From 90e93db548cc43fb3174516141786c1881c00e30 Mon Sep 17 00:00:00 2001
From: luoyu-intel <yu.luo@intel.com>
Date: Thu, 18 Jul 2024 15:04:41 +0800
Subject: [PATCH] add warm up also for promp_len=32, warm up both gemm and gemv

---
 common/common.cpp | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/common/common.cpp b/common/common.cpp
index dbb724fbb..de68fdf94 100644
--- a/common/common.cpp
+++ b/common/common.cpp
@@ -2126,6 +2126,25 @@ std::tuple<struct llama_model *, struct llama_context *> llama_init_from_gpt_par
         llama_kv_cache_clear(lctx);
         llama_synchronize(lctx);
         llama_reset_timings(lctx);
+        tmp.clear();
+        for (int i = 0; i < 32; i++)
+        {
+            tmp.push_back(bos);
+        }
+        tmp.push_back(eos);
+        if (llama_model_has_encoder(model)) {
+            llama_encode(lctx, llama_batch_get_one(tmp.data(), tmp.size(), 0, 0));
+            llama_token decoder_start_token_id = llama_model_decoder_start_token(model);
+            if (decoder_start_token_id == -1) {
+                decoder_start_token_id = bos;
+            }
+            tmp.clear();
+            tmp.push_back(decoder_start_token_id);
+        }
+        llama_decode(lctx, llama_batch_get_one(tmp.data(), std::min(tmp.size(), (size_t) params.n_batch), 0, 0));
+        llama_kv_cache_clear(lctx);
+        llama_synchronize(lctx);
+        llama_reset_timings(lctx);
     }
 
     return std::make_tuple(model, lctx);