I did my own small benchmark, with the contents of my current quarantine directory (just 48 infected files, but you get the idea) I did the following:
$ clamdscan -V ClamAV 0.80/876/Thu May 12 18:14:29 2005 $ time clamdscan /tmp/test [snip (the 48 hits)] ----------- SCAN SUMMARY ----------- Infected files: 48 Time: 0.202 sec (0 m 0 s) real 0m0.237s user 0m0.046s sys 0m0.015s $ time clamdscan /tmp/test [snip] ----------- SCAN SUMMARY ----------- Infected files: 48 Time: 0.192 sec (0 m 0 s) real 0m0.227s user 0m0.046s sys 0m0.015s $ time clamdscan /tmp/test [snip] ----------- SCAN SUMMARY ----------- Infected files: 48 Time: 0.184 sec (0 m 0 s) real 0m0.219s user 0m0.046s sys 0m0.031s === Average: 0.193 s (0.228 s real time) === $ clamdscan -V ClamAV 0.85/876/Thu May 12 18:14:29 2005 $ time clamdscan /tmp/test [snip] ----------- SCAN SUMMARY ----------- Infected files: 48 Time: 0.216 sec (0 m 0 s) real 0m0.252s user 0m0.046s sys 0m0.061s $ time clamdscan /tmp/test [snip] ----------- SCAN SUMMARY ----------- Infected files: 48 Time: 0.201 sec (0 m 0 s) real 0m0.235s user 0m0.030s sys 0m0.031s $ time clamdscan /tmp/test [snip] ----------- SCAN SUMMARY ----------- Infected files: 48 Time: 0.201 sec (0 m 0 s) real 0m0.235s user 0m0.030s sys 0m0.031s === Average: 0.206 s (0.241 s real time) === Difference btw. 0.80 and 0.85: less than 7 % (0.80 being faster) This doesn't test a mail server under average conditions (infected and non infected, different attachments and formats, etc.), and the test is probably too small. But this is the idea of doing a benchmark and getting real data, it looks like clamd hasn't changed performance with real viri, it doesn't show anything about performance with clean messages. Regards. -- René Berber _______________________________________________ http://lurker.clamav.net/list/clamav-users.html
