Skip to content

Latest commit

 

History

History
354 lines (312 loc) · 18.7 KB

README.md

File metadata and controls

354 lines (312 loc) · 18.7 KB
Raw
                       GLVis visualization tool

                 _/_/_/  _/      _/      _/  _/
              _/        _/      _/      _/        _/_/_/
             _/  _/_/  _/      _/      _/  _/  _/_/
            _/    _/  _/        _/  _/    _/      _/_/
             _/_/_/  _/_/_/_/    _/      _/  _/_/_/

                         https://glvis.org

Release Build License Doxygen License License

GLVis is an OpenGL tool for visualization of finite element meshes and functions. It is a multiplatform application that can be built on Linux/Unix systems, including macOS, and under Windows. It can also be used in a Jupyter notebook, or in a web browser, see https://glvis.org/live.

  • For building instructions, see INSTALL.

  • The GLVis key commands and mouse functions are documented below.

  • GLVis is distributed under the terms of the BSD-3 license. All new contributions must be made under this license. See LICENSE and NOTICE for details.

We welcome contributions and feedback from the community. Please see the file CONTRIBUTING.md for additional details about our development process.

When started without any options, glvis starts a server which waits for a socket connections (on port 19916 by default) and visualizes any received data. This way the results of simulations on a remote (parallel) machine can be visualized on the local user desktop.

GLVis can also be used to visualize a mesh with or without a finite element function (solution), as in

glvis -m cube.mesh3d

For parallel computations, GLVis supports input from several parallel socket connections as well as the visualization of parallel meshes and grid functions saved in separate files from the command line as in

glvis -np 4 -m mesh -g solution

When given parallel input, GLVis will stitch the results to show the global mesh and solution. GLVis can also run a batch sequence of commands (GLVis scripts), or display previously saved socket streams.

For a complete list of command line options, type

glvis -h

Depending on the data type, variety of manipulations can be performed by using the mouse and by typing (case sensitive) keystrokes in the GLVis window. Below is a partial list of the available functionality. Some of these keys can also be provided as input, using the -k command-line option and the keys script command.

For high-order meshes and/or solution data, GLVis performs element subdivision to try to represent the data more accurately. However, for highly-varying data or large meshes, the auto-selected subdivision factor (see the Auto-refinement section below) may not be sufficient -- use the keys o / O, described below, to manually adjust the subdivision factor.

SPDX-License-Identifier: BSD-3-Clause
LLNL Release Number: LLNL-CODE-443271
DOI: 10.11578/dc.20171025.1249

Mouse functions

Basic

  • Left – Rotate the viewpoint
  • Right – Zoom in (up) / Zoom out (down)
  • Middle – Translate the viewpoint
  • Left + Shift – Start spinning the viewpoint (according to the dragging vector)

Advanced

  • Left + Alt – Tilt
  • Left + Ctrl – Spherical rotation
  • Left + Ctrl + Shift – z-spinning
  • Right + Ctrl – Object scaling (see also [ and ])
  • Right + Shift – Change light source position (see \)
  • Middle + Ctrl – Object translation (moves the camera left/right/up/down)
  • Middle + Ctrl + Alt – Moves the camera forward/backward (vertical mouse motion) and tilts the camera left/right (horizontal mouse motion)
  • Middle + Ctrl + Shift – Object translation (turns the camera left/right/up/down)

Key commands

Basic

  • h – Print a short help message in the terminal
  • r – Reset the plot to 3D view
  • R – Cycle through the six 2D projections (camera looking above/below in x/y/z directions)
  • j – Turn on/off perspective
  • s – Turn on/off unit cube scaling
  • A – Turn on/off the use of anti-aliasing/multi-sampling
  • L – Turn on/off logarithmic scale
  • c – Toggle the colorbar and caption display state
  • C – Change the main plot caption
  • p / P – Cycle forward/backwards through color palettes (lots of options, use F6 for a menu)
  • t – Cycle materials and lights (5 states)
  • i – Toggle cutting plane (different options in 2D and 3D, see below)
  • o / O – Control element subdivisions (different options in 2D and 3D, see below)
  • l – Turn on/off the light
  • g – Toggle background color (white/black)
  • a – Toggle the bounding box axes. The options are:
    • none
    • bounding box with coordinates of the corners
    • bounding box without coordinates
    • red, green, blue colored main x, y, z axes + dashed axes
  • m – Toggle the mesh state. The options are:
    • no mesh or level lines
    • draw the element edges (i.e. the mesh)
    • draw the level lines (use F5 to modify the level lines)
  • e – Toggle the elements state (see below for vector functions). The options are:
    • show surface elements (corresponding to the function)
    • show no surface elements
    • (2D only) show element attributes
    • (2D only) show element det(J)
    • (2D only) show element 1/det(J)
    • (2D only) show element \kappa
    • (2D only) show element \kappa + 1/\kappa
  • S – Take an image snapshot or record a movie (in spinning mode). By default, the screenshots are taken in png format, using libpng. When GLVis is compiled with libtiff support (see INSTALL) then the screenshots are taken internally and saved in TIFF format (.tif extension). If both of these options are disabled during the build process, GLVis will use SDL to take screenshots in bmp format, which it will then convert to png if ImageMagick's convert tool is available.
  • G – 3D scene export to glTF format
  • Ctrl + p – Print to a PDF file using gl2ps. Other vector formats (SVG, EPS) are also possible, but keep in mind that the printing takes a while and the generated files are big.
  • Q – Cycle between representations of the visualized quadrature data. The options are:
    • piece-wise constant refined (LOR)
    • L2 element dof collocation (interpolation)
    • L2 element projection (L2 projection)
  • q – Exit

