mirror of
https://github.com/LostRuins/koboldcpp.git
synced 2026-09-04 03:51:17 +02:00
07ff000551
ACL graph capture disallows host-to-device memcpy and device memory malloc/free on the captured stream. Pre-load the RoPE cache before capture so that: - Host-to-device copies and allocations run on the non-captured stream - Cache metadata is populated and memory pool is warmed up - During capture, only on-device computations are recorded; host-side and allocation branches are skipped