|
128 | 128 | <div class="contents"> |
129 | 129 | <div class="textblock"><p><a class="anchor" id="autotoc_md36"></a> </p> |
130 | 130 | <h1><a class="anchor" id="autotoc_md37"></a> |
131 | | -Shu-Osher problem (1D)</h1> |
132 | | -<p>Reference: C. W. Shu, S. Osher, Efficient implementation of essentially non-oscillatory shock-capturing schemes, Journal of Computational Physics 77 (2) (1988) 439–471. doi:10.1016/0021-9991(88)90177-5.</p> |
| 131 | +2D Riemann Test (2D)</h1> |
| 132 | +<p>Reference: Chamarthi, A., & Hoffmann, N., & Nishikawa, H., & Frankel S. (2023). Implicit gradients based conservative numerical scheme for compressible flows. arXiv:2110.05461</p> |
133 | 133 | <h2><a class="anchor" id="autotoc_md38"></a> |
| 134 | +Density Initial Condition</h2> |
| 135 | +<div class="image"> |
| 136 | +<img src="alpha_rho1_initial-2D_riemann_test-example.png" alt=""/> |
| 137 | +<div class="caption"> |
| 138 | +Density</div></div> |
| 139 | + <h2><a class="anchor" id="autotoc_md39"></a> |
| 140 | +Density Final Condition</h2> |
| 141 | +<div class="image"> |
| 142 | +<img src="alpha_rho1_final-2D_riemann_test-example.png" alt=""/> |
| 143 | +<div class="caption"> |
| 144 | +Density Norms</div></div> |
| 145 | + <h1><a class="anchor" id="autotoc_md40"></a> |
| 146 | +Titarev-Toro problem (1D)</h1> |
| 147 | +<p>Reference: V. A. Titarev, E. F. Toro, Finite-volume WENO schemes for three-dimensional conservation laws, Journal of Computational Physics 201 (1) (2004) 238–260.</p> |
| 148 | +<h2><a class="anchor" id="autotoc_md41"></a> |
134 | 149 | Initial Condition</h2> |
135 | 150 | <div class="image"> |
136 | | -<img src="initial-1D_shuosher-example.png" alt=""/> |
| 151 | +<img src="initial-1D_titarevtorro-example.png" alt=""/> |
137 | 152 | <div class="caption"> |
138 | 153 | Initial Condition</div></div> |
139 | | - <h2><a class="anchor" id="autotoc_md39"></a> |
| 154 | + <h2><a class="anchor" id="autotoc_md42"></a> |
140 | 155 | Result</h2> |
141 | 156 | <div class="image"> |
142 | | -<img src="result-1D_shuosher-example.png" alt=""/> |
| 157 | +<img src="result-1D_titarevtorro-example.png" alt=""/> |
143 | 158 | <div class="caption"> |
144 | 159 | Result</div></div> |
145 | | - <h1><a class="anchor" id="autotoc_md40"></a> |
146 | | -3D Weak Scaling</h1> |
147 | | -<p>The <a href="case.py"><b>3D_weak_scaling</b></a> case depends on two parameters:</p> |
148 | | -<ul> |
149 | | -<li><b>The number of MPI ranks</b> (<em>procs</em>): As <em>procs</em> increases, the problem size per rank remains constant. <em>procs</em> is determined using information provided to the case file by <code>mfc.sh run</code>.</li> |
150 | | -<li><b>GPU memory usage per rank</b> (<em>gbpp</em>): As <em>gbpp</em> increases, the problem size per rank increases and the number of timesteps decreases so that wall times consistent. <em>gbpp</em> is a user-defined optional argument to the <a href="case.py">case.py</a> file. It can be specified right after the case filepath when invoking <code>mfc.sh run</code>.</li> |
151 | | -</ul> |
152 | | -<p>Weak scaling benchmarks can be produced by keeping <em>gbpp</em> constant and varying <em>procs</em>.</p> |
153 | | -<p>For example, to run a weak scaling test that uses ~4GB of GPU memory per rank on 8 2-rank nodes with case optimization, one could:</p> |
154 | | -<div class="fragment"><div class="line">./mfc.sh run examples/3D_weak_scaling/case.py 4 -t pre_process simulation \</div> |
