tests: fix CI riscv timeout in test-llama-context (#31265)

In --models mode the thread count per model is computed and printed,
but never set in the params, so every context runs on the common
default, the number of physical cores. With several models in
parallel this oversubscribes the CPU, and the busy-wait barriers of
the CPU backend make it collapse.

Set the count in cpuparams and cpuparams_batch.

tests/test-llama-context.cpp

@@ -409,6 +409,10 @@ int main(int argc, char ** argv) {
 
     const int32_t n_threads = std::max<int32_t>(1, n_hw_2 / n_pool);
 
+    // every context of the sweep runs on its share of the cores
+    params.cpuparams.n_threads       = n_threads;
+    params.cpuparams_batch.n_threads = n_threads;
+
     // a column is as wide as its case name, but never narrower than the widest verdict
     const size_t verdict_width = std::max({ test_status_width(test_status::PASS),
                                             test_status_width(test_status::FAIL),