<div dir="ltr">The published Phoronix tests are often somewhere between poorly done and wrong; the test measure things that have nothing to do with an operating system (linear disk write speed) or math tests that appear poor because of the particular libs on the system, not because there's anything intrinsic to the operating system.</div>
<div class="gmail_extra"><br><br><div class="gmail_quote">On Sun, May 4, 2014 at 3:22 AM, Wojciech Puchar <span dir="ltr"><<a href="mailto:wojtek@wojtek.tensor.gdynia.pl" target="_blank">wojtek@wojtek.tensor.gdynia.pl</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">examples are<br>
<br>
<a href="http://www.phoronix.com/scan.php?page=news_item&px=MTA3MDA" target="_blank">http://www.phoronix.com/scan.<u></u>php?page=news_item&px=MTA3MDA</a><br>
<br>
that shows very slow test of uncompressing large tarball. I observed the reverse - uncompressing large tarball to hammer filesystem is VERY fast relative to UFS.<br>
<br>
other is here<br>
<br>
<a href="http://www.phoronix.com/scan.php?page=article&item=dragonfly_hammer&num=2" target="_blank">http://www.phoronix.com/scan.<u></u>php?page=article&item=<u></u>dragonfly_hammer&num=2</a><br>
<br>
i've seen some other test before that make me completely ignore DragonflyBSD hammer filesystem.<br>
<br>
<br>
the final thought would be - do your own tests and don't believe anyone.<br>
<br>
<br>
<br>
by the way i've read somewhere that Matt planned to include database like functionality in hammer filesystem directly - simply to allow btree structure to be used for any databases. (TO BE CLEAR - no SQL)<br>
<br>
Would be great to have it and use it as database handler for harbour.<div class="HOEnZb"><div class="h5"><br>
<br>
On Sat, 3 May 2014, Justin Sherrill wrote:<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
<br>
<br>
On Sat, May 3, 2014 at 6:22 AM, Wojciech Puchar <<a href="mailto:wojtek@wojtek.tensor.gdynia.pl" target="_blank">wojtek@wojtek.tensor.gdynia.<u></u>pl</a>> wrote:<br>
<br>
Congratulations. Given all these articles about how bad DragonFly is in performance i see a completely different picture. I did few my own tests about non-I/O performance and<br>
it seems to be at least not worse than FreeBSD.<br>
<br>
<br>
Wait, which articles? There's one set of wildly inaccurate articles I can think of, but nothing else.<br>
<br>
<br>
</blockquote>
</div></div></blockquote></div><br></div>