155 | | -<div class="line"> -e batch -p mypartition -N 8 -n 2 -w "01:00:00" -# "MFC Weak Scaling" \</div> |
156 | | -<div class="line"> --case-optimization -j 32</div> |
157 | | -</div><!-- fragment --><h1><a class="anchor" id="autotoc_md41"></a> |
| 160 | + <h1><a class="anchor" id="autotoc_md43"></a> |
158 | 161 | Isentropic vortex problem (2D)</h1> |
159 | 162 | <p>Reference: Coralic, V., & Colonius, T. (2014). Finite-volume Weno scheme for viscous compressible multicomponent flows. Journal of Computational Physics, 274, 95–121. <a href="https://doi.org/10.1016/j.jcp.2014.06.003">https://doi.org/10.1016/j.jcp.2014.06.003</a></p> |
160 | | -<h2><a class="anchor" id="autotoc_md42"></a> |
| 163 | +<h2><a class="anchor" id="autotoc_md44"></a> |
161 | 164 | Density</h2> |
162 | 165 | <div class="image"> |
163 | 166 | <img src="alpha_rho1-2D_isentropicvortex-example.png" alt=""/> |
164 | 167 | <div class="caption"> |
165 | 168 | Density</div></div> |
166 | | - <h2><a class="anchor" id="autotoc_md43"></a> |
| 169 | + <h2><a class="anchor" id="autotoc_md45"></a> |
167 | 170 | Density Norms</h2> |
168 | 171 | <div class="image"> |
169 | 172 | <img src="density_norms-2D_isentropicvortex-example.png" alt=""/> |
170 | 173 | <div class="caption"> |
171 | 174 | Density Norms</div></div> |
172 | | - <h1><a class="anchor" id="autotoc_md44"></a> |
173 | | -2D Hardcodied IC Example</h1> |
174 | | -<h2><a class="anchor" id="autotoc_md45"></a> |
| 175 | + <h1><a class="anchor" id="autotoc_md46"></a> |
| 176 | +Shock Droplet (2D)</h1> |
| 177 | +<p>Reference: Panchal et. al., A Seven-Equation Diffused Interface Method for Resolved Multiphase Flows, JCP, 475 (2023)</p> |
| 178 | +<h2><a class="anchor" id="autotoc_md47"></a> |
175 | 179 | Initial Condition</h2> |
176 | 180 | <div class="image"> |
177 | | -<img src="initial-2D_hardcodied_ic-example.png" alt=""/> |
| 181 | +<img src="initial-2D_shockdroplet-example.png" alt=""/> |
178 | 182 | <div class="caption"> |
179 | 183 | Initial Condition</div></div> |
180 | | - <h2><a class="anchor" id="autotoc_md46"></a> |
| 184 | + <h2><a class="anchor" id="autotoc_md48"></a> |
181 | 185 | Result</h2> |
182 | | -<p><img src="result-2D_hardcodied_ic-example.png" alt="" class="inline" title="Result"/> </p> |
183 | | -<h1><a class="anchor" id="autotoc_md47"></a> |
184 | | -Lid-Driven Cavity Problem (2D)</h1> |
185 | | -<p>Reference: Bezgin, D. A., & Buhendwa A. B., & Adams N. A. (2022). JAX-FLUIDS: A fully-differentiable high-order computational fluid dynamics solver for compressible two-phase flows. arXiv:2203.13760</p> |
186 | | -<p>Reference: Ghia, U., & Ghia, K. N., & Shin, C. T. (1982). High-re solutions for incompressible flow using the Navier-Stokes equations and a multigrid method. Journal of Computational Physics, 48, 387-411</p> |
187 | | -<p>Video: <a href="https://youtube.com/shorts/JEP28scZrBM?feature=share">https://youtube.com/shorts/JEP28scZrBM?feature=share</a></p> |
188 | | -<h2><a class="anchor" id="autotoc_md48"></a> |
189 | | -Final Condition</h2> |
190 | | -<div class="image"> |
191 | | -<img src="final_condition-2D_lid_driven_cavity-example.png" alt=""/> |
192 | | -<div class="caption"> |
193 | | -Final Condition</div></div> |
194 | | - <h2><a class="anchor" id="autotoc_md49"></a> |
195 | | -Centerline Velocities</h2> |
196 | | -<div class="image"> |
197 | | -<img src="centerline_velocities-2D_lid_driven_cavity-example.png" alt=""/> |
198 | | -<div class="caption"> |
