• ### Popular Now

• 12
• 14
• 13
• 10
• 11

#### Archived

This topic is now archived and is closed to further replies.

# Measuring performance increase?

This topic is 5123 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

## Recommended Posts

Meh, stupid question and it''s probably been asked before, but I did a few forum searches and nothing came up. What is the ''proper'' way to calculate a performance increase? My first thought was (new_time - old_time)/old_time, for a percentage. Is this correct, or are there pitfalls I do not see? What is considered the ''standard'' way to calculate this? Again, sorry if this has been addressed. -Lutrosis

##### Share on other sites
If the new version of a routine is supposed to be faster, then new_time will be smaller than old_time (takes less time to execute), so the increase would be (old_time - new_time)/old_time * 100 percent.

---
tommy online: http://users.pandora.be/tommycarlier