Sync master with upstream ggml-org/llama.cpp (82bb48500) - #80
Merged
Conversation
…it (vectorized LD/ST perf opt) for Adreno GPUs (ggml-org#25810) * opencl: read MoE dp4a activation tile as 128-bit local loads * opencl: vectorize MoE dp4a activation staging as 128-bit loads
…ation (ggml-org#25787) DeepSeek-V4's ffn_gate_tid2eid tensor is an i32 token-id -> expert-id index table, not weights. It was never added to the name-based exclusion list alongside ffn_gate_inp.weight, so llama-quantize tries to quantize it and fails since i32 cannot convert to a float type. Fixes ggml-org#25754 Signed-off-by: Yash Raj Pandey <yashpn62@gmail.com>
* dflash: rotate injected K/V cache when using K/V quantization * Update src/models/dflash.cpp Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> * clearer format * remove trailing whitespace --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
…r Adreno Q8_0 MUL_MAT for llama-server multi-stream (ggml-org#25910) * opencl: handle broadcast for adreno gemm/gemv_noshuffle * opencl: honor view_offs for adreno noshuffle gemm/gemv * opencl: general GEMM/GEMV support broadcast * opencl: remove unnecessary tests * opencl: remove unnecessary comments --------- Co-authored-by: Li He <lih@qti.qualcomm.com>
…org#25865) The agentic gate counted MCP servers, builtin and custom tools but not frontend tools, so with the JS sandbox as the only tool source the agentic flow was skipped, no tools field reached the server and the chat template rendered without the tool system prompt. The sandbox is fully client-side: frontendTools derives from the Developer settings toggle alone, counting it in the gate restores that single source of truth.
* ui: fix Show tool call in progress toggle ignored The showToolCallInProgress setting was disconnected from the render path during the agentic content rework: getDefaultExpanded() returns a hardcoded false for tool call sections and an unconditional effect auto-expands the currently executing tool call regardless of the setting. Drive default expansion of all tool call section types from the setting and remove the now redundant auto-expand effect. Manual toggling still takes precedence over the default in both directions. * ui: rename Show tool call in progress to Always show tool call content The previous name suggested symmetry with Show thought in progress, which only applies while inference is running, but tool call content stays expanded after completion. Rename the label, the settings key and the syncable server key to alwaysShowToolCallContent. The synced parameter never worked under its previous name so no migration is needed.
Edge paragraph margins are now zeroed at the source in markdown-content.css, but the user bubble and the system message still carried the -my-4 compensation for them. The uncompensated negative margins shrank the wrapper 2rem below its content, collapsing single-line user bubbles into a scrollable sliver and skewing the system message expand threshold.
* llama_dsv4: write only used rows in state * add TODO about conflating token pos with kv rows
…gml-org#25815) * feat: WIP * feat: Replace conversation rename flow with unified AlertDialog component * feat: Add radio group component and consolidate title generation settings * refactor: Remove JS Sandbox global toggle and migrate legacy user state * chore: Formatting * refactor: Cleanup Co-authored-by: Aleksander Grygier <aleksander.grygier@gmail.com> * refactor: Cleanup * refactor: Marquee selection hook * feat: UI improvements * refactor: Bulk db operations * fix: optimize bulk conversation deletion to handle ancestor chains * refactor: remove pairedKey mechanism from settings system * fix: remove redundant onclick handler from dialog cancel button * chore: pin @lucide/svelte to exact version * feat: Run JavaScript tool disabled by default * fix: correct active conversation deletion tracking in bulk delete * feat: improve shift-key multi-selection support in sidebar via keyboard * refactor: Retrieve JS Tool enabling via Developer Settings * nits: sync, dialog wording, cycle guard, and lockfile follow-ups - Restore titleGenerationUseLLM registry entry so it syncs across devices again - Mention fork cascade in the bulk delete confirmation dialog - Clear newParent on cycle guard break so children never point at a deleted conversation - Align @lucide/svelte in package-lock.json with the exact pin in package.json --------- Co-authored-by: Pascal <admin@serveurperso.com>
