ggml-backend: re-enable graph reuse with pipeline parallelism (ggml-org#20927)

Author: am17an

src/llama-context.cpp

@@ -342,14 +342,6 @@ llama_context::llama_context(
 
         if (cparams.pipeline_parallel) {
             LLAMA_LOG_INFO("%s: pipeline parallelism enabled\n", __func__);
-
-            if (!graph_reuse_disable) {
-                // TODO: figure out a way to make graph reuse work with pipeline parallelism
-                // ref: https://github.com/ggml-org/llama.cpp/pull/20463
-                LLAMA_LOG_WARN("%s: graph reuse is currently not compatible with pipeline parallelism - disabling\n", __func__);
-
-                graph_reuse_disable = true;
-            }
         }
 
         sched_reserve();
@@ -1189,6 +1181,13 @@ llm_graph_result * llama_context::process_ubatch(const llama_ubatch & ubatch, ll
     if (!graph_reuse_disable && res->can_reuse(gparams)) {
         //LLAMA_LOG_DEBUG("%s: reusing previous graph\n", __func__);
 
+        // with pipeline parallelism, the previous graph_compute_async may still be running
+        // on the GPU. we must synchronize before set_inputs to avoid overwriting input tensors
+        // that the previous compute is still reading.
+        if (cparams.pipeline_parallel) {
+            ggml_backend_sched_synchronize(sched.get());
+        }
+
         n_reused++;
     } else {
         res->reset();