Advanced

  • f – Change the shading type (the way the elements and mesh are drawn). The options are:
    • one triangle / quad per element with a constant normal
    • one triangle / quad per element with normals averaged at the vertices
    • multiple triangles / quads per element, also allowing for the visualization of discontinuous functions and curvilinear elements (use o / O to control subdivisions)
  • \ – Set light source position (see Right + Shift)
  • * / / – Zoom in/out
  • + / - – Stretch/compress in z-direction
  • [ / ] – Shrink/enlarge the bounding box (relative to the colorbar)
  • ( / ) – Shrink/enlarge the visualization window
  • . – Start/stop z-spinning (speed/direction can be controlled with 0 / Enter)
  • ←, →, ↑, ↓ – Manual rotation
  • 1, 2, 3, 4, 5, 6, 7, 8, 9 – Manual rotation along coordinate axes
  • Alt + a – Set axes number format
  • Alt + c – Set colorbar number format
  • Ctrl + ←, →, ↑, ↓ – Translate the viewpoint
  • Ctrl + o – Toggle an element ordering curve
  • n / N – Cycle through numberings. The options are:
    • none
    • show element numbering
    • show edge numbering
    • show vertex numbering
  • ` – Toggle a ruler, with initial origin at the center of the bounding box. The origin can be later moved with ~. The options are:
    • none
    • coordinate axes lines
    • coordinate axes planes
  • ~ - Enter new ruler origin
  • k / K - Adjust the transparency level. The balance of transparency can be further adjusted with , and <.
  • ! - Toggle the use of (1D) texture (smooth interpolation of colors). The options are:
    • use discrete texture, the number of colors used depends on the current palette
    • use smooth texture (interpolated from current palette)
  • F5 – Change the range and number of the level lines
  • F6 – Palette menu (negative repeat number flips the palette)
  • F7 – Change the minimum and maximum values
  • Shift + F7 – Set the bounding box from the terminal
  • Ctrl + a – Cycle through the auto-scaling options:
    • off: do not change the bounding box and the value range
    • on (default): recompute both the bounding box and the value range
    • value: recompute only the value range
    • mesh: recompute only the bounding box

2D scalar data

  • i – Toggle cutting (clipping) plane in 2D
  • y / Y – Rotate cutting plane (\theta) in 2D
  • z / Z – Translate cutting plane in 2D
  • o / O – Control element subdivisions in 2D
    • there are two subdivision factors in this case: element (s1) and boundary (s2).
    • O cycles through the following "subdivision functions", (prints a message in the terminal when changed):
      • Increase element subdivision factor: s1 += s2
      • Decrease element subdivision factor: s1 -= s2
      • Increase boundary subdivision factor: s2++
      • Decrease boundary subdivision factor: s2--
    • o – performs the currently selected function
  • b – Toggle the boundary in 2D scalar mode. The options are:
    • no boundary
    • black boundary
    • boundary colored with the boundary attribute
    • Use Shift + F9 / F10 to cycle through the boundary attributes.
  • F3 / F4 – Shrink/Zoom each element towards its center, in order to better visualize the different element shapes
  • F8 – List of material subdomains to show
  • F9 / F10 – Walk through material subdomains
  • F11 / F12 – Shrink/Zoom material subdomains (to the centers of the attributes)

3D scalar data

  • i – Toggle cutting (clipping) plane in 3D. The options are:
    • none
    • cut through the elements
    • show only elements behind the cutting plane
  • I – Toggle the cutting plane algorithm used when the option cut through the elements is selected. The two algorithms are:
    • slower, more accurate algorithm for curved meshes (default)
    • faster algorithm suitable for meshes with planar faces
  • x / X – Rotate cutting plane (\phi) in 3D
  • y / Y – Rotate cutting plane (\theta) in 3D
  • z / Z – Translate cutting plane in 3D
  • E – Display/Hide the elements in the cutting plane
  • M – Display/Hide the mesh in the cutting plane
  • o / O – Refine/de-refine elements in 3D
  • u / U – Move level surface value up/down
  • v / V – Add/Delete a level surface value
  • w / W – Move boundary elements up/down in direction of the normal (i.e. "plot" the boundary values in normal direction)
  • F3 / F4 – Shrink/Zoom boundary elements (to the centers of the attributes)
  • F8 – List of boundary subdomains to show
  • F9 / F10 – Walk through boundary subdomains
  • F11 / F12 – Shrink/Zoom material subdomains (to the centers of the attributes)

