Silicon Atlas(@Silicon_Atlas):<Silicon Atlas Quick Note> HBF is not cheaper HBM. At Hot Chips, Oxmiq modeled an all-HBF versus all-HBM 72-GPU rack serving Kimi-K2 (1T parameters, FP4; 1M tokens in, 1K out). At equal modeled rack cost and power, HBF provided ~14× the capacity (294.9 vs 20.7 TB), but only ~0.6× the aggregate bandwidth (922 vs 1,584 TB/s). Peak HBF bandwidth requires 64 KB accesses. These are Oxmiq’s simulation results, not measured silicon. My take: HBF is a software-managed capacity tier for low-batch MoE expert weights and sparse-attention KV cache, not a bandwidth substitute for HBM. Source:

Silicon Atlas

@Silicon_Atlas

Evidence-first AI semiconductor analysis: what new silicon claims prove, where bottlenecks move, and whether gains survive at system and economic scale.

参加 March 2021

123 フォロー中 2.9K ファン

Silicon Atlas@Silicon_Atlas

2026.08.31 02:45

HBF is not cheaper HBM. At Hot Chips, Oxmiq modeled an all-HBF versus all-HBM 72-GPU rack serving Kimi-K2 (1T parameters, FP4; 1M tokens in, 1K out). At equal modeled rack cost and power, HBF provided ~14× the capacity (294.9 vs 20.7 TB), but only ~0.6× the aggregate bandwidth (922 vs 1,584 TB/s). Peak HBF bandwidth requires 64 KB accesses. These are Oxmiq’s simulation results, not measured silicon. My take: HBF is a software-managed capacity tier for low-batch MoE expert weights and sparse-attention KV cache, not a bandwidth substitute for HBM. Source: