For a given output position j on the time axis, only input positions
i such that i*s0 <= j < i*s0 + K contribute -- i.e.
i in [ceil((j - K + 1)/s0), floor(j/s0)] intersected with [0, IL-1].
That's at most ceil(K/s0) values (typically 2 for stride==K/2
transposed convs).
The current kernel iterates the full IL range and filters with an
`if`, amplifying per-thread work by IL/ceil(K/s0) (~160x for IL=320,
K=10, s0=5 -- a representative codec-decoder shape). On Apple M1
the wasted work trips the macOS GPU watchdog
(kIOGPUCommandBufferCallbackErrorImpactingInteractivity) on long
graphs.
Compute i_min, i_max analytically before the inner loop and iterate
only [i_min, i_max]. Output is bit-identical (same multiplies and
adds in the same order); loop bound shrinks by IL/ceil(K/s0).
Tested on M1 with a downstream consumer running a TTS codec at full
T_codec; end-to-end codec decode ~3-4x faster, zero watchdog hits
across long synthesis runs vs ~30% pre-patch.
* spec: support MTP
* fix batch size
* rename files
* cont : simplify (#7)
* MTP: clean-up (#9)
* MTP: clean-up
* review: use llama_context_type instead of llama_graph_type
* review: remove llama_model_has_mtp
* review: fix convert issues
* convert: fix pycheck
* review: formatting
* use `mtp-` for identifying mtp models
* convert: fix mtp conversion
* mtp -> draft-mtp
* remove unused llama_arch
* add need_embd in speculative
* llama: allow partial seq_rm for GDN models for speculative decoding
Currently speculative checkpoint needs to restart from a checkpoint
after some draft tokens are not accepted, this leads to some wastage in
running the target again. This PR adds the ability to rollback upto
`draft_max` by storing the GDN intermediates.
* fix pending state
* vulkan: add GDN partial rollback
* meta: extend check to axis 1
* metal: add GDN partial rollback
Extend the gated delta net kernel to store intermediate states for
partial rollback support on the Metal backend.
- Add K (snapshot slot count) as a function constant
- Read input state from slot 0 of the 3D state tensor
- Write intermediate states to different slots during token loop
- For K=1, maintain backward-compatible single-slot behavior
Ref: https://github.com/ggml-org/llama.cpp/commit/8c05923630110223669f069af2000e9cf10c02bc
Assisted-by: llama.cpp:local pi
* delta_net_base: use ggml_pad instead of new_tensor
* review: add need_rs_seq
* review: rename part_bounded to n_rs
* review: deslop comments
* review: rename, add asserts
* server : adjust checkpoint logic (#11)
* server : adjust checkpoint logic
* cont : rm asserts
* server-context: fix early exit
* spec : fix compatibility with n-gram and add TODOs (#13)
* metal : cleanup
* llama : fix faulty bitwise check in recurrent memory
* server : disable RS-based MTP in combination with other spec types
* spec : add TODOs
* cont : fix comment
* cont : update comment
* common : fix logic for ngram + mtp compat
* llama-memory: enable checkpointing with partial rollback
* cont: add test-case for loading into a dirty ctx
* llama-memory-recurrent: clear rs_idx in clear
* download: fix mtp path
* llama-arch: fix enorm op
* docs: update docs
* conversion: fix type annotations
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
The MUL_MAT test loop iterates over base_types[] to generate non-contig
permutation cases (3 standard permutations across n in {1, 8, 16}).
BF16 is absent from base_types[], so these 9 cases were never generated
for BF16 even though every other type covered by base_types[] tests them.
Add the missing 9 cases explicitly: BF16 x F32, m=16, k=256, bs=[2,3],
permutations {0,2,1,3}, {0,1,3,2}, {0,3,2,1}, with n in {1, 8, 16}.
Suggested-by: @jeffbolznv
* move conversion code to a dedicated conversion directory and split the files akin to the src/models architecture
---------
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
Add Aime2026Dataset class loading from MathArena/aime_2026 on
HuggingFace. 30 problems (two sets of 15), single config/split.
Usage: --dataset aime2026
Assisted-by: llama.cpp:local pi
* Support for Codex CLI by skipping unsupported Responses tools
* Warn on skipped Responses tools and preserve gpt-oss apply_patch rejection
* Revert gpt-oss apply_patch special handling
Adds RDNA3 support to the CUDA mma FA kernel. To make the RDNA3 tensor cores work with the FP16 accumulation for VKQ the tiles they need to be 32 logical units long in direction of the attention head; for head sizes 80 and 112 that are not exactly divided by 32 the regular length of 16 with FP32 accumulation is used instead. The longer tiles also enable more efficient transposition for a warp size of 32 which is why it's also used for RDNA4. However, this scrambles the data layout of the accumulators along the attention head dimension. To prevent accidental misuse I added another entry to ggml_cuda_mma::data_layout.
I also tuned the kernel parameters for RDNA3, RDNA4, and CDNA1 in general, during which I discovered that the kernel can be made to work for head sizes up to 256 for CDNA. For RDNA3/4 I was not able to get better performance that the tile kernel for head sizes > 128.
* update test scripts
* align CI behavior between linux and android
* remove automatically cancel in 15min
* enable cancel-in-progress
* fix ty check issue
* update and fix pylint issue
* update runner such that we are not restricted by the 15min limit rule
* fix flake8 lint issue
* update runner according to review feedback
* code update according to review feedback
* switch from llama-cli to llama-completion binary with -no-cnv flag
Wagner Bruna
Concedo
Aleksander Grygier
Georgi Gerganov
CrispStrobe
Steve Lhomme
Holger Voormann
Alessandro de Oliveira Faria (A.K.A.CABELO)
Aman Gupta
kubawoo
viggy
Pascal
ynankani
askmyteapot
Sigbjørn Skjæret
Omer Ozarslan
Julien Chaumond
Xuan-Son Nguyen
Piotr Wilkin (ilintar)
Reithan
Sid Shaytay
KITAITI Makoto
Pranav Dhinakar
Johannes Gäßler
Zack Li
Reese Levine