<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 12pt;
font-family:Calibri
}
--></style></head>
<body class='hmmessage'><div dir='ltr'>Hi Felix,<br><br>try o5m for both input and output, it is much faster. <br>The command <br>osmcovert --drop-version europe-latest.osm.pbf -o=europe.o5m<br>runs quite fast (~70 seconds on my machine), <br>the o5m file is ~2.430 MB, the pbf file has 2.104 MB.<br><br>Splitter is much faster reading from o5m when <br>the keep-complete option is in use.<br>(210 secs for the o5m, 441 for pbf)<br><br>With --output=pbf both are slower, and mkgmap is also much slower.<br><br>All times with I/O on one normal hard disk. Even better results if you have <br>two different disks for reading and writing.<br><br>Gerd<br><br><div><hr id="stopSpelling">Date: Wed, 7 May 2014 11:37:58 +0200<br>From: extremecarver@gmail.com<br>To: mkgmap-dev@lists.mkgmap.org.uk<br>Subject: Re: [mkgmap-dev] splitter r325: improved split algo and new option<br><br>
  
    
  
  
    Well I still use pbf and not o5m.<br>
    First pbf is smaller..<br>
    Second - Geofabrik only offers pbf - that's why I stayed with it. <br>
    <br>
    I don't think I can cut a lot of time by first converting to 05m,
    then hand it over to splitter...<br>
    Actually I also let splitter output pbf... Maybe I could change that
    in future to 05m..<br>
    <div class="ecxmoz-cite-prefix">On 07.05.2014 11:36, Gerd Petermann
      wrote:<br>
    </div>
    <blockquote cite="mid:DUB112-W144294AF4FCD19988D46539E4E0@phx.gbl">
      <style><!--
.ExternalClass .ecxhmmessage P {
padding:0px;
}

.ExternalClass body.ecxhmmessage {
font-size:12pt;
font-family:Calibri;
}

--></style>
      <div dir="ltr">Hi Felix,<br>
        <br>
        well, nowadays splitter performance mostly depends on I/O if you
        use o5m format<br>
        for input and output and give enough heap. <br>
        <br>
        Reg. mkgmap performance improvements: yes, that's what I
        expected.<br>
        In short, the branch improved the evaluation of tags and the
        creation of the NOD file.<br>
        <br>
        Gerd<br>
        <br>
        <br>
        <div>
          <hr id="ecxstopSpelling">Date: Wed, 7 May 2014 11:29:10 +0200<br>
          From: <a class="ecxmoz-txt-link-abbreviated" href="mailto:extremecarver@gmail.com">extremecarver@gmail.com</a><br>
          To: <a class="ecxmoz-txt-link-abbreviated" href="mailto:mkgmap-dev@lists.mkgmap.org.uk">mkgmap-dev@lists.mkgmap.org.uk</a><br>
          Subject: Re: [mkgmap-dev] splitter r325: improved split algo
          and new option<br>
          <br>
          Well - I'll update all my maps on Thursday again, to recheck.
          Maybe it has to do with increasing-maxnodes? Though I thought
          the higher the max-nodes, the faster...<br>
          And I only meant splitter. I upgraded mkgmap at the same time
          (now integrating performance branch changes) - so mkgmap by
          itself got faster (though it depends on the country - seems
          like well mapped countries profit a lot more (e.g. Austria
          like 30% time off), than countries where few continue commands
          will be in action cause their mapping is basic like Asia).<br>
          <br>
          I'm not using any pre-split files or cached files of any sort
          either...<br>
          <div class="ecxmoz-cite-prefix">On 07.05.2014 06:49, Gerd
            Petermann wrote:<br>
          </div>
          <blockquote cite="mid:DUB112-W46EB8056828BFF004A1CD09E4E0@phx.gbl">
            <style><!--
.ExternalClass .ecxhmmessage P {
padding:0px;
}

.ExternalClass body.ecxhmmessage {
font-size:12pt;
font-family:Calibri;
}


--></style>
            <div dir="ltr">Hi Felix,<br>
              <br>
              reg. speed: I can't reproduce that. I compared a split of
              Germany, <br>
              both versions (r321 and r325) are more or less running the
              same time.<br>
              (I've executed both programs two times to make sure that
              disk caches <br>
              are not causing big differences)<br>
              <br>
              Or did you mean the combination of splitter + mkgmap to
              process e.g. Asia?<br>
              <br>
              Gerd<br>
              <br>
              <div>
                <hr id="ecxstopSpelling">Date: Tue, 6 May 2014 18:22:00
                +0200<br>
                From: <a class="ecxmoz-txt-link-abbreviated" href="mailto:extremecarver@gmail.com">extremecarver@gmail.com</a><br>
                To: <a class="ecxmoz-txt-link-abbreviated" href="mailto:mkgmap-dev@lists.mkgmap.org.uk">mkgmap-dev@lists.mkgmap.org.uk</a><br>
                Subject: Re: [mkgmap-dev] splitter r325: improved split
                algo and new option<br>
                <br>
                Seems to be much better now. I don't think I can
                increase the max-nodes value though, but for most maps
                the new algo creates less tiles for the same max-nodes
                value (e.g. Austria from 43 down to 35 for me, with the
                smallest tile now around 5MB instead of 2.8, and the
                biggest 12MB instead of 11MB, for Asia I simultaneously
                increased max-nodes from 800k to 900k- so I'm down from
                624 tiles to 493.... and size from 970KB-16MB to now ).
                So it still seems to depend on the country, but it's
                already a lot better...<br>
                It's a bit slower (about 10% more time) <br>
                <br>
                <div class="ecxmoz-cite-prefix">On 06.05.2014 13:56,
                  Gerd Petermann wrote:<br>
                </div>
                <blockquote cite="mid:DUB112-W111F5858AD65083970BEE389E4F0@phx.gbl">
                  <style><!--
.ExternalClass .ecxhmmessage P {
padding:0px;
}

