2024-04-06 05:40:47 +02:00
|
|
|
import sse from 'k6/x/sse'
|
2024-03-09 23:41:49 +01:00
|
|
|
import {check, sleep} from 'k6'
|
|
|
|
import {SharedArray} from 'k6/data'
|
|
|
|
import {Counter, Rate, Trend} from 'k6/metrics'
|
|
|
|
import exec from 'k6/execution';
|
|
|
|
|
|
|
|
// Server chat completions prefix
|
|
|
|
const server_url = __ENV.SERVER_BENCH_URL ? __ENV.SERVER_BENCH_URL : 'http://localhost:8080/v1'
|
|
|
|
|
|
|
|
// Number of total prompts in the dataset - default 10m / 10 seconds/request * number of users
|
|
|
|
const n_prompt = __ENV.SERVER_BENCH_N_PROMPTS ? parseInt(__ENV.SERVER_BENCH_N_PROMPTS) : 600 / 10 * 8
|
|
|
|
|
|
|
|
// Model name to request
|
|
|
|
const model = __ENV.SERVER_BENCH_MODEL_ALIAS ? __ENV.SERVER_BENCH_MODEL_ALIAS : 'my-model'
|
|
|
|
|
|
|
|
// Dataset path
|
|
|
|
const dataset_path = __ENV.SERVER_BENCH_DATASET ? __ENV.SERVER_BENCH_DATASET : './ShareGPT_V3_unfiltered_cleaned_split.json'
|
|
|
|
|
|
|
|
// Max tokens to predict
|
|
|
|
const max_tokens = __ENV.SERVER_BENCH_MAX_TOKENS ? parseInt(__ENV.SERVER_BENCH_MAX_TOKENS) : 512
|
|
|
|
|
|
|
|
// Max prompt tokens
|
|
|
|
const n_prompt_tokens = __ENV.SERVER_BENCH_MAX_PROMPT_TOKENS ? parseInt(__ENV.SERVER_BENCH_MAX_PROMPT_TOKENS) : 1024
|
|
|
|
|
|
|
|
// Max slot context
|
|
|
|
const n_ctx_slot = __ENV.SERVER_BENCH_MAX_CONTEXT ? parseInt(__ENV.SERVER_BENCH_MAX_CONTEXT) : 2048
|
|
|
|
|
|
|
|
export function setup() {
|
|
|
|
console.info(`Benchmark config: server_url=${server_url} n_prompt=${n_prompt} model=${model} dataset_path=${dataset_path} max_tokens=${max_tokens}`)
|
|
|
|
}
|
|
|
|
|
|
|
|
const data = new SharedArray('conversations', function () {
|
|
|
|
const tokenizer = (message) => message.split(/[\s,'".?]/)
|
|
|
|
|
|
|
|
return JSON.parse(open(dataset_path))
|
|
|
|
// Filter out the conversations with less than 2 turns.
|
|
|
|
.filter(data => data["conversations"].length >= 2)
|
|
|
|
.filter(data => data["conversations"][0]["from"] === "human")
|
|
|
|
.map(data => {
|
|
|
|
return {
|
|
|
|
prompt: data["conversations"][0]["value"],
|
|
|
|
n_prompt_tokens: tokenizer(data["conversations"][0]["value"]).length,
|
|
|
|
n_completion_tokens: tokenizer(data["conversations"][1]["value"]).length,
|
|
|
|
}
|
|
|
|
})
|
|
|
|
// Filter out too short sequences
|
|
|
|
.filter(conv => conv.n_prompt_tokens >= 4 && conv.n_completion_tokens >= 4)
|
|
|
|
// Filter out too long sequences.
