIsVvVATQ

Dell Technologies walks through its IR7000 Integrated Rack approach: a factory-tested, preconfigured 50U rack-scale platform (21-inch, OCP ORv3-oriented) built to host up to 36 compute nodes plus eight power supplies, with separate switching for the high-speed data fabric and the management network. The goal is “rack as a system” modularity, so the same chassis can be populated with NVIDIA Grace Blackwell hardware such as GB200 NVL4, future GB300-class payloads, or CPU-only nodes for more conventional simulation workload. https://www.dell.com/en-us/shop/dell-integrated-rack-scalable-systems/sf/integrated-rack-scalable-systems

—
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.
—

Cooling is treated as a first-class constraint. IR7000 supports direct liquid cooling with cold plates on CPUs and GPUs fed by either an in-rack CDU (coolant distribution unit) or an in-row CDU, while a rear-door heat exchanger interfaces to facility water for the room loop. That two-step scheme targets near-total heat capture at the rack, reducing dependence on cold data-hall air and shifting planning toward coolant supply temperature, flow rate, pressure drop, monitoring, and serviceability in water.

Power density is the other theme that dominates buyer questions. The demo references racks running around 264 kW, with a design path toward roughly 480 kW as accelerator sleds, NVLink-class domains, and higher-TDP CPUs raise per-node draw. In practice that means sizing busway and PDU capacity, choosing a redundancy model, validating breaker and cable derating, and ensuring cooling distribution scales across rows so new racks can be commissioned with predictable electrical and thermal capacity.

Recorded at Supercomputing SC25 in St. Louis, the message is less about one GPU SKU and more about making rack-scale AI infrastructure repeatable. When compute, networking, and liquid cooling are integrated as one deployable unit, operators can focus on topology, scheduling, and uptime targets rather than rebuilding the data center for each new generation of accelerated compute in HPC.

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

source https://www.youtube.com/watch?v=mx9o2NwRIGU

This talk follows a summer SULI internship at Oak Ridge National Laboratory evaluating whether Mojo can reduce the “two-language tax” in GPU computing: keep Python-level ergonomics while staying close to CUDA/HIP performance, and stay portable when moving between NVIDIA and AMD hardware. The focus is practical performance portability using real kernels on NVIDIA H100 and AMD MI300A. https://www.modular.com/mojo

—
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.
—

Mojo is built on MLIR (LLVM’s multi-level IR) and keeps a Python-like surface syntax while targeting systems-level codegen, memory safety, and tight control over types, layouts, and parallel execution. In the current stack, GPU work is still fairly low level: you write explicit kernels, reason about thread/block structure, and manage host↔device memory and synchronization through Mojo’s GPU APIs (often used inside MAX custom operations), so “portable” does not mean automatic.

The poster ports four scientific workloads, split into memory-bound and compute-bound behavior: BabelStream-style vector ops and a 7-point stencil over a 3D buffer, plus miniBUDE and a Hartree-Fock kernel with multiple atomic operations. For the bandwidth-driven kernels, Mojo reaches competitive memory throughput: on H100 it can beat a CUDA baseline on several vector routines, while dot is harder to match because the CUDA version relies on device-specific tuning. On MI300A, Mojo largely tracks C++/HIP for these memory-bound kernels, with similar bandwidth per routine.

Compute-bound kernels are where compiler maturity shows up. In miniBUDE, performance sits between unoptimized and heavily optimized CUDA as per-thread work (PPWI) rises, suggesting Mojo still needs more aggressive fast-math and scheduling for arithmetic intensity. For Hartree-Fock, atomics can look strong on small H100 cases but degrade sharply at the largest size; on MI300A, atomics may be far slower and the biggest test can fail, highlighting gaps in atomic codegen and runtime behavior at scale.

Filmed at Supercomputing SC25 in St. Louis, the takeaway is that Mojo already looks credible for memory-bound HPC and AI-adjacent kernels where bandwidth dominates, while compute-heavy and atomic-heavy code still needs iteration. The follow-on plan mentioned here—building a BLAS-style library in Mojo while benchmarking best-case paths on NVIDIA and AMD—maps well to how performance-portable stacks usually mature, GPU.

