Performance Comparison of Php three accelerators hardware environment: Server: DL180G6 & nbsp; (16 & nbsp; Intel (R) Xeon (R) CPU & nbsp; & nbsp; E5620 & nbsp; @ 2.40 GHz) & nbsp; & nbsp; & nbs Php performance comparison of three accelerators
Hardware environment:
Server: DL180G6 (16 Intel (R) Xeon (R) CPU E5620 @ 2.40 GHz) 16 GB memory
Press: DL180G6 (16 Intel (R) Xeon (R) CPU E5620 @ 2.40 GHz) 16 GB memory
Software environment:
Nginx1.2.5php5.4.9 Max 600php-fpm
Network Environment: Gigabit
Test scenario:
Scenario 1: (without any extension ):
Scenario 2: (enable APC 3.1.14 and allocate a cache of 512 MB ):
Scenario 3: (enable XCache 3.0.1 and allocate a cache of 512 MB ):
Scenario 4: (EAccelerator 1.0-dev, cache allocation 32 M): because 512 cannot be set, the cache generated on the test page is 1.15m, which has little impact.
The above four scenarios, the same number of concurrent users 20 (the best number of users), access the same URL (http://playback. I .ifeng.com/hehe.pi? Vt = 5), each scenario runs for 5 minutes, compare the performance of the three Php caches under the same pressure.
Test results:
Extension |
CPU (us/sy) |
Load average |
Response time (s) |
QPS/s |
None |
67%, 26% |
12 |
0.18 |
108 |
Xcache |
56%, 34% |
15 |
0.16 |
120 |
EAccelerator |
64%, 27% |
15 |
0.13 |
150 |
Apc |
60%, 30% |
13 |
0.11 |
170 |