Name and Version

build: 5272 (3e959f0)

Compiled with #define GGML_SCHED_MAX_BACKENDS 64

Error:
Device 14: NVIDIA GeForce RTX 3090, compute capability 8.6, VMM: yes
Device 15: NVIDIA GeForce RTX 3090, compute capability 8.6, VMM: yes
error while handling argument "-ts": got 16 input configs, but system only has 16 devices

Operating systems

Linux

Which llama.cpp modules do you know to be affected?

llama-server

Command line

./llama-server ... -ts 10,11,11,11,11,11,11,11,11,11,11,9,9,9,9,9

Problem description & steps to reproduce

Build llama for more than 16 devices, try to use -ts

First Bad Commit

No response

Relevant log output