Qeravio
Canonical AI event

Hexagon multi-sequence handling added to concat_2d in llama.cpp

Hexagon multi-sequence handling added to concat_2d. Supports Linux Snapdragon CPU, Adreno GPU, Hexagon NPU.

25 Sept 20261 verified claims1 sources1 observations
What happened

The official source reports this update: b11170: hexagon: handle multi-sequence in concat_2d.

Why it matters

This official update documents a development concerning b11170: hexagon: handle multi-sequence in concat_2d. Its practical significance depends on the scope and evidence stated by the source.

What to watch next

Read the official source update and verify its stated scope, evidence and timing before acting on it.

Connected knowledge

Entities affected by this event

Llama Cppsubject
Continue this topic
infrastructureLegacy Vulkan build issue resolved with cooperativeMatrix API support added.26 Sept 2026infrastructureMetal FWHT kernels now support block widths up to 8192.26 Sept 2026infrastructureMetal kernels now split into per-dtype libraries.26 Sept 2026infrastructureHexagon updates DMA for contiguous dim1 CONCAT.26 Sept 2026infrastructureLlama.cpp updates get_alloc_size cache key and floors result at ggml_nbytes26 Sept 2026infrastructureCUDA kernel fusion reduces kernel launches by 96 per batch on Qwen3.8-27B26 Sept 2026
Evidence trail

Sources behind the event

llama.cpp official releasesb11170: hexagon: handle multi-sequence in concat_2d
Editorial presentation

Read the public story

Open story