Tensor parallelism (-sm tensor) combined with -ncmoe (CPU-offloaded MoE
experts) aborts during warm-up on MoE models with
GGML_ASSERT(ggml_is_contiguous(tensor)) in ggml-backend-meta.cpp.
The failing tensor is the MoE router output (ffn_moe_topk): it is mirrored
(GGML_BACKEND_SPLIT_AXIS_MIRRORED, replicated across backends since routing
must be identical) and happens to be a non-contiguous view.
ggml_backend_meta_buffer_{get,set}_tensor asserted contiguity before
consulting the split state, so a mirrored non-contiguous tensor tripped the
assert even though the GGML_BACKEND_SPLIT_AXIS_MIRRORED case right below
already handles it.
Move the split-state lookup above the assert and allow the mirrored case in
both get_tensor and set_tensor.
cuda: enable topk-moe fusion for 288 experts (llama/25267)
* cuda: enable topk-moe fusion for 288 experts
The topk-moe fusion only accepted power-of-2 expert counts (or the
special-cased 576), so models with 288 experts (e.g. Step-3.7-Flash)
fell back to the unfused per-layer routing chain: softmax/sigmoid,
argsort, get_rows, sum_rows, div, clamp, scale. At batch size 1 that
is ~330 extra tiny graph nodes per token.
288 is a multiple of the warp size, so the existing kernel already
handles it; this adds the missing template instantiation and accepts
288 in the eligibility check.
Measured on gfx1151 with Step-3.7-Flash IQ4_XS (llama-bench,
-b 4096 -ub 4096 -fa 1 -dio 1 -ctk q8_0 -ctv q8_0; machine idle,
before/after paired so pp4096 stays matched as a load control):
Prompt processing is unaffected (the fusion only touches decode
routing). The decode gain is ~+2.4% at shallow context and fades with
depth: by 30k tokens each step is attention-bound over the KV cache,
so removing the fixed routing overhead is no longer visible.
Assisted-By: Claude Fable 5 <redacted>
* Update tests/test-backend-ops.cpp
Co-authored-by: Oliver Simons <redacted>
* Add comment for case 288 in topk-moe.cu
Remove redundant CUDA copies after gated_delta_net. (llama/23940)
* Remove redundant CUDA copies after gated_delta_net.
Currently, GDN writes recurrent state snapshots into its output tail, then the graph immediately copies those snapshots into ssm_states_all. With MTP draft length 3, target decode uses K=4, so that becomes 4 extra ggml_cuda_cpy calls.
The change detects that gated_delta_net -> view -> cpy pattern and makes the CUDA GDN kernel write the state snapshot(s) directly into the recurrent cache, skipping the intermediate tail writes and copy kernels when safe.
Hongqiang Wang [Sat, 27 Jun 2026 22:36:06 +0000 (15:36 -0700)]
opencl: flash attention improvement (llama/25069)
* opencl: rework FA kernel for f16 and f32
* opencl: flash-attention prefill prepass kernels
- flash_attn_kv_pad_f16 pads the tail KV tile to a BLOCK_N multiple
- flash_attn_mask_pad_f16 pads the matching mask tile
- flash_attn_blk_f16 classifies each KV tile per query block as
fully masked / mixed / fully unmasked, so
the main kernel can skip fully-masked tiles
and the mask lookup for fully-unmasked ones
* opencl: FA kernels for q4_0 and q8_0
* opencl: `set_rows` for f32 to q8_0/q4_0
* opencl: dequant kernels for q4_0 and q8_0
* opencl: add FA tile tuning table with override
* opencl: wire host side for FA
* opencl: q4_0 MoE tensors are also SOA'ed
* opencl: cosmetic fix
* opencl: refactor, also clarify some code paths in comments
Gaurav Garg [Sat, 27 Jun 2026 12:16:21 +0000 (17:46 +0530)]
Added a cudaMemcpy2DAsync fast path to ggml_cuda_cpy (llama/25057)
* [CUDA] Added a cudaMemcpy2DAsync fast path to ggml_cuda_cpy
Add a CUDA ggml_cpy fast path for same-type, same-shape strided copies that are just 2D pitched block copies.
When tensors are not fully contiguous but each row is contiguous, it now uses cudaMemcpy2DAsync instead of the slow element-wise scalar copy kernel.
