ren [Fri, 27 Mar 2026 07:05:21 +0000 (00:05 -0700)]
metal : Fix dimension constraint violation in matmul2d descriptor (#21048)
Updates Metal tensor API test probe to fix the dimension constraint violation in the matmul2d descriptor (at least one value must be a multiple of 16).
Adrien Gallouët [Thu, 26 Mar 2026 11:04:37 +0000 (12:04 +0100)]
common : fix split model migration (#21019)
Sadly the manifest does not list all required files, i honestly thought
it was the case
Without the files listed we don't have the sha256, so if the first file
is valid, and all others have the correct size, then we can assume we
are good and do the migration...
$ build/bin/llama-server
================================================================================
WARNING: Migrating cache to HuggingFace cache directory
Old cache: /home/angt/.cache/llama.cpp/
New cache: /home/angt/.cache/huggingface/hub
This one-time migration moves models previously downloaded with -hf
from the legacy llama.cpp cache to the standard HuggingFace cache.
Models downloaded with --model-url are not affected.
================================================================================
migrate_file: migrated angt_test-split-model-stories260K_stories260K-f32-00001-of-00002.gguf -> /home/angt/.cache/huggingface/hub/models--angt--test-split-model-stories260K/snapshots/68c3ea2061e8c7688455fab07597dde0f4d7f0db/stories260K-f32-00001-of-00002.gguf
migrate_file: migrated angt_test-split-model-stories260K_stories260K-f32-00002-of-00002.gguf -> /home/angt/.cache/huggingface/hub/models--angt--test-split-model-stories260K/snapshots/68c3ea2061e8c7688455fab07597dde0f4d7f0db/stories260K-f32-00002-of-00002.gguf
migrate_old_cache_to_hf_cache: migration complete, deleting manifest: /home/angt/.cache/llama.cpp/manifest=angt=test-split-model-stories260K=latest.json
Michael Wand [Thu, 26 Mar 2026 08:54:03 +0000 (01:54 -0700)]
ggml-cuda: Add NVFP4 dp4a kernel (#20644)
Added check for dst_t to cuda_cast template for float
Restored ggml_cuda_ue4m3_to_fp32, changed vecdot ints to int32ts
Added CUDART/HIP Check and HIP/fp8 include
Added NVFP4 to Test-backend-ops
Added hip_fp8_e4m3 to __nv_fp8_e4m3 typedef
SamareshSingh [Thu, 26 Mar 2026 07:14:36 +0000 (02:14 -0500)]
imatrix : fix crash when using --show-statistics with zero counts (#19532)
* imatrix: fix crash when using --show-statistics with zero counts
Fixes division by zero that caused floating point exceptions when processing imatrix files with zero count values. Added checks to skip zero counts and handle empty activation vectors.
Fix for the bug #19190
* imatrix: lower log level for zero-count skip message to DBG
Yihao Wang [Thu, 26 Mar 2026 02:19:14 +0000 (19:19 -0700)]
CUDA & CPU: support F32 kernel type for `CONV_TRANSPOSE_2D` (#17094)
* Refactor CUDA 2D transpose implementation to support multiple kernel types and improve parameter handling
- Introduced a `conv2d_transpose_params` struct for better parameter management.
- Updated `conv2d_transpose_kernel` to be templated for different kernel types (float and half).
- Modified `ggml_cuda_conv_2d_transpose_p0` to handle both F16 and F32 kernel types.
- Enhanced test cases to validate functionality for both kernel types.
* Refactor test cases for 2D convolution transpose to support dynamic kernel types
- Updated `test_conv_transpose_2d` structure to improve parameter handling by reordering constructor arguments.
- Enhanced test case generation to iterate over kernel types, allowing for flexible testing of different configurations.
- Removed hardcoded kernel type instances in favor of a loop for better maintainability and scalability.
* Refactor ggml_compute_forward_conv_transpose_2d to support both F16 and F32 tensor types.
* Refactor conv2d transpose kernel to use a template for kernel type, enhancing flexibility for different data types.
Update test cases to include both F16 and F32 tensor types for comprehensive coverage.
* Update ggml/src/ggml-cuda/conv2d-transpose.cu
Co-authored-by: Aman Gupta <redacted>
* Update ggml/src/ggml-cpu/ggml-cpu.c
Co-authored-by: Aman Gupta <redacted>
* Refactor conv2d transpose implementation by removing the conv2d_transpose_params struct and dispatching with direct kernel launch.
* Enhance cpu conv2d transpose implementation by introducing a templated kernel type for improved flexibility with F16 and F32 data types.
Chenguang Li [Mon, 23 Mar 2026 07:24:06 +0000 (15:24 +0800)]
CANN: add RoPE cache preload before ACL graph capture (#20747)
ACL graph capture disallows host-to-device memcpy and device memory
malloc/free on the captured stream. Pre-load the RoPE cache before
capture so that:
- Host-to-device copies and allocations run on the non-captured stream
- Cache metadata is populated and memory pool is warmed up
- During capture, only on-device computations are recorded; host-side
and allocation branches are skipped
Gaurav Garg [Sun, 22 Mar 2026 08:49:35 +0000 (14:19 +0530)]
[CUDA] Increase number of output elements per-thread block if the K-dimension is small (#20635)
* Increase per-thread work if the K-dimension is small
With tensor parallelism, the K-dimension of the FFN-down matrices is split, which makes it quite small, especially for MOEs. For example, Qwen3-30b-A3B has a K-dimension of 768, and Qwen3235B-A22B has k-dimension of 1536.
The current heuristic uses a group of 4 warps irrespective of K-dimension size, resulting in some of the threads being idle. This results in poor performance for these matrices.
This change increases the number of output elements per block for such cases.