k_get_rows_float did a scalar one-element-per-thread copy and recomputed the row-invariant work (index load, fast_div_modulo, src/dst row pointers) for every element. Hoist that out of the per-element loop, and add a vectorized path (k_get_rows_float_vec) that copies one int4 (16 B) per thread for the contiguous same-type (no-cast) case. The vectorized path is gated at compile time (is_same<src0_t, dst_t>) and at runtime on 16-byte alignment of the base pointers and all row strides and on ne00 % VEC == 0. Vectorizing divides the block count by VEC, so a small single-row gather can drop below the device CU count and regress; an occupancy gate keeps those on the block-rich scalar path. On Strix Halo (gfx1151) the DeltaNet recurrent-state gather (ne00=524288) drops 18.6us -> 13.0us (rocprofv3 HW timestamps), faster than the Vulkan backend, with no regression on the small conv-state gather; total get_rows -27%. test-backend-ops GET_ROWS passes (47/47). Assisted-by: Claude Opus 4.8
…nd (ggml-org#25795) This adds the missing `GGML_BACKEND_DL_IMPL()` macro invocation, that other backends have. Fixes ggml-org#25586 for me
…les (ggml-org#23570) * Refactor vk_queue to use per-instance mutexes and unique handles * integrates VK_KHR_internally_synchronized_queues, abstracting the queue submission into a polymorphic interface that completely bypasses host-side mutex locking when driver-side synchronization is supported * fix compilation error * fix duplicate pNext chain for VkPhysicalDeviceInternallySynchronizedQueuesFeaturesKHR * add fallback defines for VK_KHR_internally_synchronized_queues * add null checks for queues in vk_device_struct destructor * use unique_ptr for outer queues to enforce exclusive ownership and optimize lifetime * use static constexpr for eInternallySynchronizedKHR * add lock guard to ggml_vk_create_aliased_queue for thread safety * initialize sync_query_features.internallySynchronizedQueues to VK_FALSE * reuse sync_query_features for internallySynchronizedQueues and simplify chaining * refactor internallySynchronizedQueues detection * fix internallySynchronizedQueues query guard * use eInternallySynchronizedKHR constant * fix self-referential alias for eInternallySynchronizedKHR * use macro for eInternallySynchronizedKHR fallback * fix internallySynchronizedQueues query timing in ggml-vulkan.cpp to prevent device creation mismatch * reset sync_query_features.pNext before reusing in device creation chain, also removed the redundant second probe call * refactor internally synchronized queues detection to use chained feature query and avoid redundant API calls * Update ggml/src/ggml-vulkan/ggml-vulkan.cpp Co-authored-by: Jeff Bolz <jbolz@nvidia.com> * Update ggml/src/ggml-vulkan/ggml-vulkan.cpp Co-authored-by: Jeff Bolz <jbolz@nvidia.com> * Update ggml/src/ggml-vulkan/ggml-vulkan.cpp Co-authored-by: Jeff Bolz <jbolz@nvidia.com> * Update ggml/src/ggml-vulkan/ggml-vulkan.cpp Co-authored-by: Jeff Bolz <jbolz@nvidia.com> * Update ggml/src/ggml-vulkan/ggml-vulkan.cpp Co-authored-by: Jeff Bolz <jbolz@nvidia.com> * Update ggml/src/ggml-vulkan/ggml-vulkan.cpp Co-authored-by: Jeff Bolz <jbolz@nvidia.com> * rename sync_enable_features to internally_synchronized_queues_features * queue_flags is still computed before has_internally_synchronized_queues is set * fix trailing whitespace * replace eInternallySynchronizedKHR macro with static constexpr * preserve source queue semantics in single-queue aliased transfer queue * vulkan: fix cmd_pool access via pointer for compute_queue unique_ptr * vulkan: lock queue during debug label emission when not internally synchronized --------- Co-authored-by: Jeff Bolz <jbolz@nvidia.com>
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
…tion-Id (ggml-org#25760) * server: return 400 instead of 500 on validation error with X-Conversation-Id set_req() attaches the spipe as soon as the header is present, before the request body is parsed. When params validation throws, set_next() never runs and next_orig stays empty, so on_complete() called it and crashed with std::bad_function_call, turning the prepared 400 JSON into a generic 500. on_complete() now treats an empty next_orig as "streaming never started" and evicts the session installed by set_req(), so a failed request leaves nothing behind for discovery or replay. This also covers valid requests that carry the header but do not stream, which previously left an empty finalized session in the map until the GC TTL. * ui: do not send the backend_sampling placeholder On a fresh profile the syncable settings hold the empty string placeholder meaning "let the server decide". Every neighbor field goes through the hasValue() guard that filters it, except backend_sampling, which sent the placeholder verbatim and made every default settings completion fail validation. Guard the field with hasValue() like its neighbors. hasValue(false) is true, so an explicit false still reaches the server and the intent of ggml-org#18781 (send both true and false) is preserved. Only the placeholder is filtered.
