bigdl-llm stress test for stable version (#9781)

* 1k-512 2k-512 baseline

* add cpu stress test

* update yaml name

* update

* update

* clean up

* test

* update

* update

* update

* test

* update

.github/workflows/llm_tests_for_stable_version.yml

@@ -1,4 +1,4 @@
-name: LLM Performance Test for Stable Version
+name: LLM Test for Stable Version
 
 # Cancel previous runs in the PR when you push new commits
 concurrency:
@@ -21,7 +21,7 @@ jobs:
   llm-cpp-build:
     uses: ./.github/workflows/llm-binary-build.yml
 
-  llm-performance-test-on-arc:
+  llm-perf-regression-test-on-arc:
     needs: llm-cpp-build
     strategy:
       fail-fast: false
@@ -104,7 +104,7 @@ jobs:
           python csv_to_html.py -f $CSV_SAVE_PATH/fp8 -b $CSV_SAVE_PATH/fp8/transformer_int4_gpu-results-1baseline.csv -t 5.0
 
 
-  llm-performance-test-on-spr:
+  llm-perf-regression-test-on-spr:
     needs: llm-cpp-build
     strategy:
       fail-fast: false
@@ -152,9 +152,61 @@ jobs:
           # hide time info
           sed -i 's/str(end - st)/"xxxxxx"/g' run.py
           python run.py
-          cp ./*.csv /models/nightly_perf_cpu/
+          cp ./*.csv /models/stable_version_perf_regression_test_cpu/
           cd ../../../test/benchmark
           python -m pip install pandas==1.5.3
-          python csv_to_html.py -f /models/nightly_perf_cpu/ -b /models/nightly_perf_cpu/transformer_int4-results-1baseline.csv -t 5.0
+          python csv_to_html.py -f /models/stable_version_perf_regression_test_cpu/ -b /models/stable_version_perf_regression_test_cpu/transformer_int4-results-1baseline.csv -t 5.0
   
 
+  llm-stress-test-on-spr:
+    needs: llm-perf-regression-test-on-spr
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.9"]
+    runs-on: [self-hosted, llm, spr01-perf]
+    env:
+      OMP_NUM_THREADS: 16
+      THREAD_NUM: 16
+      ANALYTICS_ZOO_ROOT: ${{ github.workspace }}
+    steps:
+      - uses: actions/checkout@v3
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v4
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Install dependencies
+        shell: bash
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install --upgrade wheel
+          python -m pip install --upgrade omegaconf
+          python -m pip install --upgrade pandas
+          python -m pip install --upgrade einops
+          python -m pip install --upgrade tiktoken
+          python -m pip install --upgrade transformers_stream_generator
+
+      - name: Download llm binary
+        uses: ./.github/actions/llm/download-llm-binary
+
+      - name: Run LLM install (all) test
+        uses: ./.github/actions/llm/setup-llm-env
+
+      - name: Test on cpu
+        shell: bash
+        run: |
+          mv python/llm/test/benchmark/stable-version-cpu-stress-test.yaml python/llm/dev/benchmark/all-in-one/config.yaml
+          cd python/llm/dev/benchmark/all-in-one
+          export http_proxy=${HTTP_PROXY}
+          export https_proxy=${HTTPS_PROXY}
+          source bigdl-llm-init -t
+          export OMP_NUM_THREADS=48
+          # hide time info
+          sed -i 's/str(end - st)/"xxxxxx"/g' run-stress-test.py
+          python run-stress-test.py
+          cp ./*.csv /models/stable_version_stress_test_cpu/
+          cd ../../../test/benchmark
+          python -m pip install pandas==1.5.3
+          python csv_to_html.py -f /models/stable_version_stress_test_cpu/

