Skip to content

improved qkv speed by removing cont op

aaf479a
Select commit
Loading
Failed to load commit list.
Open

UPSTREAM PR #1306: improved flux attention qkv unpacking #71

improved qkv speed by removing cont op
aaf479a
Select commit
Loading
Failed to load commit list.
LOCI Review / Performance Per Binary #71 succeeded Mar 2, 2026 in 0s

Performance improved in some areas, stable in others (within threshold)

1 binary improved · 0 binaries unchanged · 1 binary stable ~ within threshold · 0 binaries degraded ~ beyond threshold

Binary Δ % Response Δ % Throughput Performance (based on response time)
build.bin.sd-cli 0.13 0 stable
build.bin.sd-server -0.01 -0.02 improved

Performance threshold: 30%
Default configuration used.
Note: Performance status is evaluated only from Δ% Response. Throughput is displayed for reference.

Explore the complete analysis inside the Loci Inspector.
Open the Pull Request linked to this check-run.