With -hfd pointing to a repo shipping speculative sidecars, the draft resolved to the main model of that repo, since find_best_model() excludes sidecar files, and the explicit draft plan suppressed the sidecar discovery on the -hf repo. The draft plan already discovers its sidecars, they were just never consumed. Wire them as the draft, following the fallback pattern of the main plan, so this now works as expected: llama-server -hf repo -hfd repo --spec-type draft-dflash
…olation (ggml-org#25781) The Qwen3-VL learned position embedding is interpolated to the runtime patch grid with the default bilinear+antialias (align_corners=False) sampling, while the transformers reference uses align_corners=True (torch.linspace(0, side-1, T)). The mismatch scales grounding coordinates about the image center, growing with image size and per-axis for non-square images (see ggml-org#16880).
Signed-off-by: wendadawen <wendadawen@qq.com>
* cuda: add k-quant support to GET_ROWS Device-side embedding lookups require GET_ROWS to handle the k-quants used by common GGUF recipes (Q4_K_M stores token_embd as q6_K). Without it the backend rejects the op and the scheduler falls back to the host, copying the full embedding matrix back on every token in single-device graphs. Factor the super-block dequantizers out of the dequantize_block kernels in convert.cu into shared device functions in dequantize.cuh and reuse them from a new k_get_rows_kq kernel : one thread block dequantizes one (dst row, super-block) pair with the existing thread layouts, 32 threads for q4_K and 64 for the other k-quants. Covers q2_K to q6_K in get_rows_cuda and supports_op. i-quants are left as a TODO. * cuda: add i-quant support to GET_ROWS Extends the shared super-block dequantizers to the nine i-quants and reuses them from k_get_rows_kq with the 32-thread layout of the matching convert.cu kernels. supports_op gates the k-quant and i-quant path on ne0 being a multiple of QK_K, which iq4_nl does not guarantee on its own (QK4_NL sub-blocks). mxfp4 is left as a TODO. * cuda: add mxfp4 support to GET_ROWS Moves the mxfp4 dequantizer into the shared super-block helpers and reuses it from k_get_rows_kq with the 32-thread layout of the matching convert.cu kernel. mxfp4 joins the ne0 % QK_K gate in supports_op since its 32-value sub-blocks do not guarantee QK_K-aligned rows on their own. This closes GET_ROWS type coverage on CUDA: every quantized GGML type now takes the direct device path. * cuda: gate the GET_ROWS row size only for 32-value sub-block types Address review from @pwilkin: the i-quant commit replaced the return shared by the whole supported type cascade, so f16/f32/bf16/i32 and the legacy quants also inherited the ne0 % QK_K == 0 gate and any row size that is not a multiple of 256 fell back to the scheduler. Split the cascade: unconditional support is restored everywhere, the gate stays only on iq4_nl and mxfp4 whose 32-value sub-blocks do not guarantee the QK_K super-blocks the kernel iterates on.