python/llm/dev/benchmark/all-in-one/run-stress-test.py

@@ -0,0 +1,256 @@
+#
+# Copyright 2016 The BigDL Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+
+# this code is copied from llama2 example test, and added performance test
+import torch
+import time
+import gc
+import traceback
+import threading
+
+import numpy as np
+from datetime import date
+
+import os
+current_dir = os.path.dirname(os.path.realpath(__file__))
+benchmark_util_path = os.path.join(current_dir, '..')
+import sys
+sys.path.append(benchmark_util_path)
+from benchmark_util import BenchmarkWrapper
+from bigdl.llm.utils.common.log4Error import invalidInputError
+
+LLAMA_IDS = ['meta-llama/Llama-2-7b-chat-hf','meta-llama/Llama-2-13b-chat-hf',
+             'meta-llama/Llama-2-70b-chat-hf','decapoda-research/llama-7b-hf',
+             'decapoda-research/llama-65b-hf','lmsys/vicuna-7b-v1.5',
+             'lmsys/vicuna-13b-v1.3','project-baize/merged-baize-30b']
+
+CHATGLM_IDS = ['THUDM/chatglm-6b', 'THUDM/chatglm2-6b', 'THUDM/chatglm3-6b']
+
+LLAVA_IDS = ['liuhaotian/llava-v1.5-7b']
+
+results = []
+excludes = []
+
+def run_model_in_thread(model, in_out, tokenizer, result, warm_up, num_beams, input_ids, out_len, actual_in_len, num_trials):
+    for i in range(num_trials + warm_up):
+        st = time.perf_counter()
+        output_ids = model.generate(input_ids, do_sample=False, max_new_tokens=out_len,
+                                    num_beams=num_beams)
+        torch.xpu.synchronize()
+        end = time.perf_counter()
+        output_ids = output_ids.cpu()
+        print("model generate cost: " + str(end - st))
+        output = tokenizer.batch_decode(output_ids)
+        print(output[0])
+        actual_out_len = output_ids.shape[1] - actual_in_len
+        if i >= warm_up:
+            result[in_out].append([model.first_cost, model.rest_cost_mean, model.encoder_time,
+                                actual_in_len, actual_out_len])
+
+def run_model(repo_id, test_api, in_out_pairs, local_model_hub=None, warm_up=1, num_trials=3, num_beams=1, low_bit='sym_int4', cpu_embedding=False):
+    # TODO: make a parameter
+    result= {}
+    if test_api == 'transformer_int4':
+        result = run_transformer_int4(repo_id, local_model_hub, in_out_pairs, warm_up, num_trials, num_beams, low_bit)
+    elif test_api == 'transformer_int4_gpu':
+        result = run_transformer_int4_gpu(repo_id, local_model_hub, in_out_pairs, warm_up, num_trials, num_beams, low_bit)
+
+
+    for in_out_pair in in_out_pairs:
+        if result and result[in_out_pair]:
+            results.append([repo_id,
+                            round(np.mean(result[in_out_pair], axis=0)[0]*1000.0, 2),
+                            round(np.mean(result[in_out_pair], axis=0)[1]*1000.0, 2),
+                            round(np.mean(result[in_out_pair], axis=0)[2]*1000.0, 2),
+                            in_out_pair,
+                            f'{int(np.mean(result[in_out_pair], axis=0)[3])}' +
+                            f'-{int(np.mean(result[in_out_pair], axis=0)[4])}',
+                            num_beams,
+                            low_bit,
+                            cpu_embedding if 'win' in test_api else 'N/A',
+                            result[in_out_pair][-1][5] if 'win' in test_api else 'N/A']) # currently only peak mem for win gpu is caught here
+
+
+def get_model_path(repo_id, local_model_hub):
+    if local_model_hub:
+        repo_model_name = repo_id.split("/")[1]
+        local_model_path = local_model_hub + os.path.sep + repo_model_name
+        invalidInputError(os.path.isdir(local_model_path),
+                          local_model_path + " not exists!, Please check your models' folder.")
+        return local_model_path
+    else:
+        return repo_id
+
+
+
+
+
+def run_transformer_int4(repo_id,
+                         local_model_hub,
+                         in_out_pairs,
+                         warm_up,
+                         num_trials,
+                         num_beams,
+                         low_bit):
+    from bigdl.llm.transformers import AutoModel, AutoModelForCausalLM
+    from transformers import AutoTokenizer, LlamaTokenizer
+
+    model_path = get_model_path(repo_id, local_model_hub)
+    # Load model in 4 bit,
+    # which convert the relevant layers in the model into INT4 format
+    st = time.perf_counter()
+    if repo_id in CHATGLM_IDS:
+        model = AutoModel.from_pretrained(model_path, load_in_low_bit=low_bit, trust_remote_code=True, torch_dtype='auto').eval()
+        tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
+    elif repo_id in LLAMA_IDS:
+        model = AutoModelForCausalLM.from_pretrained(model_path, load_in_low_bit=low_bit, trust_remote_code=True,
+                                                     use_cache=True).eval()
+        tokenizer = LlamaTokenizer.from_pretrained(model_path, trust_remote_code=True)
+    else:
+        model = AutoModelForCausalLM.from_pretrained(model_path, load_in_low_bit=low_bit, trust_remote_code=True,
+                                                     use_cache=True).eval()
+        tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
+    end = time.perf_counter()
+    print(">> loading of model costs {}s".format(end - st))
+
+    model = BenchmarkWrapper(model)
+
+    result = {}
+    with torch.inference_mode():
+        for in_out in in_out_pairs:
+            in_out_len = in_out.split("-")
+            in_len = int(in_out_len[0])
+            out_len = int(in_out_len[1])
+            i = 0
+            with open("prompt/stress_test.txt", 'r') as file:
+                for input_str in file:
+                    # As different tokenizer has different encodings,
+                    # slice the input_ids to ensure the prompt length is required length.
