Modern clusters look uniform from the scheduler, but tiny hardware differences can quietly dominate time-to-solution. In this talk, University of Utah researcher Sowmya Yellapragada describes summer work at Lawrence Berkeley National Lab that treats heterogeneity as a first-class signal, not a nuisance, using empirical node profiles to avoid “fast nodes idle, slow nodes overloaded” behavior. https://www.utah.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 study targets architectural and intra-generation variation, including GPU–GPU differences where you might assume identical behavior. On NERSC Perlmutter, the same NVIDIA A100 family (40 GB vs 80 GB) can shift kernel runtime by roughly 9–26%, especially across memory-bound AMReX kernels, which means naïve load balance can waste expensive accelerators even when every job “fits” on paper.
Two schedulers are compared against a classic homogeneous knapsack baseline. Performance-aware scheduling compresses each node into a 1D relative-speed vector (better/worse than a reference), while relation-aware scheduling lifts that into a 2D relative performance matrix describing how every node compares to every other node, which is useful when topology, memory size, or contention creates non-transitive order.
The evaluation uses 14 representative AMReX kernels spanning compute-bound and memory-bound behavior, and reports near-perfect scheduling efficiency with measurable speedups in moderate heterogeneity and dramatic gains when CPU and GPU resources coexist. Recorded at Supercomputing SC25 in St. Louis, it also frames heterogeneity as a broader orchestration problem that can extend to cloud placement, Kubernetes clusters, and ML-assisted runtime prediction, in the same kind of on-site interview format you also capture at events like Web Summit Lisbon 2025.
A key next step is I/O contention and data movement: PCIe/NVLink traffic, burst buffers, and shared parallel filesystems can turn a “balanced” compute schedule into a stalled pipeline. Modeling those middleware effects alongside compute profiling would let future schedulers optimize makespan with a more realistic cost model, and keep utilization high without guessing.
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
