summaryrefslogtreecommitdiffstats
path: root/tools/cuda-ab/victory_strided.cu
diff options
context:
space:
mode:
authorArthur Hanson Rasmusson <arthur@vgpu.io>2026-10-24 02:18:23 -0400
committerArthur Hanson Rasmusson <arthur@vgpu.io>2026-10-24 02:18:23 -0400
commitce6a7560e6e5003eb1c351490bac70125668a4fb (patch)
tree8b3f68d57747ffe97cb8adc9d860aeaa43234bbe /tools/cuda-ab/victory_strided.cu
parent64cca43699d6b74a1f2f2c5e087d33c7aba59de1 (diff)
downloadopen-iov-ce6a7560e6e5003eb1c351490bac70125668a4fb.tar.gz
open-iov-ce6a7560e6e5003eb1c351490bac70125668a4fb.zip
vgsp-sema r595: run a plausible GR push's ops - the driver's relocation copies live on a compute channel
The r568 b-copy class is rooted in the op gate: allow_ops was (ch->engine_type != 0), so a GR-class channel's own memory copies (LAUNCH_DMA on h=0x5c000059, eng=0) never executed unless the push page came from an injected pin. Their releases were still applied by the op-hold path, the client's blocking cudaMemcpy returned, and the launch read the pre-copy page (c = a + 0). Admit ops when the push content is a plausible method stream - the same content arbiter (cap_trust) that already gates GR launch capture - never on a replay or a rehold. Measured: the d_a/d_b relocation copies now run ~0.5-1.5 s after the driver submits them instead of tens of seconds.
Diffstat (limited to 'tools/cuda-ab/victory_strided.cu')
0 files changed, 0 insertions, 0 deletions