Hello Christoph,<div><br></div><div>Right now I'm using kernel 3.2. </div><div>My test program writes 4G of data to a file that has been preallocated and each time it writes a page of data randomly to the file. It's always overwriting, and no appending. The offset of each write is always aligned to the page size. There is no overlapping between writes as I said before. </div>
<div>I also tried mixing reads and writes and got the similar result.</div><div>Can you write data to a file with multiple threads and the performance can scale up with more threads?</div><div><br></div><div>Thanks,</div>
<div>Da<br><br><div class="gmail_quote">On Mon, Jan 16, 2012 at 12:48 PM, Christoph Hellwig <span dir="ltr"><<a href="mailto:hch@infradead.org">hch@infradead.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im">On Sun, Jan 15, 2012 at 07:01:42PM -0500, Zheng Da wrote:<br>
> Hello,<br>
><br>
> I surprisedly found that writing data to a file (no appending) with direct<br>
> IO and with multiple threads has the same performance as a single thread.<br>
> Actually, it seems there is only one core is working at a time. In my case,<br>
> each time I write a page to a file and the offset is always aligned to the<br>
> page size, so there is no overlapping between writes.<br>
<br>
</div>What kernel version are you using? Also, what is the exact I/O<br>
pattern?<br>
<br>
</blockquote></div><br></div>