Pages: [1]
|
|
|
Author
|
Topic: Problems with CVS server? (Read 1908 times)
|
|
|
|
Nic.
Guest
|
All things being equal, yes it is better to download only the differences; however, CVS is not at all equal to HTTP/FTP/etc.
For one, they run on different servers; which means if someone is downloading via a non-CVS method, they're leaving the CVS server alone. By definition, that is better for the CVS server.
Non-CVS methods of retrieval are almost exclusively I/O-bound (server-wise), meaning that the only thing keeping you from your file is the speed (or utter lack thereof) of the network between you. CVS servers, on the other hand, become more and more CPU-bound as the number of concurrent connections rises, since the CVS server needs to determine what has changed and what needs to be sent to each user. Each instance requires a bit of RAM/CPU, and on the scale with which SourceForge operates, it really adds up fast. HTTP servers, by contrast, are told exactly what to do by the client, and as such have far lower overhead.
So basically my point is, it's better for them if you download whole files, not necessarily better for you.
|
|
|
Logged
|
|
|
|
Culture20
Enlightened
Offline
Posts: 917
Thraddash Flower Child
|
I'm usually a little nicer, and put a "sleep 1" in the while loop (of course, that means it takes a while before I connect).
|
|
|
Logged
|
|
|
|
Pages: [1]
|
|
|
|
|