Changeset 197 for palm/trunk/DOC/app/chapter_4.2.html
- Timestamp:
- Sep 16, 2008 3:29:03 PM (16 years ago)
- File:
-
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
palm/trunk/DOC/app/chapter_4.2.html
r189 r197 2575 2575 2576 2576 2577 <tr> 2577 <tr><td align="undefined" valign="undefined"><span style="font-style: italic; color: rgb(255, 0, 0);">p</span></td><td align="undefined" valign="undefined">Perturbation pressure (in Pa)</td></tr><tr> 2578 2578 2579 2579 … … 3717 3717 assigned. This 3718 3718 yields a loss of accuracy, but the file size is clearly reduced. The 3719 parameter <a href="chapter_4.2.html#do3d_ precision">do3d_precision</a>3719 parameter <a href="chapter_4.2.html#do3d_comp_prec">do3d_comp_prec</a> 3720 3720 can be used to separately define the number of significant digits for 3721 3721 each quantity.<br> … … 3747 3747 3748 3748 3749 <p><a name="do3d_ precision"></a><b>do3d_precision</b></p>3749 <p><a name="do3d_comp_prec"></a><b>do3d_comp_prec</b></p> 3750 3750 3751 3751 … … 3817 3817 3818 3818 3819 <p>The default assignment is <span style="font-weight: bold;">do3d_ precision</span>3819 <p>The default assignment is <span style="font-weight: bold;">do3d_comp_prec</span> 3820 3820 = <span style="font-style: italic;">'u2'</span>, <span style="font-style: italic;">'v2'</span>, <span style="font-style: italic;">'w2'</span>, <span style="font-style: italic;">'p5'</span>, <span style="font-style: italic;">'pt2'</span>.</p> 3821 3821 … … 4817 4817 4818 4818 <td style="vertical-align: top;">Subgrid-scale sensible 4819 heat flux near the surface (horizontal 4820 average) 4821 between z = 0 and z = z<sub>p</sub> = zu(1) (there it 4822 corresponds to 4823 the total heat flux) (in K m/s).</td> 4819 heat flux at k=0 (horizontal 4820 average), constant within Prandtl-layer (in K m/s).</td> 4824 4821 4825 4822 … … 4856 4853 4857 4854 4858 <tr> 4855 <tr><td align="undefined" valign="undefined"><span style="font-style: italic;">w"u"0</span></td><td align="undefined" valign="undefined">Subgrid-scale momentum flux (u-component) at k=0 (horizontal average), constant within Prandtl-layer (in m<sup>2</sup>/s<sup>2</sup>).</td></tr><tr><td align="undefined" valign="undefined"><span style="font-style: italic;">w"v"0</span></td><td align="undefined" valign="undefined">Subgrid-scale momentum flux (v-component) at k=0 (horizontal average), constant within Prandtl-layer (in m<sup>2</sup>/s<sup>2</sup>).</td></tr><tr><td align="undefined" valign="undefined"><span style="font-style: italic;">w"q"0</span></td><td align="undefined" valign="undefined">Subgrid-scale 4856 humidity flux at k=0 (horizontal average), constant within 4857 Prandtl-layer (in kg/kg m/s). Zero values are output if humidity is not 4858 used.</td></tr><tr> 4859 4859 4860 4860 … … 4886 4886 4887 4887 4888 <tr> 4889 4890 4891 4892 <td style="vertical-align: top; font-style: italic;">splptx</td> 4893 4894 4895 4896 <td style="vertical-align: top;">Percentage of grid points 4897 using upstream scheme along x with 4898 upstream-spline advection switched on.</td> 4899 4900 4901 </tr> 4902 4903 4904 <tr> 4905 4906 4907 4908 <td style="vertical-align: top; font-style: italic;">splpty</td> 4909 4910 4911 4912 <td style="vertical-align: top;">Percentage of grid points 4913 using upstream scheme along y with 4914 upstream-spline 4915 advection switched on.</td> 4916 4917 4918 </tr> 4919 4920 4921 <tr> 4922 4923 4924 <td style="vertical-align: top; font-style: italic;">splptz</td> 4925 4926 4927 4928 <td style="vertical-align: top;">Percentage of grid points 4929 using upstream scheme along z with 4930 upstream-spline 4931 advection switched on.