This fixes the GDN recurrent snapshot update with -np 4, where rollback slots are separated by cache stride gaps.
* Add new tests that execute the new optimized strided copy path
* Return unsupported for strided copy in OpenVINO, as new tests are failing
Andreas Kieslinger [Fri, 26 Jun 2026 14:18:30 +0000 (16:18 +0200)]
sched : reintroduce less synchronizations during split compute (llama/20793)
* CUDA: Improve performance via less synchronizations between token (llama/17795)
* Adds CPU-to-CUDA copy capability to
ggml_backend_cuda_cpy_tensor_async()
* Adds function to relax sync requirements between input copies on
supported backends (CUDA for now)
* Exchanges synchronous copy with async copy function.
* Adds macro guards to allow compilation in non-CUDA builds
* Reworked backend detection in ggml-backend.cpp to avoid linking
conflicts
* Relax requirement of checks in async CUDA copies from backend and buffer type to just buffer type, to avoid linking issues
* Minor cleanup
* Makes opt-in to relax use of explicit syncs more general. Backends like
vulkan which require a synchronization between HtoD copies and graph
execution could also adopt this change now.
* Reintroduces stricter check for CPU->CUDA backend async copy via
GGML_DEVICE_TYPE_CPU.
* Corrects initialization of ggml_backend_sync_mode in
ggml_backend_sched_split initialization
* Simplifies synchronizations to adhere to `saaasg` pattern.
* Apply suggestion from @ggerganov (src->buffer to buf_src)
Co-authored-by: Georgi Gerganov <redacted>
* Apply suggestion from @ggerganov (src->buffer to buf_src) v2
Co-authored-by: Georgi Gerganov <redacted>
---------
Co-authored-by: Georgi Gerganov <redacted>
* Apply suggestions from @johannesgaessler code review
Co-authored-by: Johannes Gäßler <redacted>
* Adds single-GPU synchronizations to multi-GPU settings to fix hip backend pipeline parallel bugs.
* Scheduler Hardening: Exclude hip/MUSA from copy_from_host CPU split ->
GPU split optimization
* Scheduler Hardening: Re-adding original additional synchronizations for
non-async backends
* Adds disclaimer to hip/musa exclusion of copy_from_host. Highlights that it is out of
precaution, but that no perf-impact is visible, and that it can be
revisited separately anytime.
---------
Co-authored-by: Georgi Gerganov <redacted> Co-authored-by: Johannes Gäßler <redacted>
Lin Xiaodong [Wed, 1 Jul 2026 11:10:16 +0000 (19:10 +0800)]
whisper : expose internal VAD speech segments (#3916)
When transcribing with params.vad = true, whisper already computes the speech
segments and keeps them in the state. Expose them so callers can reuse those
boundaries (for example to align or clip subtitles to speech) instead of running
a second, separate VAD pass.
Times are on the original audio timeline in centiseconds; the count is 0 when VAD
was not used. test-vad-full.cpp checks the segments are ordered and non-empty.
Daniel Bevenius [Wed, 1 Jul 2026 07:04:21 +0000 (09:04 +0200)]
ci, examples : add package-lock.json to examples addon node (#3919)
* ci, examples : add package-lock.json to examples addon node
This commit commits the package-lock.json file for examples/addon.node.
This is needed to make the CI workflow work properly and after we
started pinning the action versions in the workflow, the CI started
failing because the package-lock.json was not committed. At least that
is what I think is happening.
* ci: use node 20.x and 22.x instead of 16.x and 18.x
* fix CMakeLists.txt target properties for addon.node
Lin Xiaodong [Wed, 1 Jul 2026 06:23:50 +0000 (14:23 +0800)]
whisper : map token timestamps to original time when VAD is enabled (#3910)
whisper_full_get_token_data().t0/t1 are in the VAD-processed timeline (silences
removed), so they don't line up with the original audio. Add
whisper_full_get_token_t0/t1 that map them back.
A token inside a speech segment is interpolated within it; a token that falls in a
removed inter-segment silence snaps to the nearest boundary, so it never lands in
the middle of a cut-out gap. Without VAD the raw times are returned unchanged.