199 | | -Centerline Velocities</div></div> |
200 | | - <h1><a class="anchor" id="autotoc_md50"></a> |
201 | | -2D Riemann Test (2D)</h1> |
202 | | -<p>Reference: Chamarthi, A., & Hoffmann, N., & Nishikawa, H., & Frankel S. (2023). Implicit gradients based conservative numerical scheme for compressible flows. arXiv:2110.05461</p> |
| 186 | +<p><img src="result-2D_shockdroplet-example.png" alt="" class="inline" title="Result"/> </p> |
| 187 | +<h1><a class="anchor" id="autotoc_md49"></a> |
| 188 | +3D Weak Scaling</h1> |
| 189 | +<p>The <a href="case.py"><b>3D_weak_scaling</b></a> case depends on two parameters:</p> |
| 190 | +<ul> |
| 191 | +<li><b>The number of MPI ranks</b> (<em>procs</em>): As <em>procs</em> increases, the problem size per rank remains constant. <em>procs</em> is determined using information provided to the case file by <code>mfc.sh run</code>.</li> |
| 192 | +<li><b>GPU memory usage per rank</b> (<em>gbpp</em>): As <em>gbpp</em> increases, the problem size per rank increases and the number of timesteps decreases so that wall times consistent. <em>gbpp</em> is a user-defined optional argument to the <a href="case.py">case.py</a> file. It can be specified right after the case filepath when invoking <code>mfc.sh run</code>.</li> |
| 193 | +</ul> |
| 194 | +<p>Weak scaling benchmarks can be produced by keeping <em>gbpp</em> constant and varying <em>procs</em>.</p> |
| 195 | +<p>For example, to run a weak scaling test that uses ~4GB of GPU memory per rank on 8 2-rank nodes with case optimization, one could:</p> |
| 196 | +<div class="fragment"><div class="line">./mfc.sh run examples/3D_weak_scaling/case.py 4 -t pre_process simulation \</div> |
| 197 | +<div class="line"> -e batch -p mypartition -N 8 -n 2 -w "01:00:00" -# "MFC Weak Scaling" \</div> |
| 198 | +<div class="line"> --case-optimization -j 32</div> |
| 199 | +</div><!-- fragment --><h1><a class="anchor" id="autotoc_md50"></a> |
| 200 | +Lax shock tube problem (1D)</h1> |
| 201 | +<p>Reference: P. D. Lax, Weak solutions of nonlinear hyperbolic equations and their numerical computation, Communications on pure and applied mathematics 7 (1) (1954) 159–193.</p> |
203 | 202 | <h2><a class="anchor" id="autotoc_md51"></a> |
204 | | -Density Initial Condition</h2> |
| 203 | +Initial Condition</h2> |
205 | 204 | <div class="image"> |
206 | | -<img src="alpha_rho1_initial-2D_riemann_test-example.png" alt=""/> |
| 205 | +<img src="initial-1D_laxshocktube-example.png" alt=""/> |
207 | 206 | <div class="caption"> |
208 | | -Density</div></div> |
| 207 | +Initial Condition</div></div> |
209 | 208 | <h2><a class="anchor" id="autotoc_md52"></a> |
210 | | -Density Final Condition</h2> |
| 209 | +Result</h2> |
211 | 210 | <div class="image"> |
212 | | -<img src="alpha_rho1_final-2D_riemann_test-example.png" alt=""/> |
| 211 | +<img src="result-1D_laxshocktube-example.png" alt=""/> |
213 | 212 | <div class="caption"> |
214 | | -Density Norms</div></div> |
| 213 | +Result</div></div> |
215 | 214 | <h1><a class="anchor" id="autotoc_md53"></a> |
216 | | -Shock Droplet (2D)</h1> |
217 | | -<p>Reference: Panchal et. al., A Seven-Equation Diffused Interface Method for Resolved Multiphase Flows, JCP, 475 (2023)</p> |
| 215 | +2D Hardcodied IC Example</h1> |
218 | 216 | <h2><a class="anchor" id="autotoc_md54"></a> |
219 | 217 | Initial Condition</h2> |
220 | 218 | <div class="image"> |
221 | | -<img src="initial-2D_shockdroplet-example.png" alt=""/> |