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

source https://www.youtube.com/watch?v=HKqeMg9NZ8s

ENERZAi shows how far you can push on-device AI when memory bandwidth and DRAM size are the real bottleneck. The core idea is extreme low-bit quantization plus hardware-aware graph and kernel optimization, so models stay usable on CPUs/NPUs instead of needing a GPU server or cloud round-trip. In this demo, the focus is practical edge inference: smaller activation footprints, faster decode loops, and tight runtimes that still keep accuracy within a tolerable delta. https://enerzai.com/

—
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.
—

On Synaptics Astra (Astra Machina), they compare a “normal” Whisper deployment against their optimized Whisper variant: the optimized build cuts memory use by about 4x and reduces latency by roughly 2x, with only a small reported accuracy drop. The workflow isn’t just post-training compression; it’s quantization-aware training that explicitly models low-bit error, then compiles for the target using their Optimium inference backend so the operator graph, scheduling, and kernels match the SoC profile there.

They also show a speech-to-vision pipeline where Whisper transcribes a spoken command and triggers a YOLO detector on a Renesas RZ/V2 board. The interesting bit is heterogeneous compute: Whisper runs on the Arm Cortex-A CPU, while YOLO is offloaded to the DRP-AI accelerator, hitting a real-time 30 fps inference loop even if the demo UI takes longer to draw overlays. It’s a clean example of “voice as a control plane” for low-latency perception at the edge.

A second setup uses a Raspberry Pi to control Philips smart lighting by voice, chaining Whisper with a lightweight language/intent model that turns text into device actions. They note this isn’t just a lab trick: similar voice pipelines have been commercialized in IPTV set-top boxes (commands like channel control) and deployed at scale in Korea, which is a strong signal about footprint, cost, and reliability constraints being met today.

The final demo extends the same pattern to live captions and translation: Whisper generates subtitles from a CNBC stream, then a translation model renders Spanish in near real time, again on edge-class hardware. The conversation is filmed at Embedded World North America 2025, and it fits a broader theme you see across recent conference coverage: compress the model, optimize the runtime, and keep data local so latency, privacy, and bandwidth stay predictable.

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

source https://www.youtube.com/watch?v=5pGDGSMI6yU

RIKEN discusses a 2023–2028 effort to make quantum processors usable alongside national supercomputers by connecting QPUs and HPC through a shared software layer, not a bespoke hardware interconnect. The aim is a hybrid execution model: classical nodes do orchestration, data prep, and iterative optimization, while quantum backends are invoked only for specific kernels where they may help even in the error-prone NISQ era of compute. https://www.r-ccs.riken.jp/en/

—
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 key detail is heterogeneity by design. RIKEN is linking a superconducting IBM system (IBM Quantum System Two powered by the 156-qubit Heron processor) with a trapped-ion machine from Quantinuum (Reimei / H-Series), and treating both as callable resources. The HPC side centers on Fugaku in Kobe, with connectivity intended to extend to other Japanese university clusters, so one application can target multiple quantum and classical backends through one software stack.

The practical question becomes where hybrid actually pays off. Many steps stay classical—linear algebra, tensor contractions, gradient updates, Monte-Carlo control, pre/post-processing—while quantum calls are explored for sampling, quantum simulation, and certain optimization motifs (VQE, QAOA, Hamiltonian dynamics, probabilistic inference). That shifts attention to compilation/transpilation, circuit batching to manage latency, error mitigation, and careful benchmarking so performance claims stay grounded.

To future-proof the work, the project emphasizes an interface/protocol layer (often described as an SQC interface) that standardizes job submission, QPU selection, scheduling, authentication, and telemetry across both HPC schedulers and quantum control planes. RIKEN also mentions a test-user program to surface integration issues early—API semantics, queueing behavior, data movement, reproducibility, and measurement statistics—so future supercomputers and future quantum machines can plug in with minimal rework at scale.

Filmed at Supercomputing SC25 in St. Louis, the chat also hints at the talent pipeline: students stopping by, and the growing need for people who understand compiler stacks, numerical methods, cryogenic constraints, and cluster operations in one mental model. The hybrid platform sits alongside Japan’s longer roadmap toward a post-Fugaku system around 2030, where tighter quantum-HPC coupling could widen the range of experiments that feel computationally in scope.

