feat(llm): LLM threads use Workload::Llm + GPU offload detection

Wires the LLM call site through the new tuning helper. gpu_offloaded
reflects intent (use_gpu) cross-checked against the loaded model's
layer count: u32::MAX (when use_gpu) is trivially >= any model's
n_layer, but the explicit comparison is future-proofed if we ever
pass a specific N instead of u32::MAX.

Note: the call site is in generate() not load_model() as the plan
suggested. Context params (and thus thread count) are constructed
per-inference, not per model load, since n_ctx depends on prompt
size. The implementer adapted correctly.

The old magnotia_core::constants::inference_thread_count import is
replaced. Task 6.1 removes the constants helper now that both call
sites have migrated.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

crates/llm/src/lib.rs

@@ -9,6 +9,7 @@ use llama_cpp_2::llama_batch::LlamaBatch;
 use llama_cpp_2::model::params::LlamaModelParams;
 use llama_cpp_2::model::{AddBos, LlamaChatMessage, LlamaChatTemplate, LlamaModel};
 use llama_cpp_2::sampling::LlamaSampler;
+use magnotia_core::tuning::{inference_thread_count, Workload};
 use serde::{Deserialize, Serialize};
 
 pub mod grammars;
@@ -163,7 +164,10 @@ impl LlmEngine {
         }
 
         let n_ctx = preflight_context_window(prompt_tokens.len(), config.max_tokens)?;
-        let thread_count = i32::try_from(magnotia_core::constants::inference_thread_count())
+        let use_gpu = self.loaded_model().map(|s| s.use_gpu).unwrap_or(false);
+        let gpu_layers = if use_gpu { u32::MAX } else { 0u32 };
+        let gpu_offloaded = use_gpu && gpu_layers >= model.n_layer();
+        let thread_count = i32::try_from(inference_thread_count(Workload::Llm, gpu_offloaded))
             .unwrap_or(4);
         let ctx_params = LlamaContextParams::default()
             .with_n_ctx(Some(