Lin Xiaodong [Wed, 1 Jul 2026 06:21:35 +0000 (14:21 +0800)]
whisper : make voice_length() utf-8 aware for CJK (#3915)
* whisper : make voice_length() utf-8 aware for CJK
voice_length() weights each token by how long its text takes to say, which drives
how a segment's time is shared between its tokens. It looped over raw bytes, so
every CJK character (3 bytes) was counted ~3x and full-width punctuation never
matched, skewing token timestamps for Chinese/Japanese.
Decode one utf-8 code point at a time and give full-width ，。！？ etc. the same
weights as their ASCII counterparts. Pure-ASCII text is unaffected.
* whisper : one statement per line in voice_length()
Daniel Bevenius [Tue, 30 Jun 2026 15:47:22 +0000 (17:47 +0200)]
ci : set GGML_NATIVE=OFF for sanitize builds (#3920)
This commit sets GGML_NATIVE=OFF for all sanitize builds to avoid
the case where the compilation happens using a runner that support
instructions that may not be available on all runners leading to illegal
instruction and failed CI runs.
Tarek Dakhran [Fri, 26 Jun 2026 07:41:56 +0000 (09:41 +0200)]
ggml-cpu: fix SVE leftover path in ggml_vec_dot_f32 (llama/24699)
* ggml-cpu: fix SVE leftover path in ggml_vec_dot_f32
2D convolutions with kernel size 9 produced different results on SVE
enabled ARM devices. After debugging it turned out that ggml_vec_dot_f32
was using data from inactive lanes.
Use svmla_f32_m(pg, sum1, ax1, ay1) so inactive lanes retain sum1.
David Spruill [Thu, 25 Jun 2026 05:35:21 +0000 (01:35 -0400)]
sycl : support --split-mode tensor (llama/24152)
* Sycl tp stage1 (llama/1)
* SYCL: tensor parallelism (--split-mode tensor) for dual-GPU
Adds the comm_init/comm_free/comm_allreduce_tensor trio that the
meta-backend queries via get_proc_address to enable backend-specific
all-reduce, mirroring the pattern used by ggml-cuda.cu.
For N=2 (the common dual-GPU case) implements a degenerate ring
all-reduce with two size-branched paths:
* Small (nelem < 32768): FP32 direct memcpy + per-device ADD kernel
chained via depends_on(memcpy_event). 4 SYCL submissions/call.
* Large (nelem >= 32768): BF16-compressed. Each device compresses
FP32 -> BF16 in a local outbox, cross-device memcpys to the peer's
inbox (HALF the PCIe bytes), then decompresses + adds into the
local FP32 partial. 6 SYCL submissions/call but PCIe bytes halved
-- wins for any tensor where PCIe dominates kernel time.
Threshold and BF16 path pattern mirror the CUDA NCCL allreduce.
Storage: ONE persistent uint8_t buffer per device, 4 * nelem bytes
(matches both path layouts: FP32 nelem floats; BF16 outbox+inbox =
2 * nelem uint16_t each). Single alloc+free per device keeps the
SYCL pool's strict-LIFO invariant trivial.
Initial impl handles N=2 FP32 contiguous tensors. Other cases return
false, causing the meta-backend to use its generic butterfly fallback.
Per-call sync is intentionally omitted. SYCL in-order queue semantics
ensure that the meta-backend's next compute on the same per-device
queue waits for our final ADD, and the next allreduce's first op on
the same persistent buffer waits via the same queue. Only comm_free
does an explicit final wait.
OneCCL is NOT used: OneCCL 2021.17 hardcodes single-device-per-process
in communicator_impl.hpp:47 (condition devices.size() == 1), which is
incompatible with llama.cpp's single-process multi-GPU model.
Measured on dual Intel Arc Pro B70 (NEO 26.05.x, oneAPI 2025.3 +
DPC++ nightly):
Llama-3.3-70B in SYCL TP now comfortably beats production layer mode
on both prefill and decode. Coder-Next-80B-A3B (MoE) also wins on
both — the BF16 path is what unlocks the many-medium-allreduces
prefill pattern.
Build/CMake: no changes. No new dependencies. ~210 lines added across
ggml-sycl.h and ggml-sycl.cpp.
