Description
TL;DR: Complete liquid loop matched to the HYGON DCU K100 AI accelerator. Full-cover block routes coolant across compute die, on-package memory, and power stages. 360mm radiator handles sustained inference and training heat. Dual-card thick-row 360mm variant offered for two-accelerator builds. G1/4 ports f
Complete liquid loop matched to the HYGON DCU K100 AI accelerator. Full-cover block routes coolant across compute die, on-package memory, and power stages. 360mm radiator handles sustained inference and training heat. Dual-card thick-row 360mm variant offered for two-accelerator builds. G1/4 ports for integration into custom loop.
Q: Which accelerator?
A: HYGON DCU K100 only — not a generic GPU cooler.
Q: Radiator sizes?
A: 360mm; thick-row 360mm option for dual-card builds.
Q: Sustained AI workloads?
A: Yes, full-cover block + 360mm radiator for AI inference and LLM batch.