llama.cpp releases · · 1 min read

b9901

Mirrored from llama.cpp releases for archival readability. Support the source by reading on the original site.

sycl : set K_QUANTS_PER_ITERATION to 1 on DMMV path (#25063)

  • sycl: add supported types to ggml_sycl_supports_reorder_dmmv

The reordered feature is implemented in ggml_sycl_op_dequantize_mul_mat_vec,
but gated by ggml_sycl_supports_reorder_dmmv. This commit fixes the gate.

Signed-off-by: Todd Malsbary [email protected]

  • sycl: set K_QUANTS_PER_ITERATION=1 to improve utilization

When combined with opening the reorder gate, this improves GPU
utilization on B70, giving a significant boost to tg t/s.

Signed-off-by: Todd Malsbary [email protected]

  • sycl: replace QK_WARP_SIZE with WARP_SIZE for QK_5

Signed-off-by: Todd Malsbary [email protected]

  • sycl: add missing types to ggml_backend_sycl_buffer_init_tensor

Without this, the extra field is not allocated and the reorder path
will not take effect.

Signed-off-by: Todd Malsbary [email protected]


Signed-off-by: Todd Malsbary [email protected]

macOS/iOS:

Linux:

Android:

Windows:

openEuler:

  • DISABLED
  • openEuler x86 (310p)
  • openEuler x86 (910b, ACL Graph)
  • openEuler aarch64 (310p)
  • openEuler aarch64 (910b, ACL Graph)

UI:

Discussion (0)

Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.

Sign in →

No comments yet. Sign in and be the first to say something.

More from llama.cpp releases