New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Array kernel #13528
Array kernel #13528
Conversation
@lindsayad This is just rebasing and make |
I also moved the description of array kernels in newsletter from June to July. |
Job Documentation on 0bc4aee wanted to post the following: View the site here This comment will be updated on new commits. |
https://mooseframework.inl.gov/docs/PRs/13528/site/source/variables/ArrayMooseVariable.html is the main link to the documentation. Unfortunately I cannot figure out where the added newsletter in MOOSE Docs is due to the issue I just created in #13529. Tag @friedmud |
The failure in Pronghorn demonstrated the dangerousness of using global params. All the failing tests in Pronghorn has a global parameter |
Neat!
…On Wed, Jun 19, 2019 at 2:42 AM Yaqi ***@***.***> wrote:
I did some tests and verified the speed-up of using array kernels with
Rattlesnake. Attached are the results with a simple setup with 1089 nodes:
[image: array-kernel]
<https://user-images.githubusercontent.com/6607496/59742270-863df580-922a-11e9-8a3c-bbb0da5742bf.jpg>
[image: array-kernel2]
<https://user-images.githubusercontent.com/6607496/59742281-8ccc6d00-922a-11e9-8465-d95911d6b896.jpg>
The first plot shows the grind times, defined as the total cpu time in
residual evaluations divided by number of evaluations and the number of
DoFs. The second plot shows the ratio of grind times. We clearly see the
benefit of using array kernels.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#13528?email_source=notifications&email_token=ACOGA4BLFS43KYE5T6CJMIDP3HIL7A5CNFSM4HWUEKO2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYA3OMY#issuecomment-503428915>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ACOGA4FKASDKEG2DA4TT7EDP3HIL7ANCNFSM4HWUEKOQ>
.
|
Added a test for array kernel save-in and did some optimizations revealed necessary for the rattlesnake test. |
Job Precheck on bfb7306 wanted to post the following: Your code requires style changes. A patch was auto generated and copied here
Alternatively, with your repository up to date and in the top level of your repository:
|
When can this be reviewed? |
Looks like a rebase is needed. If that's cleaned up, I will merge unless @friedmud still wants to review |
Thanks @lindsayad I do not know why I did not notice your reply! |
…tended to array variable idaholab#6881
…s with large number of components
@lindsayad I think we can merge this now. Then I can start fixing apps. |
The failing apps are: Pronghorn, Bighorn, R7, Moltres, etc. Unfortunately, I can only access Pronghorn and Moltres, will need to have somebody in MOOSE team for others. |
That was a convenient piece of information to leave out until this was
merged...
…On Wed, Jul 10, 2019 at 4:46 PM Yaqi ***@***.***> wrote:
The failing apps are: Pronghorn, Bighorn, R7, Moltres, etc. Unfortunately,
I can only access Pronghorn and Moltres, will need to have somebody in
MOOSE team for others.
—
You are receiving this because you modified the open/close state.
Reply to this email directly, view it on GitHub
<#13528?email_source=notifications&email_token=ACOGA4CARTXAQZJELIPUK53P6ZKFFA5CNFSM4HWUEKO2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZU2UJA#issuecomment-510241316>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ACOGA4EVMNZS5ROYBYINLWTP6ZKFFANCNFSM4HWUEKOQ>
.
|
The fix should be fairly quick if I can sit on the side. I believe the failures are mostly caused by the changes in |
No worries. The fixes have been easy
… On Jul 10, 2019, at 3:35 PM, Yaqi ***@***.***> wrote:
The fix should be fairly quick if I can sit on the side. I believe the failures are mostly caused by the changes in AddVariableAction. Sorry @lindsayad This has been here too long, I forgot something.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.
|
I am wrong, I can fix R7, thm. ;-) |
I am already working on thm and am about to submit the patch
…On Wed, Jul 10, 2019 at 11:38 PM Yaqi ***@***.***> wrote:
I am wrong, I can fix R7, thm. ;-)
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#13528?email_source=notifications&email_token=ACOGA4BXDK5JI4O6DSSYKALP63IOTA5CNFSM4HWUEKO2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZVVNLY#issuecomment-510351023>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ACOGA4GOT4BUOVTJ4QYKOL3P63IOTANCNFSM4HWUEKOQ>
.
|
Oh, I am . working on thm too, lol. I will let you do it then. |
Thanks man. Have a good sleep! |
Haha, you too!
…On Thu, Jul 11, 2019 at 12:00 AM Yaqi ***@***.***> wrote:
Oh, I am . working on thm too, lol. I will let you do it then.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#13528?email_source=notifications&email_token=ACOGA4DNLXBWWSW7S3DND5TP63K7FA5CNFSM4HWUEKO2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZVW2EY#issuecomment-510356755>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ACOGA4CTH7UWHPNI6TRHGTDP63K7FANCNFSM4HWUEKOQ>
.
|
/// Holds the solution at current quadrature points | ||
const ArrayVariableValue & _u; | ||
/// The component | ||
const unsigned int _component; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
this would have been better named as _components
or _num_components
and the corresponding input parameter as well. We uses component
to denote the index of a single component (I believe that is what caused the pronghorn failure - not the danger of global parameters, but an unfortunately named parameter in this PR).
Sounds like a fix-up PR would be good...?
…On Thu, Jul 11, 2019 at 7:46 AM Daniel Schwen ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In framework/include/auxkernels/ArrayVariableComponent.h
<#13528 (comment)>:
> +
+template <>
+InputParameters validParams<ArrayVariableComponent>();
+
+class ArrayVariableComponent : public AuxKernel
+{
+public:
+ ArrayVariableComponent(const InputParameters & parameters);
+
+protected:
+ virtual Real computeValue() override;
+
+ /// Holds the solution at current quadrature points
+ const ArrayVariableValue & _u;
+ /// The component
+ const unsigned int _component;
this would have been better named as _components or _num_componentsand
the corresponding input parameter as well. We uses component to denote
the index of a single component (I believe that is what caused the
pronghorn failure - not the danger of global parameters, but an
unfortunately named parameter in this PR).
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#13528?email_source=notifications&email_token=ACOGA4HYLZSHG4B6VB72N5DP65BV7A5CNFSM4HWUEKO2YY3PNVWWK3TUL52HS4DFWFIHK3DMKJSXC5LFON2FEZLWNFSXPKTDN5WW2ZLOORPWSZGOB6FLZUY#pullrequestreview-260750547>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ACOGA4DF7M4I4ITYSQN3CL3P65BV7ANCNFSM4HWUEKOQ>
.
|
Possibly. When I add this parameter, I kind of simplified the word to make it essential, like |
Close #6881. Superseding #13314 and #13459.