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

source https://www.youtube.com/watch?v=Sm0sxR7P4eY

Jonathan Wright outlines AlmaLinux as a RHEL-compatible community distro that fills the “classic CentOS” role: a downstream rebuild aimed at stable enterprise and HPC fleets, but with its own community decisions. He mentions the recent 9.7 release and the push toward 10.1, framed as Linux for everything from a basement web server to large research labs. https://almalinux.org/

—
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 key theme is the post-2023 Red Hat/CentOS landscape: Fedora remains the upstream incubator, CentOS Stream now tracks the next RHEL minor release, and AlmaLinux follows as a downstream rebuild for teams that want reproducible baselines, long lifecycle updates, and predictable ABI/API behavior across nodes. For cluster admins, that translates into fewer surprises in toolchains, drivers, and userland dependencies when scaling workloads or rebuilding images today.

The most technical detour is Btrfs (often said as “butterfs”) arriving in AlmaLinux 10.1 even though it is not shipped in current RHEL builds. Wright frames it as ZFS-adjacent: checksumming, inline compression, copy-on-write, reflinks, subvolumes, and snapshot workflows that pair well with rollback, golden images, and fast cloning; scrub plus send/receive also become practical primitives. Filmed on the Supercomputing SC25 floor in St. Louis, the demo runs AlmaLinux on an Ampere ARM server while a live graph ramps CPU load up and down, illustrating performance-per-watt under load.

On governance, he emphasizes why AlmaLinux exists: it started with CloudLinux and former CentOS users, then transitioned into an independent nonprofit foundation with community elections, a board, a Technical Steering Committee, and SIG-style working groups. Sponsors range from hyperscalers (AWS, Azure) to silicon vendors (AMD) and research institutions like CERN, and that mix is presented less as marketing and more as a resilience route.

The takeaway is that AlmaLinux is positioning itself as a practical landing zone for organizations navigating CentOS Stream changes while still adopting modern storage and emerging CPU architectures. If you care about fleet management, secure update pipelines, and consistent OS images across x86_64 and ARM64, this conversation connects the ecosystem diagram to the day-to-day reality of keeping clusters and edge racks reliable while the Linux supply chain keeps moving ahead.

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

source https://www.youtube.com/watch?v=OW72wy19cyY

Reconfigurable compute is a double-edged sword: FPGAs let you deploy custom datapaths for low-latency inference, networking, and acceleration, but the bitstream itself can become an attack surface in multi-tenant cloud and shared embedded platforms. This talk explains a practical “bitstream vetting” idea where users could upload a compiled .bit file, then an offline/on-device classifier screens it for hardware-Trojan style payloads before configuration happens. https://sc.edu/

—
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.
—

The core method is intentionally lightweight and static: treat the FPGA bitstream as raw binary, extract byte-frequency features, then compress the feature space with Truncated SVD (TSVD) so inference stays cheap. Training data is built from Trust-Hub benchmark designs (AES-128 and RS232 variants) synthesized into labeled benign/malicious/empty bitstreams, with class imbalance handled via SMOTE and model selection done with k-fold cross validation. The best-performing model is a Random Forest, which is a good fit here because it handles noisy, high-dimensional distributions without needing deep learning on the target node or a big GPU.

A key point is deployment realism: the pipeline is demonstrated on the Digilent PYNQ-Z1 (AMD/Xilinx Zynq-7000), using the PYNQ Python stack and a Jupyter workflow rather than custom HDL changes. On-device results show about 3.35 seconds average latency per classification (feature extraction dominates, prediction is ~15–17 ms), while a hold-out test reports ~97.14% true-positive rate with ~0.8% false-positive rate, which matters when “false alarm” means re-running a long build or blocking a tenant. The interview is filmed at Supercomputing SC25 in St. Louis, which is a fitting venue for the cloud-to-edge security angle that shows up here.

What’s interesting going forward is interpretability and developer ergonomics: the next step described is pairing the detector with an NLP-style explanation layer so a misclassification can be translated into human-readable “why” signals, instead of a bare label. Combined with newer academic boards like AMD University Program AUP-ZU3 (Zynq UltraScale+ XCZU3EG class platforms), this kind of binary-level screening hints at a deployable security control that doesn’t require netlists, source RTL, or reverse engineering, and can sit right at the boundary between CI/CD and the FPGA fabric roadmap.