* webgpu : add CONV_2D_DW (depthwise conv2d) kernel Implement GGML_OP_CONV_2D_DW for the WebGPU backend, ported from the Vulkan backend's conv2d_dw.comp. Assisted-by: Claude Opus-4.8 * Remove unnecessary comments in webgpu support * update supported ops tables, triggered by adding webgpu CONV_2D_DW
* ggml: enable PowerPC backend variants on AIX Allow the PowerPC CPU backend variants to be built on AIX by extending the platform check in the CMake configuration. This reuses the existing PowerPC backend implementations without changing their behavior. Also fix a missing semicolon in the PowerPC Q0 matmul implementation. * Fix missing semicolon in sgemm.cpp
* server: don't walk Windows junctions in file_glob_search std::filesystem reports a junction as a plain directory, so the symlink guard misses it and a junction pointing back at an ancestor is walked until the path length gives out read the reparse tag and treat a symlink and a mount point as links, leaving any other reparse point walkable so cloud placeholders and dedup stubs still get searched look junk directory names up case insensitively on Windows, where NTFS makes Build the same directory as build test that a junk directory stays selectable while its contents stay out of search results * server: report a directory the walk could not read a directory that fails to open or to iterate was skipped in silence, so a caller got a listing that looked complete while a whole subtree was missing: a path over the platform limit, a volume going away, a name the filesystem rejects skip_permission_denied never reaches this path, so an error here is an incomplete answer rather than a deliberate omission, and it now sets the truncated flag * server: simplify the file_glob_search listing plumbing return a small result struct instead of two out params and a caller path that only fed an error string, taking list_entries from six parameters down to three scope the error code to the directory being read, act on the status code the entry lookups already returned, and treat an unreadable link state as a link so the walk never descends on a guess check the deadline when a directory is popped, not only per entry, so a tree of empty directories cannot outlive the budget read the path parameter once, and reject an invalid limit the way an invalid type is already rejected, instead of silently falling back normalize the resolved path, so a "." or ".." a caller typed reaches neither git nor the client, and return the generic path form with '/' separators on every platform, so the base sent to clients no longer needs a local fixup * ui: expire cached picker searches the cache grew for the lifetime of the component: entries went stale after the TTL but were never removed, so every distinct query typed in a session stayed in memory drop expired entries when a new result is stored * server: address review from @ngxson trim comments to one line each, and drop two that restate the code rename junk_lookup_name to get_effective_name, and move it and the link check to private static members next to junk_dir_names merge the Windows and Linux link checks into one is_link, so symlinks are checked everywhere and junctions only add to it on Windows * server: convert tool paths as UTF-8 on Windows a narrow path uses the active code page there, so a file name came back mangled and a path with an accent could not be opened at all convert explicitly at every crossing between a std::string, which always carries UTF-8 here, and fs::path read the home directory through the wide environment, since the narrow one returns the profile path in the active code page too the walker no longer normalizes separators by hand, since paths now come back in generic form * server: fold the platform branch inside console_output_to_utf8 match the shape of the other helpers, one definition with the #if inside, instead of two definitions wrapped in #if and #else inline the single caller helper and trim the comment
* ggml: add ggml_build_forward_order ggml_build_forward_expand marks the tensor and all its ancestors for compute, so using it as a pure ordering hint (keeping q, k and v together) defeats ggml_build_forward_select: the unselected branch is forced to run with inputs that were never uploaded. In the mtmd audio graph this makes GEN_WAV calls execute the GEN_CODE branch with a stale inp_code0, hitting the get_rows bound assert on CPU. Add ggml_build_forward_order, which inserts nodes without the compute flag; the flag is restored when the branch is actually selected. Switch the q/k/v hints in clip_graph::build_attn to it. * nit: reduce comments (AGENTS.md)
…causes of DeviceLost drivers errors (ggml-org#26371) * vulkan: add debug tooling to get more information about a DeviceLost error * fix submission threshold applied too late * use logging macros, throw instead of aborting * clean up circular dependency
…6665) This commit adds the --model-name flag to the causual and embedding model conversion scripts. The motivation for this is that this is the name used for the metadata field general.name and it can be useful to specify this explicitely if the default (the basename of the model path) is not what we want.
