For comparison:<br><br>Gosmore compiles the who planet in 7hours 16minutes using only 1 core.<br><br>Before I run gosmore I split the planet into 58 rectangles which will probably be another 4 hours of osmosis. Splitting the planet into countries will require a similar amount of computing power.<br>
<br><div class="gmail_quote">2010/1/21 Tobias Wendorff <span dir="ltr"><<a href="mailto:tobias.wendorff@uni-dortmund.de">tobias.wendorff@uni-dortmund.de</a>></span><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Stefan Dettenhofer (StefanDausR) schrieb:<div class="im"><br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Rechenzeiten (Entpacken osm.bz2, maps erzeugen, packen als 7z, txt-Datei für OpenLayers erzeugen):<br>
</blockquote>
<br></div>
Verwendest Du 7z mit Multicore oder normal? Was ist denn besser:<br>
a) schnell über mehrere Cores packen<br>
b) langsam über einen Core packen?<div><div></div><div class="h5"><br></div></div></blockquote><br></div>My suspicion is that single core will use less CPU seconds because of less overhead e.g. synchronizing. But multicore may mean less cache misses as the compression job "crowds out" other jobs. Difficult to analize and and test will be somewhat subjective.<br>
<br>Regards,<br>Nic<br><br>