wip-4b3

crates/clawhdf5-bench/Cargo.toml

@@ -75,3 +75,6 @@ libhdf5-compare = ["hdf5"]
 mpi-io = ["clawhdf5-io/mpi-io", "mpi"]
 # Real MiniLM embeddings for longmemeval_bench, so the vector stage is not inert.
 embeddings = ["candle-core", "candle-nn", "candle-transformers", "tokenizers"]
+# CUDA-accelerated embedding. MiniLM on a CPU takes hours over the full
+# longmemeval_s haystack; on a GPU it is minutes.
+embeddings-cuda = ["embeddings", "candle-core/cuda", "candle-nn/cuda", "candle-transformers/cuda"]

crates/clawhdf5-bench/src/bin/longmemeval_bench/embedder.rs

@@ -34,7 +34,19 @@ impl Embedder {
     /// `config.json` is read when present; otherwise the published MiniLM-L6-v2
     /// architecture constants are used, which are pinned rather than guessed.
     pub fn load(dir: &Path) -> Result<Self, Box<dyn std::error::Error>> {
-        let device = Device::Cpu;
+        // CUDA when the feature is on and a device is actually present; the CPU
+        // path is correct but roughly two orders of magnitude slower, which is
+        // the difference between minutes and most of a day on the full haystack.
+        let device = match Device::new_cuda(0) {
+            Ok(d) => {
+                eprintln!("Embedder: CUDA device 0");
+                d
+            }
+            Err(e) => {
+                eprintln!("Embedder: CPU ({e})");
+                Device::Cpu
+            }
+        };
         let weights = dir.join("model.safetensors");
         let tok_path = dir.join("tokenizer.json");