Posts under GPU.
A concrete guide to GPU overprovisioning strategies, including scheduler-level oversubscription, time slicing, memory controls, MIG, vGPU, queue backfill, and operational guardrails.
A concrete guide to choosing between serverless GPUs and dedicated GPUs for startups, based on cost structure, delivery speed, performance predictability, operations burden, and team maturity.
An engineering-oriented comparison of KAI-Scheduler’s Reservation Pod approach and HAMi’s hard isolation path, including trade-offs, failure modes (noisy neighbor), and how the two layers can complement each other.
An engineering-oriented guide to hetGPU: how a compiler + runtime stack can make one GPU binary run across NVIDIA/AMD/Intel/Tenstorrent, including SIMT vs MIMD, memory model gaps, and live kernel migration.
A walkthrough into Kubernetes GPU virtualization through gpu-manager startup flow, including device interception, topology awareness, scheduling, and allocation mechanics.
All posts in reverse chronological order.
A concrete guide to GPU overprovisioning strategies, including scheduler-level oversubscription, time slicing, memory controls, MIG, vGPU, queue backfill, and operational guardrails.
A concrete guide to choosing between serverless GPUs and dedicated GPUs for startups, based on cost structure, delivery speed, performance predictability, operations burden, and team maturity.
An engineering-oriented comparison of KAI-Scheduler’s Reservation Pod approach and HAMi’s hard isolation path, including trade-offs, failure modes (noisy neighbor), and how the two layers can complement each other.
An engineering-oriented guide to hetGPU: how a compiler + runtime stack can make one GPU binary run across NVIDIA/AMD/Intel/Tenstorrent, including SIMT vs MIMD, memory model gaps, and live kernel migration.
A walkthrough into Kubernetes GPU virtualization through gpu-manager startup flow, including device interception, topology awareness, scheduling, and allocation mechanics.