+                    input_ids = tokenizer.encode(input_str, return_tensors="pt")
+                    input_ids = input_ids[:, :in_len]
+                    true_str = tokenizer.batch_decode(input_ids)[0]
+                    input_ids = tokenizer.encode(true_str, return_tensors="pt")
+                    actual_in_len = input_ids.shape[1]
+                    result[in_out] = []
+                    st = time.perf_counter()
+                    output_ids = model.generate(input_ids, do_sample=False, max_new_tokens=out_len,
+                                                num_beams=num_beams)
+                    end = time.perf_counter()
+                    print("model generate cost: " + str(end - st))
+                    output = tokenizer.batch_decode(output_ids)
+                    print(output[0])
+                    actual_out_len = output_ids.shape[1] - actual_in_len
+                    if i >= warm_up:
+                        result[in_out].append([model.first_cost, model.rest_cost_mean, model.encoder_time,
+                                            actual_in_len, actual_out_len])
+                    i += 1
+
+    return result
+
+
+def run_transformer_int4_gpu(repo_id,
+                             local_model_hub,
+                             in_out_pairs,
+                             warm_up,
+                             num_trials,
+                             num_beams,
+                             low_bit):
+    from bigdl.llm.transformers import AutoModel, AutoModelForCausalLM
+    from transformers import AutoTokenizer, GPTJForCausalLM, LlamaTokenizer
+    import intel_extension_for_pytorch as ipex
+    model_path = get_model_path(repo_id, local_model_hub)
+    # Load model in 4 bit,
+    # which convert the relevant layers in the model into INT4 format
+    st = time.perf_counter()
+    if repo_id in CHATGLM_IDS:
+        model = AutoModel.from_pretrained(model_path, load_in_low_bit=low_bit, optimize_model=True,
+                                          trust_remote_code=True, use_cache=True).eval()
+        tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
+        model = model.to('xpu')
+    elif repo_id in LLAMA_IDS:
+        model = AutoModelForCausalLM.from_pretrained(model_path, load_in_low_bit=low_bit, trust_remote_code=True,
+                                                     use_cache=True).eval()
+        tokenizer = LlamaTokenizer.from_pretrained(model_path, trust_remote_code=True)
+        model = model.to('xpu')
+    else:
+        model = AutoModelForCausalLM.from_pretrained(model_path, optimize_model=True, load_in_low_bit=low_bit,
+                                                     trust_remote_code=True, use_cache=True).eval()
+        tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
+        model = model.to('xpu')
+        if isinstance(model, GPTJForCausalLM):
+            # For gpt-j model family, this optimization can provide a better performance.
+            model = ipex.optimize(model.eval(), inplace=True)
+    end = time.perf_counter()
+    print(">> loading of model costs {}s".format(end - st))
+
+    model = BenchmarkWrapper(model)
+
+    result = {}
+    with torch.inference_mode():
+        for in_out in in_out_pairs:
+            in_out_len = in_out.split("-")
+            in_len = int(in_out_len[0])
+            out_len = int(in_out_len[1])
+            # As different tokenizer has different encodings,
+            # in_len.txt maybe shorter than we need,
+            # use much longer context to make sure input length
+            test_length = min(in_len*2, 8192)
+            while test_length not in [32, 256, 1024, 2048, 8192]:
+                test_length = test_length * 2
+            input_str = open(f"prompt/{test_length}.txt", 'r').read()
+            # As different tokenizer has different encodings,
+            # slice the input_ids to ensure the prompt length is required length.
+            input_ids = tokenizer.encode(input_str, return_tensors="pt")
+            input_ids = input_ids[:, :in_len]
+            true_str = tokenizer.batch_decode(input_ids)[0]
+            input_ids = tokenizer.encode(true_str, return_tensors="pt").to('xpu')
+            actual_in_len = input_ids.shape[1]
+            result[in_out] = []
+            thread = threading.Thread(target=run_model_in_thread, args=(model, in_out, tokenizer, result, warm_up, num_beams, input_ids, out_len, actual_in_len, num_trials))
+            thread.start()
+            thread.join()
+    del model
+    torch.xpu.empty_cache()
+    return result
+
+
+
+
+if __name__ == '__main__':
+    from omegaconf import OmegaConf
+    conf = OmegaConf.load(f'{current_dir}/config.yaml')
+    today = date.today()
+    if 'exclude' in conf:
+        excludes = conf['exclude']
+    
+    import pandas as pd
+    for api in conf.test_api:
+        for model in conf.repo_id:
+            in_out_pairs = conf['in_out_pairs'].copy()
+            if excludes:
+                for in_out in conf['in_out_pairs']:
+                    model_id_input = model + ':' + in_out.split('-')[0]
+                    if model_id_input in excludes:
+                        in_out_pairs.remove(in_out)
+            run_model(model, api, in_out_pairs, conf['local_model_hub'], conf['warm_up'], conf['num_trials'], conf['num_beams'],
+                      conf['low_bit'], conf['cpu_embedding'])
+        df = pd.DataFrame(results, columns=['model', '1st token avg latency (ms)', '2+ avg latency (ms/token)', 'encoder time (ms)',
+                                            'input/output tokens', 'actual input/output tokens', 'num_beams', 'low_bit', 'cpu_embedding', 
+                                            'peak mem (GB)'])
+
+        df.to_csv(f'{current_dir}/{api}-results-{today}.csv')
+        results = []