<br> 4932 4933 4934 </td> 4935 4936 4937 </tr> 4888 4889 4890 4891 4892 4893 4894 4938 4895 4939 4896 … … 5640 5597 It is only possible to gather data from a level larger than the one 5641 5598 determined automatically. A test run may be neccessary to determine 5642 this level.< /td>5599 this level.<br><br>Setting of <span style="font-weight: bold;">mg_switch_to_pe0_level</span> = <span style="font-style: italic;">-1</span> prevents that data are collected on PE0 at all, i.e. coarsening of grids is limited by the subdomains.</td> 5643 5600 5644 5601 … … 5848 5805 5849 5806 5850 <tr> 5807 <tr><td style="vertical-align: top;"><a href="npex"><span style="font-weight: bold;"></span></a><span style="font-weight: bold;"><a name="npex"></a>npex</span></td><td style="vertical-align: top;">I</td><td style="vertical-align: top;"></td><td style="vertical-align: top;"><p>Number of processors 5808 along x-direction of the virtual 5809 processor 5810 net. </p> 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 <p>For parallel runs, the total 5825 number of processors to be used 5826 is given by 5827 the <span style="font-weight: bold;">mrun</span> 5828 option <a href="http://www.muk.uni-hannover.de/software/mrun_beschreibung.html#Opt-X">-X</a>. 5829 By default, depending on the type of the parallel computer, PALM 5830 generates a 1d processor 5831 net (domain decomposition along x, <span style="font-weight: bold;">npey</span> 5832 = <span style="font-style: italic;">1</span>) or a 5833 2d-net (this is 5834 favored on machines with fast communication network and/or large number of processors (>256)). In case of a 5835 2d-net, it is tried to make it more or less square-shaped. If, for 5836 example, 16 processors are assigned (-X 16), a 4 * 4 processor net is 5837 generated (<span style="font-weight: bold;">npex</span> 5838 = <span style="font-style: italic;">4</span>, <span style="font-weight: bold;">npey</span> 5839 = <span style="font-style: italic;">4</span>). 5840 This choice is optimal for square total domains (<a href="chapter_4.1.html#nx">nx</a> 5841 = <a href="chapter_4.1.html#ny">ny</a>), 5842 since then the number of ghost points at the lateral boundarys of 5843 the subdomains reaches a minimum. If <span style="font-weight: bold;">nx</span> 5844 and <span style="font-weight: bold;">ny</span> 5845 differ extremely, the 5846 processor net should be manually adjusted using adequate values for <span style="font-weight: bold;">npex</span> and <span style="font-weight: bold;">npey</span>. </p> 5847 5848 5849 5850 5851 5852 5853 5854 5855 5856 5857 5858 5859 5860 <p><b>Important:</b> The value of <span style="font-weight: bold;">npex</span> * <span style="font-weight: bold;">npey</span> must exactly match the 5861 value assigned by the <span style="font-weight: bold;">mrun</span>-option 5862 <tt>-X</tt>. 5863 Otherwise the model run will abort with a corresponding error 5864 message. <br> 5865 5866 5867 5868 5869 5870 5871 5872 Additionally, the specification of <span style="font-weight: bold;">npex</span> 5873 and <span style="font-weight: bold;">npey</span> 5874 may of course 5875 override the default setting for the domain decomposition (1d or 2d) 5876 which may have a significant (negative) effect on the code performance. 5877 </p></td></tr><tr><td style="vertical-align: top;"><span style="font-weight: bold;"><a name="npey"></a>npey</span></td><td style="vertical-align: top;">I</td><td style="vertical-align: top;"></td><td style="vertical-align: top;"><p>Number of processors 5878 along y-direction of the virtual 5879 processor 5880 net. </p> 5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893 5894 <p>For further information see <a href="chapter_4.2.html#npex">npex</a>.</p></td></tr><tr> 5851 5895 5852 5896
Note: See TracChangeset
for help on using the changeset viewer.