| abstract
| - It is possible to take advantage of modern multi-core processors to compile maps much faster by using multithreading. A quick way to see if your machine has got multiple cores (or multiple processors), is:
* Microsoft Windows: to take a look to the "Performance" tab of the "Task manager" tool (from Ctrl+Alt+Del). If the "CPU utilization history" is divided into two or more graphs, your machine as got at least two or more cores (physical or logical).Note: up to Windows 7, under "View" -> "CPU History" menu from Task manager "Perfomance" tab, you can find two options: "One Graph, All CPUs" and "One Graph per CPU": the second one is the one enabled by default required for this test. In case you have "One Graph, ALL CPUs" checked instead, you will NOT be able to find out the number of cores, unless you modify the option. If you are using Windows 8/10, you will not find the "CPU history" item under the "View" menu: you have to right-click on the CPU graph and choose "Change graph to" -> "Logical processors" instead (by default, a single graph is showed there!). However Windows 8/10 Task Manager, under "Performance" -> "CPU" screen, lists the number of physical processors, cores and logical processors as simple text right under the graph.
* GNU/Linux: run the "top" command from a Linux "shell" (command line interface). After pressing the "1" key, in the upper part of the screen, the "Cpu(s)" line should change into multiple lines ("Cpu0", "Cpu1", "Cpu2"...), one for each core (physical or logical). If you have got more than one core, you may find useful to compile maps (q3map/q3map2) using more threads, in other words splitting the process into separate sections that can run at the same time on multi-core systems. This way, you should see higher CPU load in Task Manager/top while compiling, and you should be able to compile maps much faster than using a single thread, because each thread should work on a separate core (with a single thread, only one core would be used). So, choose the number of threads (1, 2, 3, 4, 5, 6...) depending from how many cores you have got and how many of them you want to use for this purpose: using all available cores should give you faster compiling. On the other hand, leaving at least one physical core "free" should allow you to do not experience slowdowns using the computer do to other things (e.g. playing a game or surfing the web) while you are compiling a map; it's your decision. You may wish to create various custom pre-sets using different numbers of threads, and choose each time which one to use depending from how much processing power you wish to have free for other applications.
|
![[RDF Data]](/fct/images/sw-rdf-blue.png)