python/llm/test/benchmark/csv_to_html.py

@@ -31,7 +31,7 @@ def highlight_vals(val, max=3.0):
         return ''
 
 def is_diffs_within_normal_range(diff1, diff2, threshold=5.0):
-    return not any(diff < (-threshold) for diff in diff1 + diff2)
+    return not any(diff < (-threshold) for diff in diff1 + diff2 if isinstance(diff, float))
 
 def main():
     parser = argparse.ArgumentParser(description="convert .csv file to .html file")

python/llm/test/benchmark/stable-version-cpu-stress-test.yaml

@@ -0,0 +1,20 @@
+repo_id:
+  - 'meta-llama/Llama-2-7b-chat-hf'
+  - 'meta-llama/Llama-2-13b-chat-hf'
+  - 'THUDM/chatglm2-6b'
+  - 'THUDM/chatglm3-6b'
+  - 'baichuan-inc/Baichuan2-7B-Chat'
+  - 'baichuan-inc/Baichuan2-13B-Chat'
+  - 'Qwen/Qwen-14B-Chat'
+local_model_hub: '/models'
+warm_up: 1
+num_trials: 4
+num_beams: 1 # default to greedy search
+low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
+in_out_pairs:
+  - '1024-512'
+  - '2048-512'
+test_api:
+  - "transformer_int4"
+  # - "transformer_int4_gpu"  # on Intel GPU
+cpu_embedding: False # whether put embedding to CPU (only avaiable now for gpu win related test_api)