Skip to content

feat(llama-cpp): expose split_mode option for multi-GPU placement #9560

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
mudler merged 1 commit into from
Apr 25, 2026
+16 −0
Merged

feat(llama-cpp): expose split_mode option for multi-GPU placement #9560

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 15 additions & 0 deletions backend/cpp/llama-cpp/grpc-server.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -642,6 +642,21 @@ static void params_parse(server_context& /*ctx_server*/, const backend::ModelOpt
} else if (optval_str == "false" || optval_str == "0" || optval_str == "no" || optval_str == "off" || optval_str == "disabled") {
params.no_op_offload = false;
}
} else if (!strcmp(optname, "split_mode") || !strcmp(optname, "sm")) {
// Accepts: none | layer | row | tensor (the latter requires a llama.cpp build
// that includes ggml-org/llama.cpp#19378, FlashAttention enabled, and KV-cache
// quantization disabled).
if (optval != NULL) {
if (optval_str == "none") {
params.split_mode = LLAMA_SPLIT_MODE_NONE;
} else if (optval_str == "layer") {
params.split_mode = LLAMA_SPLIT_MODE_LAYER;
} else if (optval_str == "row") {
params.split_mode = LLAMA_SPLIT_MODE_ROW;
} else if (optval_str == "tensor") {
params.split_mode = LLAMA_SPLIT_MODE_TENSOR;
}
}
} else if (!strcmp(optname, "kv_unified") || !strcmp(optname, "unified_kv")) {
if (optval_str == "true" || optval_str == "1" || optval_str == "yes" || optval_str == "on" || optval_str == "enabled") {
params.kv_unified = true;
Expand Down
1 change: 1 addition & 0 deletions docs/content/features/text-generation.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -514,6 +514,7 @@ The `llama.cpp` backend supports additional configuration options that can be sp
| `no_op_offload` | boolean | Disable offloading host tensor operations to device. Default: `false`. | `no_op_offload:true` |
| `kv_unified` or `unified_kv` | boolean | Enable unified KV cache. Default: `false`. | `kv_unified:true` |
| `n_ctx_checkpoints` or `ctx_checkpoints` | integer | Maximum number of context checkpoints per slot. Default: `8`. | `ctx_checkpoints:4` |
| `split_mode` or `sm` | string | How to split the model across multiple GPUs: `none` (single GPU only), `layer` (default — split layers and KV across GPUs), `row` (split rows across GPUs), `tensor` (experimental tensor parallelism — requires `flash_attention: true`, no KV-cache quantization, manually set `context_size`, and a llama.cpp build that includes [#19378](https://github.com/ggml-org/llama.cpp/pull/19378)). | `split_mode:tensor` |

**Example configuration with options:**

Expand Down
Loading