Re: How to (efficiently) write an int array into a file ?
Roedy Green wrote:
On Sun, 27 Sep 2009 12:56:12 -0400, Lew <noone@lewscanon.com> wrote,
quoted or indirectly quoted someone who said :
4. You were playing prosecutor.
I put that possibility at the end, because I thought it was by far the
most plausible. I thought you had some lawyerly argument in mind. I
thought perhaps you were planning to spring an argument for example
that the additional CPU overhead was irrelevant, only elapsed I/O time
mattered. I missed that you were talking about nio. You seemed to be
contracting me, and I was talking about DataOutputStream.
You seemed to be stubbornly insisting on things that you surely knew
must be untrue. Even if by some strange twist, you did not already
know them, I had just explained them. Why were you behaving so
strangely? Were you just trying to get my goat?
So mystery explained. Possibility 3 is the answer.
You were talking about my response to Alexandre Ferrieux, so there was no way
I could've been contradicting you. He had said, "no hope of getting close to
the perf of C's write(), right ?" In my response to that I had said:
For maximum speed in Java, I suspect NIO buffers and their built-in
endian control would rule.
As for your conclusions about speed, on what evidence are they based?
I still have no clue how you derived from that that I was addressing
DataOutputStream. Clearly I was talking about NIO and M. Ferrieux's
conclusion that there's "no hope of getting close to the perf of C's
write()". Clearly I was saying nothing whatsoever about any of your statements.
--
Lew