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

source https://www.youtube.com/watch?v=7JjvXgHE9C4

Digital Logic is a Serbia-based electronics maker focused on NFC/RFID reader-writer hardware aimed at developers and system integrators building time attendance, access control, and vending stack. Their flagship µFR Zero Online is an OEM-sized reader that pairs a 13.56 MHz multi-ISO NFC front end (ISO14443 A/B and ISO15693) with an ESP32-S3 module for Wi-Fi and BLE, plus wired options like Ethernet and Power over Ethernet for clean single-cable setup. https://www.d-logic.com/ufr-zero-online-series-multi-protocol-network-nfc-rfid-readers-writers/

—
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 key idea in this demo is “tap, then authenticate”: NFC is used for near-field intent detection, and then Bluetooth can wake only when a phone is close, enabling a proximity-based exchange that feels like a single interaction to the user. With a custom app, the phone and reader can establish a secure channel, pass signed data, and treat the handset like a credential without exposing raw identifiers in view.

The broader µFR Zero lineup is intentionally modular: round housings, quarter-card (QS), half-card (HS), and larger antenna variants share the same integration philosophy while trading size for RF field strength. In the booth walkthrough recorded at Embedded World North America 2025, they show how antenna geometry maps to practical read ranges, roughly from about 10 cm on the compact unit to around 17 cm, up to about 30 cm on the XL format.

For system builders, the appeal is deployment flexibility: Wi-Fi/BLE for retrofits, Ethernet/PoE for fixed installs, and developer-friendly interfaces like USB CDC/HID, UART/SPI, and event streaming over network protocols. Add-ons such as touch displays, SD storage, GPIO/relays, and secure-element or SAM-style modules let integrators tune security posture, offline logging, and UX without redesigning the core.

A concrete example is self-serve beverage walls, where a prepaid NFC card or phone credential meters dispensing while a back end decrements credit in real time. That “reader as a networked edge node” pattern generalizes to ticketing, lockers, kiosks, and industrial access points, where the same device can handle identity, policy checks, and telemetry in a compact embedded stack.

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

source https://www.youtube.com/watch?v=JYpEplERT7A

Silicon Power Industrial focuses on industrial-grade SSDs and DRAM modules built for embedded and automation gear, where long lifecycle, qualification, and customization matter as much as throughput. The interview frames storage as a hardware+firmware stack: NAND selection (pSLC/3D TLC), ECC, wear leveling, bad-block management, and predictable latency for edge AI or vision load.
https://www.silicon-power-industrial.com/

—
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 recurring theme is harsh-environment design: wide-temperature DRAM and SSD options, vibration and shock tolerance, and tighter validation for boxes that live in cabinets, vehicles, or fanless enclosures. In the conversation they reference operating ranges down to about -40°C, and SP markets wider industrial ranges (often up to around +85°C) depending on grade and form factor.

On the software side, they highlight controlled firmware updates, secure delivery, and a rollback mindset so field upgrades do not brick devices. That maps to common industrial features such as power-failure protection, S.M.A.R.T health data, and at-rest security via self-encrypting drives (TCG Opal 2.0 / AES-256), plus vendor tools for monitoring and configuration across a fleet, reducing operational risk.

Filmed on the show floor at Embedded World North America 2025, the VP also gives the business backdrop: headquarters in Taipei with manufacturing in Taiwan, a U.S. presence in Fremont, and regional coverage across Europe and Asia-Pacific. They position the brand in the industrial memory market as roughly top-10 to top-15, serving IPC builders, automation OEMs, and embedded integrators at scale.

The practical takeaway is that modern industrial memory design is converging on three constraints: deterministic performance (avoiding latency spikes), thermal behavior (heatsinks on SATA SSDs and even DDR5 modules), and auditable device management. For system designers, that pushes storage from a part-number choice into firmware governance, security posture, and telemetry-driven maintenance strategy today.

