Sync master with upstream release b5558#111

Merged

jan-service-account merged 10 commits intodevfrom

update-dev-from-master-2025-06-01-00-11

Jun 1, 2025

Commits on May 31, 2025

CUDA: add a prop in ggml_cuda_device_infor for distinguish iGPU or dGPU in cuda (#13856) (#13895)

authored
kv-cache : refactor + add llama_memory_state_i (#13746)
ggerganov
authored
mtmd : drop `_shared` from `libmtmd` name, merge helpers into libmtmd (⚠️ breaking change) (#13917)
ngxson
authored
llama : auto-batch preparation (#13845)
ggerganov
authored
webui : Replace alert and confirm with custom modals. (#13711)

authored
llama : use n_swa + n_ubatch cells for SWA cache (#13833)
ggerganov
authored
llama : deprecate explicit kv_self defrag/update calls (#13921)
ggerganov
authored
server: allow unclosed thinking tags (#13931)
ochafik
authored
docs : Note about necessity of having libcurl installed for standard build. (#13945)
jpodivin
authored
threading: support for GGML_SCHED_PRIO_LOW, update thread info on Windows to avoid throttling (#12995)

max-krasnyansky
and
slaren
authored