You use WSL to benchmark the Swoole version. But WSL may cause a big

And the result is even better if we use 32 threads to request: <code class="notran

Question with your benchmark method about avatar-cache HOT 1 CLOSED

xtlsoft commented on June 14, 2024

Question with your benchmark method

from avatar-cache.

Comments (1)

xtlsoft commented on June 14, 2024

And the result is even better if we use 32 threads to request:
ab -c 32 -n 10000 -r http://localhost:32501/gravatar/605f8c6c64b8fcd514a0b53c6cc3680c

Server Software:        NodeName
Server Hostname:        localhost
Server Port:            32501

Document Path:          /gravatar/605f8c6c64b8fcd514a0b53c6cc3680c
Document Length:        15080 bytes

Concurrency Level:      32
Time taken for tests:   1.708 seconds
Complete requests:      10000
Failed requests:        57
   (Connect: 0, Receive: 19, Length: 19, Exceptions: 19)
Total transferred:      154685538 bytes
HTML transferred:       150513480 bytes
Requests per second:    5853.62 [#/sec] (mean)
Time per request:       5.467 [ms] (mean)
Time per request:       0.171 [ms] (mean, across all concurrent requests)
Transfer rate:          88424.77 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    1   0.4      0       5
Processing:     1    5   3.5      4      32
Waiting:        0    4   3.5      3      30
Total:          2    5   3.6      4      32
ERROR: The median and mean for the initial connection time are more than twice the standard
       deviation apart. These results are NOT reliable.

Percentage of the requests served within a certain time (ms)
  50%      4
  66%      6
  75%      6
  80%      7
  90%      9
  95%     12
  98%     17
  99%     21
 100%     32 (longest request)

from avatar-cache.

Recommend Projects