Problemsby Ibisllama.cpp lock-in makes switching costly
Cloud and tool cost3 posts from 3 people3d active+00 posts in the last 7 days, 2 the 7 days before (fading)
Posts per day
The posts behind it
3, newest firstserver : saving idle slots to the prompt cache stalls all requests for seconds with interleaved KV cells### Name and Version version: 11387 (8330e96) built with GCC 13.3.0 for x86_64-linux-gnu, CUDA backend ### Operating systems Linux ### Which llama.cpp modules do you know to be affected? llama-server ### Command line ### Problem description & steps to reproduce With unified KV,...ggml-org/llama.cpp · green-real1 repliesOpen ↗Vulkan: add_rms_fusion is disabled for all Intel devices by a vendor gate - re-enable on Xe2 (tested on Arc B580)...| 40.0 | 0% | Suggested change: Question for maintainers: the original exclusion was a workaround for a driver bug observed on Alchemist (A770). If it may still reproduce on older Intel platforms, a narrower gate (for example by architecture - enable on Xe2) would keep the...ggml-org/llama.cpp · dosermenOpen ↗Feature Request: pipeline parallel encoding on single GPU### Prerequisites - [x] I am running the latest code. Mention the version if possible as well. - [x] I carefully followed the [README.md](https://github.com/ggml-org/llama.cpp/blob/master/README.md). - [x] I searched using keywords relevant to my issue to make sure that I am...ggml-org/llama.cpp · uis246Open ↗
Companies and products named
About this problem
Evidence
3 posts from 3 people in 1 place, about 2 a week over 9 days. Mostly on ggml-org/llama.cpp. Tools named alongside: GitHub, NVIDIA.
Frustration Frustration 0 of 3· Seen on GitHub issues
How it was grouped
Posts that state a pain and match the "vendor-lock" rule about llama.cpp. First post Oct 1, 2026, latest Oct 9, 2026. Corroborated: 3 or more posts from 2 or more people or places. Method
The same problem elsewhere
Vendor lock-in makes switching tools costlyGitHub lock-in makes switching costlyKubernetes lock-in makes switching costlyClaude Code lock-in makes switching costlyClaude lock-in makes switching costlyNVIDIA lock-in makes switching costlyCodex lock-in makes switching costlyAnthropic lock-in makes switching costlypandas lock-in makes switching costlyApache Airflow lock-in makes switching costlyvLLM lock-in makes switching costlyPyTorch lock-in makes switching costly
Other llama.cpp problems
Running llama.cpp locally hits memory and speed limitsTrouble with llama.cpp miscllama.cpp users keep filing bugsllama.cpp upgrades break working setupsFine-tuning llama.cpp is costly and unpredictablellama.cpp runs out of context on real workTrouble with llama.cpp buildsTrouble with llama.cpp flashllama.cpp runs out of memoryTrouble with llama.cpp vulkan
History
steady to fading
llama.cpp lock-in makes switching costly (3 posts)