Skip to content
GitLab
Explore
Sign in
Register
Primary navigation
Search or go to…
Project
F
fluiddyn_papers
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Wiki
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Container Registry
Model registry
Operate
Environments
Monitor
Incidents
Service Desk
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Terms and privacy
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
fluiddyn
fluiddyn_papers
Commits
9bebe5b1e3d7
Commit
9bebe5b1e3d7
authored
6 years ago
by
Ashwin Vishnu
Browse files
Options
Downloads
Patches
Plain Diff
FluidSim: comparison with sequential walltimes
parent
a3bcea4ef857
Loading
Loading
No related merge requests found
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
fluidsim/fluidsim_paper.tex
+30
-4
30 additions, 4 deletions
fluidsim/fluidsim_paper.tex
with
30 additions
and
4 deletions
fluidsim/fluidsim_paper.tex
+
30
−
4
View file @
9bebe5b1
...
...
@@ -687,10 +687,12 @@
\subsubsection*
{
Scalability
}
Scalability can be quantified by speedup
$
S
$
which is a measure of the time taken
to complete
$
N
$
iterations for different number of processes,
$
n
_
p
$
. We shall
refrain from comparing sequential runs in this context, since the operators used
for the sequential mode differ from the parallel mode, especially the FFT
class. Speedup is formally defined here as:
to complete
$
N
$
iterations for different number of processes,
$
n
_
p
$
.
%We shall
%refrain from comparing sequential runs in this context, since the operators used
%for the sequential mode differ from the parallel mode, especially the FFT
%class.
Speedup is formally defined here as:
\begin{equation}
S
_
\alpha
(n
_
p) =
\frac
...
...
@@ -743,6 +745,30 @@
\fluidpack
{
fft
}
, namely
\codeinline
{
fft2d.mpi
\_
with
\_
fftw1d
}
and
\codeinline
{
fft2d.mpi
\_
with
\_
fftwmpi2d
}
.
Although the FFT classes are different for sequential and parallel
runs, for the sake of comparison, we have tabulated the actual
walltime elapsed in Table~
\ref
{
table:seqpar
}
. Clearly due to the differences
in the implementations, for
\codeinline
{
fftw1d
}
-based classes it takes a
performance hit with parallelization, while for classes based on
\codeinline
{
fftw2d
}
the performance improves.
\begin{table}
\centering
\begin{tabular}
{
lrlr
}
\hline
FFT class
&
Time (
$
n
_
p
=
1
$
)
&
FFT class
&
Time (
$
n
_
p
=
2
$
)
\\
\hline
\codeinline
{
fft2d.with
\_
fftw1d
}
&
6.63
&
\codeinline
{
fft2d.mpi
\_
with
\_
fftw1d
}
&
7.63
\\
\codeinline
{
fft2d.with
\_
fftw2d
}
&
5.59
&
\codeinline
{
fft2d.mpi
\_
with
\_
fftwmpi2d
}
&
3.91
\\
\hline
\end{tabular}
\caption
{
Elapsed times (in seconds) for twenty time steps of 1024 x 1024 case
with the 2D Navier-Stokes 2D (
\codeinline
{
fluidsim.solvers.ns2d
}
) solver.
}
\label
{
table:seqpar
}
\end{table}
\begin{figure}
[htp]
\centering
\includegraphics
[width=\linewidth]
{
tmp/fig
_
bench
_
strong2d
}
...
...
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment