Github user pkures commented on the pull request:
https://github.com/apache/ant/pull/8#issuecomment-95718851
I was thinking along the lines of ... the transfer speed is 1MB/s but
theoretically on 1 Gbit network you should see up to 100MB/s transfer speed
ideally, so I tried to increase 100 times. It's about how often Java gets a
chance to process InputStream data and how many of those 1K packets arrive
between OS context switches.
Increasing the buffer size from 1K to 100K increased the transfer speed to
75MB/s (very large file), so the buffer size really does matter a lot. I can
look how much the InputStream read method reads in one call on average - this
should tell us how large the buffer needs to be for gigabit network. I can also
test over our 15Mbit link, but I'am afraid that's slowest connection I have
available except maybe cellular internet, hmmm, maybe I can test that also.
InputStream.read - Reads up to len bytes of data from the input stream into
an array of bytes.
In my experience it means that on slow network the only risk is that the
excessive buffer space will be wasted for the duration of the transfer. As it
will not wait for the buffer to fill.
But this is really getting little bit complicated for my time constraints
:-) Maybe just adding the attribute to override the defaul buffer size is not a
bad option after all.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org
For additional commands, e-mail: dev-h...@ant.apache.org