| 219 | +<img src="initial-2D_hardcodied_ic-example.png" alt=""/> |
222 | 220 | <div class="caption"> |
223 | 221 | Initial Condition</div></div> |
224 | 222 | <h2><a class="anchor" id="autotoc_md55"></a> |
225 | 223 | Result</h2> |
226 | | -<p><img src="result-2D_shockdroplet-example.png" alt="" class="inline" title="Result"/> </p> |
| 224 | +<p><img src="result-2D_hardcodied_ic-example.png" alt="" class="inline" title="Result"/> </p> |
227 | 225 | <h1><a class="anchor" id="autotoc_md56"></a> |
228 | | -Titarev-Toro problem (1D)</h1> |
229 | | -<p>Reference: V. A. Titarev, E. F. Toro, Finite-volume WENO schemes for three-dimensional conservation laws, Journal of Computational Physics 201 (1) (2004) 238–260.</p> |
| 226 | +Lid-Driven Cavity Problem (2D)</h1> |
| 227 | +<p>Reference: Bezgin, D. A., & Buhendwa A. B., & Adams N. A. (2022). JAX-FLUIDS: A fully-differentiable high-order computational fluid dynamics solver for compressible two-phase flows. arXiv:2203.13760</p> |
| 228 | +<p>Reference: Ghia, U., & Ghia, K. N., & Shin, C. T. (1982). High-re solutions for incompressible flow using the Navier-Stokes equations and a multigrid method. Journal of Computational Physics, 48, 387-411</p> |
| 229 | +<p>Video: <a href="https://youtube.com/shorts/JEP28scZrBM?feature=share">https://youtube.com/shorts/JEP28scZrBM?feature=share</a></p> |
230 | 230 | <h2><a class="anchor" id="autotoc_md57"></a> |
231 | | -Initial Condition</h2> |
| 231 | +Final Condition</h2> |
232 | 232 | <div class="image"> |
233 | | -<img src="initial-1D_titarevtorro-example.png" alt=""/> |
| 233 | +<img src="final_condition-2D_lid_driven_cavity-example.png" alt=""/> |
234 | 234 | <div class="caption"> |
235 | | -Initial Condition</div></div> |
| 235 | +Final Condition</div></div> |
236 | 236 | <h2><a class="anchor" id="autotoc_md58"></a> |
237 | | -Result</h2> |
| 237 | +Centerline Velocities</h2> |
238 | 238 | <div class="image"> |
239 | | -<img src="result-1D_titarevtorro-example.png" alt=""/> |
| 239 | +<img src="centerline_velocities-2D_lid_driven_cavity-example.png" alt=""/> |
240 | 240 | <div class="caption"> |
241 | | -Result</div></div> |
| 241 | +Centerline Velocities</div></div> |
242 | 242 | <h1><a class="anchor" id="autotoc_md59"></a> |
243 | | -Lax shock tube problem (1D)</h1> |
244 | | -<p>Reference: P. D. Lax, Weak solutions of nonlinear hyperbolic equations and their numerical computation, Communications on pure and applied mathematics 7 (1) (1954) 159–193.</p> |
| 243 | +Shu-Osher problem (1D)</h1> |
| 244 | +<p>Reference: C. W. Shu, S. Osher, Efficient implementation of essentially non-oscillatory shock-capturing schemes, Journal of Computational Physics 77 (2) (1988) 439–471. doi:10.1016/0021-9991(88)90177-5.</p> |
245 | 245 | <h2><a class="anchor" id="autotoc_md60"></a> |
246 | 246 | Initial Condition</h2> |
247 | 247 | <div class="image"> |
248 | | -<img src="initial-1D_laxshocktube-example.png" alt=""/> |
| 248 | +<img src="initial-1D_shuosher-example.png" alt=""/> |
249 | 249 | <div class="caption"> |
250 | 250 | Initial Condition</div></div> |
251 | 251 | <h2><a class="anchor" id="autotoc_md61"></a> |
252 | 252 | Result</h2> |
253 | 253 | <div class="image"> |
254 | | -<img src="result-1D_laxshocktube-example.png" alt=""/> |
| 254 | +<img src="result-1D_shuosher-example.png" alt=""/> |
255 | 255 | <div class="caption"> |
256 | 256 | Result</div></div> |
257 | 257 | </div></div><!-- contents --> |
|
0 commit comments