<html><head><style>body{font-family:Helvetica,Arial;font-size:13px}</style></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;"><div id="bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px; color: rgba(0,0,0,1.0); margin: 0px; line-height: auto;">For continuous data dimensional compression is probably only going to do common bits removal, and might not even do that, you might just get zlib on your column, you’d have to check the debug and see what it chose.</div><div id="bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px; color: rgba(0,0,0,1.0); margin: 0px; line-height: auto;">If your patches are well correlated (data from similar locations in same patch) then common bits will have more room to take out redundant info.</div><div id="bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px; color: rgba(0,0,0,1.0); margin: 0px; line-height: auto;">Storage should be ‘external’ that is correct, and I’m surprised it got switched back to main? Once you have your own compression scheme (as we do) you don’t want main storage anymore.</div><div id="bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px; color: rgba(0,0,0,1.0); margin: 0px; line-height: auto;">If storage is ‘main’ then “uncompressed” patches of 1000 points will probably end up compressed by the postgresql backend, to fit them onto pages. PostgreSQL really wants to compress things for you, it’s hard to stop it, unless you’re set to external storage.</div><div id="bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px; color: rgba(0,0,0,1.0); margin: 0px; line-height: auto;"><br></div><div id="bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px; color: rgba(0,0,0,1.0); margin: 0px; line-height: auto;">P </div> <br><p style="color:#000;">On November 23, 2015 at 6:58:27 AM, Rémi Cura (<a href="mailto:remi.cura@gmail.com">remi.cura@gmail.com</a>) wrote:</p> <blockquote type="cite" class="clean_bq"><span><div><div></div><div>
<title></title>
<div dir="ltr">
<div class="gmail_default" style="font-family:monospace,monospace">
Hey,<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
I'm a bit suprised by the poor perf of the dimensionnal compression
on my simple XYZ data.<br>
<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
XYZ are declared as int32_t, data is centered on
(0,0,0),<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
ranging from -0.20 to 0.20 , precision 0.00001, 2 millions
points.<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
Patches contains avg 1000 points, about 1.5k rows<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
It achieves a 2x compression ratio over binary ply, and roughly the
same compression ratio as zipping the binary ply.<br>
<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
More suprisingly, uncompressed patches occupy only 14% more
(setting the storage to external).<br>
<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
Any idea why ?<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
(by the way, for the pcpatch type, the default value should be
'external' and not 'main' I guess).<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
running 1.0.1<br>
<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
Cheers,<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
Rémi C<br></div>
<div class="gmail_default" style="font-family:monospace,monospace">
<br>
<br></div>
</div>
_______________________________________________<br>pgpointcloud mailing list<br>pgpointcloud@lists.osgeo.org<br>http://lists.osgeo.org/mailman/listinfo/pgpointcloud</div></div></span></blockquote> <div id="bloop_sign_1448291086868273920" class="bloop_sign"><div style="font-family:helvetica,arial;font-size:13px">-- <br>Paul Ramsey<br>Sent with Airmail</div></div></body></html>