* Fix comments
* documentation update to address PR feedback
* Bring over my device-to-device memcpy chagnes
* move the dev2dev_memcpy calls to the upstream 7-parameter variety
liminfei-amd [Wed, 24 Jun 2026 09:42:03 +0000 (17:42 +0800)]
vulkan: fail the build when a shader fails to compile (llama/24450)
* vulkan-shaders-gen: fail the build when a shader fails to compile
vulkan-shaders-gen did not detect shader-compile subprocess failures, so a
broken libggml-vulkan could be produced while the build reported success and
the breakage only surfaced at run time. execute_command() discarded the child
exit code (POSIX waitpid passed nullptr for status; the Windows branch never
called GetExitCodeProcess) and string_to_spv decided success only from whether
stderr was empty, so a non-zero exit with empty stderr, or a subprocess that
failed to launch, was treated as success.
Return the child exit code from execute_command() (WEXITSTATUS on POSIX,
GetExitCodeProcess on Windows), treat a non-zero exit or non-empty stderr or a
launch exception as a failure, and record it in an atomic flag. main() checks
the flag after process_shaders() and returns EXIT_FAILURE before writing the
output files, so the build stops instead of emitting a broken backend.
Fixes #24393
Signed-off-by: liminfei-amd <redacted>
* vulkan-shaders-gen: simplify compile_failed access and drop unreachable return
Address review feedback on #24450:
- Access the std::atomic<bool> compile_failed directly (= / implicit bool)
instead of .store()/.load(); the flag stays atomic because the worker
threads in process_shaders() set it concurrently.
- Remove the unreachable trailing return -1 in execute_command(): on POSIX the
child _exit()s after execvp and the parent returns (fork()<0 throws); on
Windows the block returns the exit code.
Wyatt Caldwell [Tue, 23 Jun 2026 10:55:46 +0000 (03:55 -0700)]
vulkan: link ggml-cpu when GGML_VULKAN_CHECK_RESULTS / RUN_TESTS are enabled (llama/24444)
The result-checking and test debug paths in ggml-vulkan.cpp call ggml_graph_compute_with_ctx() to compute a CPU reference graph, but that symbol is defined in ggml-cpu, which ggml-vulkan does not link. Enabling -DGGML_VULKAN_CHECK_RESULTS=ON (or -DGGML_VULKAN_RUN_TESTS=ON) therefore fails to link with an unresolved external (e.g. LNK2019 on MSVC, undefined reference on GCC/Clang). This regressed after ggml-cpu was split into its own library. Link ggml-cpu under those two options so the debug builds link again.
QuantiusBenignus [Fri, 26 Jun 2026 06:09:03 +0000 (02:09 -0400)]
examples : fix argument flag for min speech duration in VAD (#3907)
Fixed the -vspd flag for vad_min_speech_duration_ms, to prevent hiding vad_min_silence_duration_ms.
In usage () clarified the output timestamp units.
Fixed a few typos.
shalinib-ibm [Fri, 19 Jun 2026 05:55:38 +0000 (11:25 +0530)]
ggml-cpu: support K tails in power10 Q8/Q4 MMA matmul (llama/24753)
* ggml-cpu: support K tails in Power10 MMA Q8/Q4 matmul
This patch removes the requirement that K be divisible by kc in the tinyBlas_Q0_PPC tiled matmul path. Process the final K panel using its actual depth and pass the reduced panel size through packing and kernel execution. This allows more workloads to use the MMA kernel and reduces fallback to mnpack.
shalinib-ibm [Wed, 17 Jun 2026 18:45:19 +0000 (00:15 +0530)]
ggml-cpu: Conditionally enable power11 backend based on compiler support (llama/24687)
* ggml: Conditionally enable power11 backend based on compiler support
Guard POWER11 backend creation behind a compiler flag check for -mcpu=power11. This avoids build failures on current GCC/Clang toolchains while preserving forward compatibility once POWER11 support becomes available.
Georgi Gerganov [Wed, 17 Jun 2026 16:38:55 +0000 (19:38 +0300)]
metal : add f16 and bf16 support for concat operator (llama/24724)
* metal : add f16 and bf16 support for concat operator
Extend the Metal backend concat operator to support f16 and bf16 tensor
types in addition to the existing f32 and i32 support.