|
|
|
|
.filter(conv => conv.n_prompt_tokens <= n_prompt_tokens && conv.n_prompt_tokens + conv.n_completion_tokens <= n_ctx_slot)
|
|
|
|
// Keep only first n prompts
|
|
|
|
.slice(0, n_prompt)
|
|
|
|
})
|
|
|
|
|
|
|
|
const llamacpp_prompt_tokens = new Trend('llamacpp_prompt_tokens')
|
|
|
|
const llamacpp_completion_tokens = new Trend('llamacpp_completion_tokens')
|
2024-04-06 05:40:47 +02:00
|
|
|
|
2024-03-09 23:41:49 +01:00
|
|
|
const llamacpp_tokens_second = new Trend('llamacpp_tokens_second')
|
2024-04-06 05:40:47 +02:00
|
|
|
const llamacpp_prompt_processing_second = new Trend('llamacpp_prompt_processing_second')
|
2024-03-09 23:41:49 +01:00
|
|
|
|
|
|
|
const llamacpp_prompt_tokens_total_counter = new Counter('llamacpp_prompt_tokens_total_counter')
|
|
|
|
const llamacpp_completion_tokens_total_counter = new Counter('llamacpp_completion_tokens_total_counter')
|
|
|
|
|
|
|
|
const llamacpp_completions_truncated_rate = new Rate('llamacpp_completions_truncated_rate')
|
|
|
|
const llamacpp_completions_stop_rate = new Rate('llamacpp_completions_stop_rate')
|
|
|
|
|
|
|
|
export const options = {
|
|
|
|
thresholds: {
|
|
|
|
llamacpp_completions_truncated_rate: [
|
|
|
|
// more than 80% of truncated input will abort the test
|
|
|
|
{threshold: 'rate < 0.8', abortOnFail: true, delayAbortEval: '1m'},
|
|
|
|
],
|
|
|
|
},
|
|
|
|
duration: '10m',
|
|
|
|
vus: 8,
|
|
|
|
}
|
|
|
|
|
|
|
|
export default function () {
|
|
|
|
const conversation = data[exec.scenario.iterationInInstance % data.length]
|
|
|
|
const payload = {
|
|
|
|
"messages": [
|
|
|
|
{
|
|
|
|
"role": "system",
|
|
|
|
"content": "You are ChatGPT, an AI assistant.",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
"role": "user",
|
|
|
|
"content": conversation.prompt,
|
|
|
|
}
|
|
|
|
],
|
|
|
|
"model": model,
|
2024-04-06 05:40:47 +02:00
|
|
|
"stream": true,
|
2024-04-04 11:57:58 +02:00
|
|
|
"seed": 42,
|
2024-04-26 09:26:16 +02:00
|
|
|
"max_tokens": max_tokens,
|
|
|
|
"stop": ["<|im_end|>"] // This is temporary for phi-2 base (i.e. not instructed) since the server expects that the model always to emit BOS
|
2024-03-09 23:41:49 +01:00
|
|
|
}
|
|
|
|
|
2024-04-06 05:40:47 +02:00
|
|
|
const params = {method: 'POST', body: JSON.stringify(payload)};
|
|
|
|
|
|
|
|
const startTime = new Date()
|
|
|
|
let promptEvalEndTime = null
|
|
|
|
let prompt_tokens = 0
|
|
|
|
let completions_tokens = 0
|
|
|
|
let finish_reason = null
|
|
|
|
const res = sse.open(`${server_url}/chat/completions`, params, function (client) {
|
|
|
|
client.on('event', function (event) {
|
|
|
|
if (promptEvalEndTime == null) {
|
|
|
|
promptEvalEndTime = new Date()
|
|
|
|
}
|
2024-03-09 23:41:49 +01:00
|
|
|
|
2024-04-06 05:40:47 +02:00
|
|
|
let chunk = JSON.parse(event.data)
|
|
|
|
let choice = chunk.choices[0]
|
|
|
|
if (choice.finish_reason) {
|
|
|
|
finish_reason = choice.finish_reason
|
|
|
|
}
|
2024-03-09 23:41:49 +01:00
|
|
|
|
2024-04-06 05:40:47 +02:00
|
|
|
if (chunk.usage) {
|
|
|
|
prompt_tokens = chunk.usage.prompt_tokens
|
|
|
|
llamacpp_prompt_tokens.add(prompt_tokens)
|
|
|
|
llamacpp_prompt_tokens_total_counter.add(prompt_tokens)
|
|
|
|
|
|
|
|
completions_tokens = chunk.usage.completion_tokens
|
|
|
|
llamacpp_completion_tokens.add(completions_tokens)
|
|
|
|
llamacpp_completion_tokens_total_counter.add(completions_tokens)
|
|
|
|
}
|
|
|
|
})
|
2024-03-09 23:41:49 +01:00
|
|
|
|
2024-04-06 05:40:47 +02:00
|
|
|
client.on('error', function (e) {
|
|
|
|
console.log('An unexpected error occurred: ', e.error());
|
|
|
|
throw e;
|
|
|
|
})
|
|
|
|
})
|
2024-03-09 23:41:49 +01:00
|
|
|
|
2024-04-06 05:40:47 +02:00
|
|
|
check(res, {'success completion': (r) => r.status === 200})
|
2024-03-09 23:41:49 +01:00
|
|
|
|
2024-04-06 05:40:47 +02:00
|
|
|
const endTime = new Date()
|
2024-03-09 23:41:49 +01:00
|
|
|
|
2024-04-06 05:40:47 +02:00
|
|
|
const promptEvalTime = promptEvalEndTime - startTime
|
|
|
|
if (promptEvalTime > 0) {
|
|
|
|
llamacpp_prompt_processing_second.add(prompt_tokens / (promptEvalEndTime - startTime) * 1.e3)
|
|
|
|
}
|
2024-03-09 23:41:49 +01:00
|
|
|
|
2024-04-06 05:40:47 +02:00
|
|
|
const completion_time = endTime - promptEvalEndTime
|
|
|
|
if (completions_tokens > 0 && completion_time > 0) {
|
|
|
|
llamacpp_tokens_second.add(completions_tokens / completion_time * 1.e3)
|
2024-03-09 23:41:49 +01:00
|
|
|
}
|
2024-04-06 05:40:47 +02:00
|
|
|
llamacpp_completions_truncated_rate.add(finish_reason === 'length')
|
|
|
|
llamacpp_completions_stop_rate.add(finish_reason === 'stop')
|
2024-03-09 23:41:49 +01:00
|
|
|
|
|
|
|
sleep(0.3)
|
|
|
|
}
|