Skip to content
KernelIndex
Search⌘K

Compare runs

cohort sha256:b2abddfea5826c5f… · source-native profile
ranking-v1
Not comparableAdd at least one more run to compare.
no ranks: not comparable
105.5 µs
Reproduction-ready · run detail
Comparison identity
operation
gqa-paged-decode-h20-kv4-d128-ps1
workload
num_pages = 3874 · batch_size = 2 · len_indptr = 3 · num_kv_indices = 2292 · bf16 · b90447b2
protocol
flashinfer-bench · 7d44d206
environment
NVIDIA B200 · af7cb68a
correctness policy
64fb869e
metric
latency mean (ns)
Context · 10 fields›
architecture
sm_100
CUDA
12.8
driver
unknown
framework
pytorch 2.9.1+cu128
samples
unknown
evidence
reproducible
license
Apache-2.0
source
available
status
passed
observed
2026-03-28

No winner: these runs didn't measure the same thing. The rows above show what differs. Why comparable?

Add run ›
Export ›
Markdown CSV JSON

flashinfer_wrapper_94b73a · sha256:4e6c4fbe7ffa9… · observed 2026-03-28