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

source https://www.youtube.com/watch?v=7C6pUxe2rII

This Microchip Technology demo frames a rain-triggered greenhouse roof as a clean embedded control problem: a wet event becomes a motor move, with a touch UI as a human override. A SAM E54 (ARM Cortex-M4) acts as the central controller, taking inputs from a dedicated liquid-detection device and a separate touch interface MCU, then commanding a stepper stage that opens the roof for natural watering. https://www.microchip.com/

—
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.
—

On the sensing side, the MTCH910 is presented as Microchip’s first liquid-detection ASSP: it encapsulates the detection firmware so the application can treat “rain detected” as a simple signal instead of tuning analog thresholds, filters, and edge cases from scratch. That event is forwarded to the SAM E54, which turns it into an actuator command, keeping timing and safety logic in one place while the roof motion remains predictable.

For manual control, a PIC32CM “MGC” device is used as the touch controller together with a QT7 Xplained Pro board, so a user can set a partial roof opening by tapping and sliding on the panel. Architecturally it’s a familiar IoT pattern: a sensor front end, a UI processor, a main MCU, and a motor driver, each doing one job and communicating through a narrow interface.

The stated purpose is toolchain flexibility rather than the prop itself: the main controller firmware is shown as developed in Zephyr, while the touch firmware uses MPLAB tooling via VS Code extensions and a touch library. In real products, subsystems are often written years apart and by different teams, so being able to mix RTOS builds, vendor SDKs, and minimal-firmware peripherals is a practical advantage for long-lived device software.

Microchip also teases an MPLAB AI Coding Assistant that runs inside VS Code: a datasheet-parsing agent that can answer targeted questions (like how an ADC works) and help you jump to the right register diagrams or tables faster. A newer “agent mode” is described as reading MCC-generated source and APIs, then drafting the application glue code so you can move from GUI-based peripheral configuration to a working main loop with fewer manual edits. The same VS Code extension bundle also exposes MPLAB Code Configurator, a peripheral GUI that writes the low-level register setup for you and generates drivers, plus a data visualizer for inspecting runtime signals; the AI assistant is currently VS Code only and is shown with an early free-token model. Filmed at Embedded World North America 2025, it’s a quick snapshot of how modern MCU development is drifting toward generated HAL code, searchable documentation, and agent-assisted bring-up in one workflow.

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

source https://www.youtube.com/watch?v=uBYZG6ml8qE

NEC explains why its long-running vector-computing lineage is now converging with a more open scalar ecosystem: building a successor to the company’s Vector Engine PCIe card that pairs NEC vector IP with a RISC-V control and system layer co-developed with Openchip. The point is pragmatic engineering rather than hype: vector math is a NEC strength, while RISC-V software stacks, tooling, and platform glue are easier to share across partners and regions. https://www.nec.com/en/global/solutions/hpc/sx/vector_engine.html

—
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.
—

To ground that story, the interview shows a working Vector Engine card and describes the underlying vector architecture that NEC has iterated for roughly four decades. The design is built around wide vector registers, high sustained throughput, and extreme memory bandwidth via on-card high-bandwidth memory, targeting workloads where data movement dominates. In the SX-Aurora TSUBASA line, Vector Engine specs reach up to 16 vector cores, up to 2.45 TB/s of memory bandwidth, and up to about 4.91 TFLOPS double-precision peak, depending on model and config, which sets the tone for the stack.

The Openchip board on the table is a mock-up, but the architectural direction is clear: a RISC-V scalar environment wrapped around a next-generation NEC vector processing unit so the card can look more like a standard Linux accelerator while keeping the vector datapath specialized. That enables cleaner compiler paths via LLVM/GCC, familiar MPI + OpenMP style programming, and room for tighter host integration through PCIe Gen5/Gen6 and CXL-era memory and coherency ideas. This segment was filmed at Supercomputing SC25 in St Louis, with the broader backdrop of Openchip–NEC work that has also been discussed publicly with HPC groups such as Barcelona Supercomputing Center, around RISC-V-based vector compute on a single card.

The same booth also hints at NEC’s dual role as both chip developer and system integrator, pointing to a Japanese “plasma” cluster concept that mixes mainstream CPUs and GPUs—think Intel Xeon-class parts alongside AMD Instinct MI300A-class accelerators—to chase strong performance per watt in the Green500 mindset. Put together, the thread is about heterogeneous compute as a design rule: CPUs for control, GPUs for throughput, and vectors for bandwidth-bound kernels, with RISC-V acting as the connective tissue for a longer-term, more portable software path.

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

source https://www.youtube.com/watch?v=DFnrIwjmu44