2D vector data

  • v – Toggle the vector state (uses vector subdivision factor, accept u / U). The options are:
    • do not show vectors
    • show vectors as displacement
    • show vector field; vectors are uniformly scaled; the color varies with the magnitude (or the current vector-to-scalar function, see keys u / U)
    • show vector field as above, but the vectors are scaled proportionally to their magnitude
    • show vector field; vectors are uniformly scaled; the color is gray; arrows are above the surface
    • show vector field as above, but the vectors are scaled proportionally to their magnitude
  • V – Change the scaling of the vectors relative to the default
  • d – Toggle the displaced mesh state: (see also keys n / b). The options are:
    • do not show displaced mesh
    • show displaced mesh
    • assuming displacement field show deformation using Cartesian lines
    • assuming displacement field show deformation using polar lines
  • n – Increase the displacement amount in 10% steps, wraps around from 100% to 0%
  • b – Decrease the displacement amount in 10% steps, wraps around from 0% to 100%
  • B – Toggle the boundary in 2D vector mode
  • e – Toggle the elements state (vector data version). The options are:
    • show surface elements corresponding to the current vector-to-scalar function
    • do not show surface elements
    • assuming a displacement field show det(J)/det(J_d)
    • assuming a displacement field show det(J_d)/det(J)
  • u / U – Change the vector-to-scalar function and the vector subdivision factor
  • U – Toggle the functionality of u (prints a message in the terminal when changed). The options are:
    • Increase the vector subdivision factor
    • Decrease the vector subdivision factor
    • Cycle through vector-to-scalar functions choices:
      • magnitude: \sqrt{v_x^2+v_y^2}
      • direction from -\pi to \pi: atan2(v_y,v_x)
      • x-component: v_x
      • y-component: v_y
      • divergence: div(v)
      • curl: curl(v) [skipped for H(div) elements]
      • anisotropy in grad(v) [skipped for H(div) elements]

3D vector data

  • v – Toggle the vector state. The options are:
    • do not show vectors
    • show vectors as displacement
    • show vector field; vectors are uniformly scaled; the color varies with the magnitude (or the current vector-to-scalar function, see key F)
    • show vector field as above, but the vectors are scaled proportionally to their magnitude
    • show the subset of the vector field with scalar function around a given value (adjusted with keys u / U and w / W)
    • show the vector field restricted to the boundary of the domain
  • V – Cycle the vector state in the opposite direction of v
  • u / U – Move the level field vectors (in the appropriate vector state)
  • w / W – Add/Delete level field vector (in the appropriate vector state)
  • d – Toggle the displaced mesh state (see also keys n / b). The options are:
    • do not show displaced mesh
    • show displaced mesh
  • n – Increase the displacement amount in 10% steps, wraps around from 100% to 0%
  • b – Decrease the displacement amount in 10% steps, wraps around from 0% to 100%
  • F – Change the vector-to-scalar function. The options are:
    • magnitude: \sqrt{v_x^2+v_y^2+v_z^2}
    • x-component: v_x
    • y-component: v_y
    • z-component: v_z

Auto-refinement

The GLVis auto-refinement algorithm selects a subdivision factor trying to achieve an accurate representation of high-order meshes and solution data while keeping the initial time to visualize the data reasonable. The algorithm can be summarized as follows:

  • GLVis draws surface elements; the number of drawn elements, ne, is either:
    • the number of elements in the mesh for 2D meshes (including surface meshes, i.e. 2D meshes embedded in 3D space), or
    • the number of boundary mesh elements described by the mesh in 3D.
  • A tentative upper limit on the number of vertices to be drawn is defined based on the maximum order of the mesh and the solution, max_order: 
    max_vert = ne * (max_order + 1) * (max_order + 1)
  • To allow more accurate representation for small meshes, this number is potentially increased: 
    max_vert = max(max_vert, 100 000)
  • To keep the time to initially visualize the data reasonable, this number is potentially reduced: 
    max_vert = min(max_vert, 2 000 000)
  • Finally, the subdivision factor ref is chosen to be the largest number such that:
    • the number of vertices needed to draw the ne surface elements with ref subdivisions does not exceed max_vert: 
      ne * (ref + 1) * (ref + 1) <= max_vert
    • for large meshes where the above limit cannot be satisfied, set ref = 1
    • for small meshes, avoid excessive refinements: 
      ref <= 16

Note that, for highly-varying data or large meshes, this auto-selected subdivision factor may not be sufficient for accurate representation. In such cases the subdivision can be manually adjusted using the keys o / O, described above.