You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
<spanid="requirements"></span><h1>Requirements and Quickstart<aclass="headerlink" href="#requirements-and-quickstart" title="Link to this heading"></a></h1>
96
96
<sectionid="hardware-and-platform">
97
97
<h2>Hardware and Platform<aclass="headerlink" href="#hardware-and-platform" title="Link to this heading"></a></h2>
98
-
<p>This page summarizes the supported mainline workflow of <codeclass="docutils literal notranslate"><spanclass="pre">Streamcenter+</span></code>.
99
-
The current mainline is no longer the old Python package set. It is centered on a
100
-
native CUDA FTLE solver together with a native Qt/VTK desktop GUI.
101
-
The archived Python implementation is still preserved under <codeclass="docutils literal notranslate"><spanclass="pre">legacy/</span></code> in the code
102
-
repository for reference and migration, but it is no longer the primary workflow
103
-
described in this documentation.</p>
98
+
<p><codeclass="docutils literal notranslate"><spanclass="pre">Streamcenter+</span></code> consists of a native CUDA FTLE solver together with a native
99
+
Qt/VTK desktop GUI.</p>
104
100
<p>The minimum practical requirements are listed as follows:</p>
105
101
<ulclass="simple">
106
102
<li><p>At least 4 GB of system memory and about 10 GB of free disk space.</p></li>
107
103
<li><p>A 64-bit Windows or Linux environment. The shipped install scripts currently target
108
104
Windows and Debian/Ubuntu-style Linux systems with <codeclass="docutils literal notranslate"><spanclass="pre">apt</span></code>.</p></li>
109
-
<li><p>An NVIDIA GPU together with a locally installed CUDA Toolkit for the mainline solver.</p></li>
105
+
<li><p>An NVIDIA GPU together with a locally installed CUDA Toolkit for the solver.</p></li>
110
106
</ul>
111
107
<p>The solver configures <codeclass="docutils literal notranslate"><spanclass="pre">CMAKE_CUDA_ARCHITECTURES=native</span></code> by default, so the build
112
108
machine is expected to have a visible local NVIDIA CUDA environment.</p>
@@ -116,14 +112,11 @@ <h2>Hardware and Platform<a class="headerlink" href="#hardware-and-platform" tit
116
112
below roughly 200,000,000 Lagrangian particles per GiB of VRAM for the default
117
113
double-sided still workflow, and lower than that for higher-order settings.</p>
118
114
</section>
119
-
<sectionid="mainline-environment">
120
-
<h2>Mainline Environment<aclass="headerlink" href="#mainline-environment" title="Link to this heading"></a></h2>
121
-
<p>The native solver and the native GUI do not require Python or <codeclass="docutils literal notranslate"><spanclass="pre">pip</span></code> for their
122
-
standard build-and-run path.
123
-
Instead, the repository provides platform-specific installation scripts that prepare
124
-
the CMake/CUDA/Qt/VTK toolchain directly.</p>
125
-
<p>On Windows, <codeclass="docutils literal notranslate"><spanclass="pre">install.cmd</span></code> and <codeclass="docutils literal notranslate"><spanclass="pre">install.ps1</span></code> provision the following mainline
126
-
dependencies:</p>
115
+
<sectionid="environment">
116
+
<h2>Environment<aclass="headerlink" href="#environment" title="Link to this heading"></a></h2>
117
+
<p>The repository provides platform-specific installation scripts that prepare the
118
+
CMake/CUDA/Qt/VTK toolchain directly.</p>
119
+
<p>On Windows, <codeclass="docutils literal notranslate"><spanclass="pre">install.cmd</span></code> and <codeclass="docutils literal notranslate"><spanclass="pre">install.ps1</span></code> provision the following dependencies:</p>
<p>For migration convenience, the renamed build scripts still accept the legacy
172
-
<codeclass="docutils literal notranslate"><spanclass="pre">PY3DFTLE_*</span></code> environment variables as fallback inputs.</p>
173
164
</section>
174
165
<sectionid="add-on-tools">
175
166
<h2>Add-on Tools<aclass="headerlink" href="#add-on-tools" title="Link to this heading"></a></h2>
@@ -191,7 +182,7 @@ <h2>Add-on Tools<a class="headerlink" href="#add-on-tools" title="Link to this h
191
182
</section>
192
183
<sectionid="quickstart">
193
184
<h2>Quickstart<aclass="headerlink" href="#quickstart" title="Link to this heading"></a></h2>
194
-
<p>The steps below follow the current native mainline repository layout.
185
+
<p>The steps below follow the current repository layout.
195
186
They use the top-level install and build entry points and then launch the GUI or solver
196
187
through the maintained wrapper scripts.</p>
197
188
<sectionid="windows">
@@ -248,8 +239,7 @@ <h3>Linux (Debian/Ubuntu)<a class="headerlink" href="#linux-debian-ubuntu" title
248
239
</section>
249
240
<sectionid="notes-for-real-cases">
250
241
<h2>Notes for Real Cases<aclass="headerlink" href="#notes-for-real-cases" title="Link to this heading"></a></h2>
251
-
<p>The mainline GUI always launches the native CUDA solver. There is no longer a
252
-
Python/native backend selector in the supported workflow.</p>
242
+
<p>The GUI launches the native CUDA solver.</p>
253
243
<p>The parameter format remains <codeclass="docutils literal notranslate"><spanclass="pre">.par</span></code>-based.
254
244
When preparing your own cases, update entries such as <codeclass="docutils literal notranslate"><spanclass="pre">vtk_folder</span></code> and
255
245
<codeclass="docutils literal notranslate"><spanclass="pre">output_dir</span></code> so that they point to locations that actually exist on your machine.
0 commit comments