- Template kernel_concat on type T with specializations for float, half,
bfloat, and int
- Add type-specific pipeline getter ggml_metal_library_get_pipeline_concat()
- Update device support check to allow f16 unconditionally and bf16 when
device supports bfloat16
- Update dispatch to select the correct kernel specialization by type
Assisted-by: pi:llama.cpp/Qwen3.6-27B
* metal : extend concat operator to support f16, bf16, i8, i16 and i64
* update stateful_kv_size correctly in mismatch case
* OpenVINO backend: enable arch test for qwen3vl
* OpenVINO backend: enable cohere2 for arch test
* OpenVINO backend: enable t5 for arch test
* OpenVINO backend: enable jamba for arch test
* OpenVINO backend: remove warning for tmp
* OpenVINO backend: enable kimi-linear for arch test
* Remove unused
* Fix gpt-oss accuracy issue
* OpenVINO backend: enable arctic for arch test
* OpenVINO backend: enable grok for arch test
* Gemma4 initial npu support (llama/179)
* Initiall gemma4 npu support
* temp. fix for gemma4 accuracy bug on npu
* Remove hardcoded names for npu-fold handling
* revert static n tokens for cont translation as it is not needed
* removed unused variable
* ggml-openvino: add GGML_OPENVINO_ENABLE_CACHE env var to control decoder cache. Add environment variable GGML_OPENVINO_ENABLE_CACHE (default: YES). When set to NO, the decoder_cache is bypassed and models are rebuilt from the cgraph on every inference call in both dynamic and static compute paths. This is useful for debugging and verifying correctness without caching interference.
* OpenVINO backend: fallback FLASH_ATTN_EXT in gemma3n to CPU backend
* Add raw ov infer profiling metric
* Add OV raw infer time metric to static compute path
Co-authored-by: virajwad <redacted>
* Modify precision of static profiling
* update to OV 2026.2, add OV windows CI
* fix editorconfig-checks
* Initiall gemma4 npu support
* temp. fix for gemma4 accuracy bug on npu
* Remove hardcoded names for npu-fold handling
* revert static n tokens for cont translation as it is not needed
* removed unused variable
* test-llama-archs fix
* Fix gemma4 flash_attn fallback
* support im2col
* fix code style
* disable add_rope_sin_cos optimization
* stateless boradcast and rope optimizations
* Enable manual gqa attn by default for stateless gpu
* manual gqa: fixed static batch
* gemma4 llama-bench ctx update fix
* Update OV win CI
* stateful rope fusion temp. fix
* OpenVINO backend: Conslolidate supported ops
* Exclude unsupported GGML_OP_SUB cases
* Exclude unsupported TOPK_MOE cases
* OpenVINO Backend: MUL_MAT enhancements
* Update OV CI
* support f16 mask input for npu
* Make GGML_OPENVINO_* env vars usage uniform
Standardize all GGML_OPENVINO_* env flags:
positive integers >0 to enable. Unset, empty, =0, or non-numeric values to disable.
This fixes cases where text values or empty strings enabled features.
* OpenVINO backend: Enhance envvar handling
* more cleanup
* move ggml_openvino_env_flag to appropriate place
Francois Dugast [Wed, 17 Jun 2026 05:54:21 +0000 (07:54 +0200)]
sycl: Add optional USM system allocations (llama/22526)
This introduces an optional feature to allocate large GPU buffers (≥ 1GB)
using USM system allocations if supported by the device. It allows using
buffers from the system allocator then letting the system manage memory
migrations between host and device as necessary.
This feature is disabled by default and requires the GGML_SYCL_USM_SYSTEM
environment variable to enable. If USM system allocations are not supported
by the device or the system, we fallback to regular allocations.
This feature can allow VRAM overcommit. For example, the test below fails
on B580 due to lack of memory for allocation, but it passes when enabling
USM system allocations:
Frosty40 [Tue, 16 Jun 2026 05:35:00 +0000 (00:35 -0500)]
sycl: support reordered Q4_K/Q5_K/Q6_K MoE MUL_MAT_ID (llama/24452)
* sycl: support reordered Q4_K and Q5_K MoE MUL_MAT_ID
Extend reordered-weight handling to fused MoE MUL_MAT_ID for Q4_K and Q5_K expert tensors and add Q5_K reordered DMMV coverage. Unsupported 3D reorder cases now fall back instead of aborting.
overview / pkg / ggml / sources / whisper.cpp / log