Revert "Disable warmup"

This reverts commit 21a30be.

See #485

llama.cpp/common.h

@@ -171,7 +171,7 @@ struct gpt_params {
     bool infill            = false; // use infill mode
     bool dump_kv_cache     = false; // dump the KV cache contents for debugging purposes
     bool no_kv_offload     = false; // disable KV offloading
-    bool warmup            = false; // warmup run
+    bool warmup            = true;  // warmup run
     bool check_tensors     = false; // validate tensor data
 
     std::string cache_type_k = "f16"; // KV cache data type for the K

llamafile/schlep.c

@@ -80,10 +80,6 @@ void llamafile_schlep(const void *data, size_t size) {
     if (size < 128 * 1024 * 1024)
         return;
 
-    // don't bother if we're using a gpu
-    if (llamafile_has_gpu())
-        return;
-
     // don't bother if stderr isn't a terminal
     if (!isatty(2))
         return;