* ci: prepare for amd rocm ci Signed-off-by: Aaron Teo <aaron.teo1@ibm.com> * ci: fix editorconfig-checker Signed-off-by: Aaron Teo <aaron.teo1@ibm.com> * ci: fix device not recognised Signed-off-by: Aaron Teo <aaron.teo1@ibm.com> * ci: rename gpu-amd to gpu-hip Signed-off-by: Aaron Teo <aaron.teo1@ibm.com> * ci: gpu-hip to gpu-rocm haha Signed-off-by: Aaron Teo <aaron.teo1@ibm.com> * CUDA: allow integrated-GPU host output buffer in debug assert On integrated GPUs (APUs), the scheduler can legitimately place a graph node's output on the host-visible buffer, which ggml_cuda_compute_forward already handles. The debug assert in ggml_cuda_graph_evaluate_and_capture required every node output to be on the device buffer, so a debug build aborts on such a node (e.g. attn_residual ADD -> ROCm_Host on RDNA3.5). The source-tensor assert directly below already permits this via the integrated + cuda_host exception; apply the same exception to the node's own output buffer. Debug-only; no effect on release/compute. Fixes test-recurrent-state-rollback on gfx1151 (Strix Halo). * ci: enable unified memory for ROCm gfx1151 job Work around a coherence issue on integrated RDNA3.5 (gfx1151) where GPU kernels reading mmap-loaded weights can return incorrect output, which makes test-llama-archs (and real inference) intermittently wrong. GGML_CUDA_ENABLE_UNIFIED_MEMORY=1 uses managed memory, which restores coherence. Remove once the underlying ROCm/HIP issue is fixed. * test-llama-archs: skip jamba on HIP backend jamba produces incorrect output (~0.55 NMSE vs CPU) on the HIP backend on RDNA3.5 (gfx1151); the SSM kernels need separate investigation. Skip it for now, matching the existing per-backend carve-outs (WebGPU), so the ROCm CI can run the test for the remaining architectures. * ci: use HIP_LAUNCH_BLOCKING for ROCm gfx1151 job The gfx1151 ROCm CI job produced incorrect inference output (qwen3 perplexity ~88 vs ~9.4) due to an async-execution correctness issue in the HIP path. Serializing kernel launches with HIP_LAUNCH_BLOCKING=1 restores correctness. This replaces the earlier GGML_CUDA_ENABLE_UNIFIED_MEMORY workaround, which did not fix batched inference. * test-backend-sampler: skip top-k subtests on HIP backend The ROCm backend does not support the TOP_K/ARGSORT op at vocab scale (no CUB; bitonic argsort is capped at ncols <= 1024), so top-k/top-p backend samplers cannot be offloaded. The penalties, set_sampler, mixed, and top_p subtests assert that offload happened, so they fail on HIP. Skip them until TOP_K is supported on the ROCm backend. * Update tests/test-backend-sampler.cpp Co-authored-by: Aaron Teo <taronaeo@gmail.com> * Update tests/test-backend-sampler.cpp Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> --------- Signed-off-by: Aaron Teo <aaron.teo1@ibm.com> Co-authored-by: Aaron Teo <aaron.teo1@ibm.com> Co-authored-by: Jim Wu <ywu@xilinx.com> Co-authored-by: Aaron Teo <taronaeo@gmail.com> Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* mtmd: add chunk save/load function * nits * add tests * rn _MAX --> _COUNT
…tning_indexer (ggml-org#26646) - In MSL, declaring an array of matrix types like `threadgroup half4x4` causes a 'no matching constructor' compilation error because MSL matrix types do not have zero-argument default constructors and threadgroup variables cannot have initializers. - Fix this by declaring a POD `threadgroup half` array instead and casting to `threadgroup half4x4 *` for matrix indexing. Signed-off-by: JamePeng <jame_peng@sina.com>
1. Abort CI if build requirements are missing. 2. Add check to make sure Git LFS has been configured. 3. Add trailing newlines to log messages.
…#26515) * support fp16 to fp16/fp32 * support all missed data types in set_rows * refactor the code to support all data types
…DSV4_HC_PRE (ggml-org#26568) * support DSv4 OPs: LIGHTNING_INDEXER,DSV4_HC_COMB,DSV4_HC_POST,DSV4_HC_PREwq * update ops.md * fix format issue
* tests: add SWIGLU perf cases perf mode had no GLU coverage. Adds SWIGLU at 17408 columns, 512 and 2048 tokens, f16 and f32, with the operands both fused and split. * sycl: consolidate fused-GLU kernels They differed only in which op_* they called, so take the op as an argument and share a common launcher. Their block sizes were all 256, so launch geometry is unchanged; SYCL_GELU_BLOCK_SIZE and SYCL_SILU_BLOCK_SIZE lose their last users so are dropped. * sycl: contiguous fast path for the fused GLU ops o0 == n and o1 == n collapse the de-interleave index math to the identity, so dispatch a flat kernel in that case. It fires for ggml_glu_split with packed operands; a fused [gate|up] tensor keeps the strided path. test-backend-ops perf -o SWIGLU on an Arc Pro B70: split +14% f16 and +4% f32, fused unchanged.
The NVFP4 quantization format stores a scaling factor for every group of 16 weights, packed into a single UE4M3 byte. The SYCL GPU code was converting these scale values using the E4M3 path, but that's *signed*, and these are unsigned values.
