-
Notifications
You must be signed in to change notification settings - Fork 12.8k
Closed
Labels
bugSomething isn't workingSomething isn't workinglow severityUsed to report low severity bugs in llama.cpp (e.g. cosmetic issues, non critical UI glitches)Used to report low severity bugs in llama.cpp (e.g. cosmetic issues, non critical UI glitches)
Description
What happened?
Previously, when model_params.n_gpu_layers = 0
metal backend was not initialized. Now even if model_params.n_gpu_layers = 0, metal backend initialization is still performed, which is terminated by the following error:
ggml_metal_init: error: load pipeline error: Error Domain=CompilerError Code=2 "only 14 constant buffers binding are supported in the simulator but 25 were used" UserInfo={NSLocalizedDescription=only 14 constant buffers binding are supported in the simulator but 25 were used}
ggml_backend_metal_device_init: error: failed to allocate context
llama_new_context_with_model: failed to initialize Metal backend
Name and Version
version: b3982
What operating system are you seeing the problem on?
No response
Relevant log output
No response
a-ghorbani and brittlewis12
Metadata
Metadata
Assignees
Labels
bugSomething isn't workingSomething isn't workinglow severityUsed to report low severity bugs in llama.cpp (e.g. cosmetic issues, non critical UI glitches)Used to report low severity bugs in llama.cpp (e.g. cosmetic issues, non critical UI glitches)