.ExternalClass body.ecxhmmessage {
font-size:12pt;
font-family:Calibri;
}


--></style>
                  <div dir="ltr">Hi all,<br>
                    <br>
                    I've applied num-tiles-v1.patch&nbsp; and improved the
                    split algo, see<br>
                    <a href="http://gis.19327.n5.nabble.com/mkgmap-ToDo-list-tp5803388p5805165.html" target="_blank">http://gis.19327.n5.nabble.com/mkgmap-ToDo-list-tp5803388p5805165.html</a><br>
                    <br>
                    <br>
                    It is now less likely that splitter creates tiles
                    with a low number of <br>
                    nodes, it is more likely that all tiles have nearly&nbsp;
                    the same number of nodes,<br>
                    and typically you will see fewer tiles.<br>
                    Maybe this also means that you can increase the
                    max-nodes value.<br>
                    <br>
                    I hope this also reduces the need for complex
                    interactions between <br>
                    spltter and mkgmap.<br>
                    <br>
                    <br>
                    <br>
                    Gerd<br>
                  </div>
                  <br>
                  <fieldset class="ecxmimeAttachmentHeader"></fieldset>
                  <br>
                  <pre>_______________________________________________
mkgmap-dev mailing list
<a class="ecxmoz-txt-link-abbreviated" href="mailto:mkgmap-dev@lists.mkgmap.org.uk">mkgmap-dev@lists.mkgmap.org.uk</a>
<a class="ecxmoz-txt-link-freetext" href="http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev" target="_blank">http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev</a></pre>
                </blockquote>
                <br>
                <pre class="ecxmoz-signature">-- 
keep on biking and discovering new trails

Felix
openmtbmap.org &amp; <a class="ecxmoz-txt-link-abbreviated" href="http://www.velomap.org" target="_blank">www.velomap.org</a></pre>
                <br>
                _______________________________________________
                mkgmap-dev mailing list <a class="ecxmoz-txt-link-abbreviated" href="mailto:mkgmap-dev@lists.mkgmap.org.uk">mkgmap-dev@lists.mkgmap.org.uk</a>
                <a class="ecxmoz-txt-link-freetext" href="http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev" target="_blank">http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev</a></div>
            </div>
            <br>
            <fieldset class="ecxmimeAttachmentHeader"></fieldset>
            <br>
            <pre>_______________________________________________
mkgmap-dev mailing list
<a class="ecxmoz-txt-link-abbreviated" href="mailto:mkgmap-dev@lists.mkgmap.org.uk">mkgmap-dev@lists.mkgmap.org.uk</a>
<a class="ecxmoz-txt-link-freetext" href="http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev" target="_blank">http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev</a></pre>
          </blockquote>
          <br>
          <pre class="ecxmoz-signature">-- 
keep on biking and discovering new trails

Felix
openmtbmap.org &amp; <a class="ecxmoz-txt-link-abbreviated" href="http://www.velomap.org" target="_blank">www.velomap.org</a></pre>
          <br>
          _______________________________________________
          mkgmap-dev mailing list
          <a class="ecxmoz-txt-link-abbreviated" href="mailto:mkgmap-dev@lists.mkgmap.org.uk">mkgmap-dev@lists.mkgmap.org.uk</a>
          <a class="ecxmoz-txt-link-freetext" href="http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev" target="_blank">http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev</a></div>
      </div>
      <br>
      <fieldset class="ecxmimeAttachmentHeader"></fieldset>
      <br>
      <pre>_______________________________________________
mkgmap-dev mailing list
<a class="ecxmoz-txt-link-abbreviated" href="mailto:mkgmap-dev@lists.mkgmap.org.uk">mkgmap-dev@lists.mkgmap.org.uk</a>
<a class="ecxmoz-txt-link-freetext" href="http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev" target="_blank">http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev</a></pre>
    </blockquote>
    <br>
    <pre class="ecxmoz-signature">-- 
keep on biking and discovering new trails

Felix
openmtbmap.org &amp; <a class="ecxmoz-txt-link-abbreviated" href="http://www.velomap.org" target="_blank">www.velomap.org</a></pre>
  

<br>_______________________________________________
mkgmap-dev mailing list
mkgmap-dev@lists.mkgmap.org.uk
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev</div>                                               </div></body>
</html>