* ui: read model modalities from the router model list The router advertises input modalities for every model, loaded or not. Reading them at list build time lets the UI accept image and audio uploads for a model selected through ?model=, which has no /props yet. * enum
Brings the fork up to upstream 82bb485, 1337 source files. The only conflicts were the 17 upstream CI workflows this fork deletes on purpose in favour of the unsloth-prebuilt set, so those stay deleted. No source file conflicted. build-wasm.yml is new upstream and is deleted for the same reason: it triggers on push to master with broad path filters and would burn runner capacity on every push here.
|
Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits. |
The rule this repo runs on is that a sync may take upstream content but must never modify, revert or delete anything we authored. That was being checked by reading a 1300-file diff, which does not scale and does not catch a renumbered enum. verify_upstream_sync.py checks it exactly instead of by eye. It computes the set of paths we touched since the merge base and requires every one to be byte-identical in the merge, requires every one of our commits to still be an ancestor, and allows a deletion only when upstream deleted the file and we never touched it. Two extra layers on top of the byte compare: an AST pass over our Python that pins every top-level name and its value, and an enum pass over ggml.h and llama.h that pins every GGML_TYPE, GGML_FTYPE and LLAMA_FTYPE id and flags collisions. The enum pass is deliberately not limited to files we touch, because a published GGUF stores those ids, so a shift is a data-corruption bug whoever caused it. Unowned headers are compared against upstream rather than against our stale copy, so upstream adding a type is not a false positive. test_verify_upstream_sync.py builds a throwaway repo of the same shape and injects the four failure modes one at a time: our file modified, our file deleted, an id renumbered, and a value inside our Python changed. All four are caught and a clean merge passes. On this PR the gate reports: 132 fork commits all still ancestors, all 68 paths we touched byte-identical, all 86 deletions upstream retiring its own files, no Python definition lost, no id changed, no collisions.
Member
Author
|
Re-checked this against the rule that a sync may take upstream content but must never modify, revert or delete anything we authored. It holds, and I have added a gate so it does not have to be re-checked by eye next time. What the fork actually owns
ResultThe strongest single statement: the merged tree differs from The 86 deletions relative to our master are upstream retiring its own files ( The gate
Run it on this PR with: python3 scripts/unsloth/verify_upstream_sync.py --repo . --merge <merge-sha>It needs a full clone; it refuses to run on a shallow one rather than report a meaningless pass. One thing to confirm
|
A _COUNT is one past the last id, not an id, so a branch that legitimately adds types must move it and flagging that is a false positive: #61 takes GGML_TYPE_COUNT from 43 to 67. It is still checked, just with the right rule. The sentinel may only grow, never shrink, and it must still bound every real id in its family. A shrink, or a value that no longer bounds the ids, is a violation as before, and no real id may move at all.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Brings the fork up to upstream
ggml-org/llama.cppmaster at82bb48500. Master was 718 commits behind with 132 of its own.Conflicts
Seventeen, all of them
.github/workflows/files that this fork deletes on purpose in favour of theunsloth-prebuilt-*set. Every one is a modify/delete: present upstream, already deleted here. They stay deleted.No source file conflicted. The merge touches 1337 source files.
build-wasm.ymlis new upstream and is deleted for the same reason as the other seventeen: it triggers on push to master with broad path filters, so it would burn runner capacity on every push to this repo.What this means for #61
After this merge #61 no longer merges cleanly, in seven files. Six are adjacency only, where upstream and the PR both add lines to the same enum or macro block and the resolution is to keep both:
ggml/include/ggml.h,include/llama.h,gguf-py/gguf/constants.py,src/llama-model-loader.cpp,ggml/src/ggml-cpu/arch-fallback.h,ggml/src/ggml-cuda/mmq.cuThe seventh is real.
ggml/src/ggml-cuda/mmq.cuhwas refactored upstream in "CUDA: refactor MMQ kernel configuration" (ggml-org#24127) plus five follow-ups, which replaced the tile API that #61's MMQ port was written against. That port needs redoing againstggml_cuda_mmq_config. I will push that as a follow-up commit on #61.Worth noting that the id choice in #61 survives the sync untouched: upstream now uses
GGML_TYPE_Q2_0 = 42andLLAMA_FTYPE_MOSTLY_Q2_0 = 41, exactly the range #61 was moved clear of, and the narrow types at 64/65/66 do not collide with anything.