Thank you for pointing that out.
After passing my solution vector as reference the assembly time has reduced
from 2 seconds to 0.2 seconds.
Thanks and regards
Wasim Niyaz
Research scholar
CE Dept.
IITM
On Sun, 8 Jan, 2023, 7:31 pm Jean-Paul Pelteret,
wrote:
> Hi Wasim,
>
> It looks like
Hi Wasim,
It looks like you're passing your solution vector by copy for each cell
that you're assembling on. You probably want to pass it by reference.
Best,
Jean-Paul
Sent from my mobile device. Please excuse my brevity and any typos.
On Sun, 08 Jan 2023, 14:38 Wasim Niyaz Munshi ce21d400, <
Thank you Prof. Munch.
I am now passing FEValues object as a parameter to the H_plus function. My
assemble time has decreased from (4-5 seconds) to around 2 seconds, but it
is still nowhere close to the time required if there is no call to h_plus
function(say eg. step-3 tutorial problem).
I
Do I need to make this change in both declaration and definition?
Also, do I need a* const *because then* fe_values.reinit(cell) *will be
discarded?
Thanks
Wasim
On Sun, Jan 8, 2023 at 3:44 PM Peter Munch wrote:
> Change:
>
> > float H_plus(Vector solution_elastic, const auto cell,const
>
Change:
> float H_plus(Vector solution_elastic, const auto cell,const
unsigned int q_point,
* FEValues<2> fe_values_damage*);
to:
> float H_plus(Vector solution_elastic, const auto cell,const
unsigned int q_point,
* const FEValues<2> & fe_values_damage*);
Peter
On
Thank you, Prof. Munch.
I tried to pass the FEValues object as a parameter to the H_plus function
as follows:
I changed the function declaration from:
float H_plus(Vector solution_elastic, const auto cell,const
unsigned int q_point);
to this:
float H_plus(Vector solution_elastic, const auto
You are creating a new instance of FEValues at each quadrature point. This
is a very expensive operation, since there all the shape functions are
evaluated. Try to reuse that by passing it to the function as a parameter.
Hope that helps!
Peter
On Sunday, 8 January 2023 at 09:58:41 UTC+1
Following is my H_plus function:
float PhaseField::H_plus(Vector solution_elastic
, const auto cell,const unsigned int q_point)
{
QGauss<2> quadrature_formula_damage(fe_damage.degree + 1);
FEValues<2> fe_values_damage(fe_damage,
quadrature_formula_damage,
There might be many things that can be done to improve the speed of this
function. You can ask yourselve the following question as guidance:
- Does the function allocate memory?
- Could it be inlined?
- Are you calling the function inside the DOF loops or inside the
quadrature loop?
Then I
On 1/6/23 22:01, Wasim Niyaz Munshi ce21d400 wrote:
H_plus is being called in Assemble_damage and not assemble_elastic. It uses
elastic solution, cell and gauss point to evaluate strain at a gauss point.
Then some quantity is evaluated based on the strain.
Similarly I have another function
Sorry for the confusion. I think I made a mistake while writing the first
email.
H_plus is being called in Assemble_damage and not assemble_elastic. It uses
elastic solution, cell and gauss point to evaluate strain at a gauss point.
Then some quantity is evaluated based on the strain.
Similarly
I use it to evaluate strain at Gauss points. Then, i evaluate some quantity
which is a function of this strain.
Wasim Niyaz
Research scholar
CE Dept.
IITM
On Sat, 7 Jan, 2023, 3:09 am Wolfgang Bangerth,
wrote:
> On 1/6/23 13:53, Wasim Niyaz Munshi ce21d400 wrote:
> > I am using 65536 elements.
On 1/6/23 13:53, Wasim Niyaz Munshi ce21d400 wrote:
I am using 65536 elements. For step-8 the assembly takes very less time
(around 0.15second) while for my assemble_elastic, it takes around 5 seconds.
The only difference between my assemble_elastic function and the assemble
function of step-8
Hello everyone.
I am trying to solve elasticity and laplace equations one after the other.
The elasticity solution feeds as input to the laplace equation (stiffness
matrix of laplace equation depends on elasticity solution) and vice versa.
I am using 65536 elements. For step-8 the assembly takes
14 matches
Mail list logo