* grammar: prevent stack overflow with nullable symbol loop
Fix a potential stack overflow in llama_grammar_advance_stack that
could occur when processing grammars with nullable symbols that lead
to infinite derivations of empty strings. The fix introduces cycle
detection by tracking visited stacks to prevent infinite recursion.
rg-edit regexp: llama_grammar_advance_stack
rg-edit extra-args: -A20
rg-edit directive: """Rewrite: fix the following segfault:
* grammar: convert recursive llama_grammar_advance_stack to iterative
This change converts the function to an iterative approach using
explicit stacks, which prevents deep recursion and eliminates the risk
of stack overflow.
rg-edit regexp: llama_grammar_advance_stack
rg-edit extra-args: -A30
rg-edit directive: """Rewrite: fix the following segfault:
v2: Added a `std::set` to perform tree-based lookups with O(N log N)
complexity. Testing with a parallel run of `test-grammar-integration`
shows a double-digit percentage increase in runtime. An
`unordered_set` with O(1) hashing was also evaluated, but the overhead
of constructing hash keys from pointers made it significantly slower
than the rbtree implementation that only requires an ordering
operator. The performance regression in the test suite appears
justified by the overall reduction in algorithmic complexity.
Co-developed-by: Piotr Wilkin (ilintar) <redacted>
* grammar: add test case for hang in repetition grammar processing
This commit adds a new test case to the grammar integration tests that
specifically targets a hang scenario in the repetition grammar parser
found while adding GBNF support to ripgrep-edit.
The change introduces a maximum repetition threshold to avoid
excessive rule expansion during grammar parsing. When parsing
repetition patterns like {m,n}, the parser now calculates the
potential number of rules that would be generated and throws an error
if the product of previous rules and new rules exceeds the threshold.
A test case was added to verify the threshold is properly enforced for
deeply nested repetition patterns that would otherwise cause hangs.
Tom Hillbrunner [Sat, 21 Mar 2026 17:35:00 +0000 (18:35 +0100)]
context : use n_embd_out for pooled embedding extraction (#20840)
The MEAN/CLS/LAST pooling paths in encode() and decode() used
n_embd_inp() (16384 for qwen3vl with deepstack) to read from the
pooled embedding tensor, which only has n_embd_out() (4096) floats
per sequence. This caused a tensor read out of bounds assertion.
Fixes embedding mode for Qwen3-VL-Embedding models.
y198 [Sat, 21 Mar 2026 13:59:43 +0000 (20:59 +0700)]
fix(rpc): prevent division by zero in deserialize_tensor (#20712)
rpc : prevent division by zero in deserialize_tensor
When receiving an RPC message with a deprecated tensor type (e.g., type 4 or 5 where `blck_size == 0`), `ggml_row_size()` will trigger a division by zero (SIGFPE) and crash the rpc-server.
This patch adds a simple validation check in `deserialize_tensor` to return `nullptr` if the requested tensor type has a block size of 0.
(Note: This was originally reported via Security Advisory and maintainer suggested dropping a patch here).
shalinib-ibm [Fri, 20 Mar 2026 23:11:45 +0000 (04:41 +0530)]
ggml-cpu: add always_inline to tinyBLAS_PPC accumulator saves (#20791)
Explicitly mark save_acc and add_save_Acc with always_inline
in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator
disassembly within kernel's register context, preventing un-necessary
stask spills.
Jeff Bolz [Fri, 20 Mar 2026 11:17:15 +0000 (06:17 -0500)]
vulkan: change gated_delta_net to shard a column across a subgroup (#20662)
* vulkan: change gated_delta_net to shard a column across a subgroup
This is based on https://github.com/ggml-org/llama.cpp/pull/20391, I used an
LLM to port the CUDA code to Vulkan, and guided to it to make various fixes to
work with Vulkan (e.g. handling different subgroup sizes, unknown mapping of
subgroup to invocation id, using subgroupAdd optionally, etc.).
This fixes a perf regression from the transposing of the values in memory
(!20443).
* vulkan: Spread columns across fewer lanes to reduce the number of workgroups
Ruikai Peng [Fri, 20 Mar 2026 09:31:34 +0000 (17:31 +0800)]
context: zero output buffer on allocation (#20781)
* context: zero output buffer on allocation
Address GHSA-wqq9-25mr-rw76.
The logits output buffer allocated in output_reserve() uses
posix_memalign(), which does not zero memory. The buffer is only
written during decode when needs_raw_logits() returns true. When
backend samplers cover all output sequences, needs_raw_logits()
returns false and the buffer is never written, but
llama_get_logits() still returns a pointer to it, exposing stale
heap content.
Zero the buffer after allocation to prevent information disclosure
through the public logits API.
Found-by: Pwno
* Update src/llama-context.cpp
Co-authored-by: Georgi Gerganov <redacted>
---------
hipudding [Fri, 20 Mar 2026 09:08:39 +0000 (17:08 +0800)]
CANN: add BF16 support for core operators (#20152)
* CANN: add BF16 support for core operators
Add BF16 (bfloat16) type support to the CANN backend for the following
operators: MUL_MAT, MUL_MAT_ID, GET_ROWS, SET_ROWS, CPY, CONT, and
OUT_PROD. This enables BF16 models to run on Ascend NPUs.
* CANN: skip NZ weight format for BF16 and add 310P compile guards
NZ weight format conversion does not support BF16 tensors, skip it
in set_tensor, get_alloc_size and mul_mat. Remove BF16 from MUL_MAT_ID
and OUT_PROD as there are no BF16 use cases. Add #ifndef ASCEND_310P
guards for all BF16 operator support since 310P does not support BF16.
overview / pkg / ggml / sources / llama.cpp / log