Re: [Ns-commits] [Git][nsnam/ns-3-dev][master] 2 commits: ci/cd: Added Clang 10 compile test

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: [Ns-commits] [Git][nsnam/ns-3-dev][master] 2 commits: ci/cd: Added Clang 10 compile test

Barnes, Peter D.
Hello Natale,

Could you please commit on this commit:

On Jun 8, 2020, at 1:09 AM, N. <[hidden email]<mailto:[hidden email]>> wrote:
N. pushed to branch master at nsnam / ns-3-dev<https://gitlab.com/nsnam/ns-3-dev>
Commits:

  *   f8e8ade4<https://gitlab.com/nsnam/ns-3-dev/-/commit/f8e8ade4dfe0bd00b53bd3d8a7b82abc36e75d11>
by Natale Patriciello at 2020-06-08T10:09:32+02:00

ci/cd: Disabled debug/release tests, doing them only in optimized mode

Why not run tests in debug and release configurations?

We do allow code dependent on the build profile, via function-like macros NS_BUILD_x.  Such code in tests won’t get exercised after this commit.

P
_____________________________________________________________
Dr. Peter D. Barnes, Jr. CASC Division, B451 R2035
Lawrence Livermore National Laboratory Computation Directorate
7000 East Avenue, L-561 email:  [hidden email]<mailto:[hidden email]>
P. O. Box 808 Voice:  (925) 422-3384
Livermore, California 94550



Reply | Threaded
Open this post in threaded view
|

Re: [Ns-commits] [Git][nsnam/ns-3-dev][master] 2 commits: ci/cd: Added Clang 10 compile test

Natale Patriciello
Hello,

the problem is that testing requires time, and we are often blocked by
the timeout set on the worker machine (which is usually 3 hours).

With that change, I hope to avoid problems such as:

"There has been a timeout failure or the job got stuck. Check your timeout limits or try again"

(example taken from https://gitlab.com/nsnam/ns-3-dev/-/jobs/574473753) by running test
only in optimized mode. The timing problem has been quite present in the
last months (you can see it from the history of the pipelines that
almost every weekly schedule has failed).

Some time ago Tom was running a dedicated nsnam worked without the 3h
timelimit for the RUN_FOREVER tasks, but it would be a good thing if we
can contribute more power. When the code will stabilize, near the
release, I will try to do a run on my private branch to check timings again.

Does that sound reasonable? We can even try it now, by removing the
nsnam tag from the worker and see if everything stays under the 3h
limit.

Nat


On 08/06/20 at 06:02pm, Barnes, Peter D. wrote:

> Hello Natale,
>
> Could you please commit on this commit:
>
> On Jun 8, 2020, at 1:09 AM, N. <[hidden email]<mailto:[hidden email]>> wrote:
> N. pushed to branch master at nsnam / ns-3-dev<https://gitlab.com/nsnam/ns-3-dev>
> Commits:
>
>   *   f8e8ade4<https://gitlab.com/nsnam/ns-3-dev/-/commit/f8e8ade4dfe0bd00b53bd3d8a7b82abc36e75d11>
> by Natale Patriciello at 2020-06-08T10:09:32+02:00
>
> ci/cd: Disabled debug/release tests, doing them only in optimized mode
>
> Why not run tests in debug and release configurations?
>
> We do allow code dependent on the build profile, via function-like macros NS_BUILD_x.  Such code in tests won’t get exercised after this commit.
>
> P
> _____________________________________________________________
> Dr. Peter D. Barnes, Jr. CASC Division, B451 R2035
> Lawrence Livermore National Laboratory Computation Directorate
> 7000 East Avenue, L-561 email:  [hidden email]<mailto:[hidden email]>
> P. O. Box 808 Voice:  (925) 422-3384
> Livermore, California 94550
>
>
>
Reply | Threaded
Open this post in threaded view
|

Re: [Ns-commits] [Git][nsnam/ns-3-dev][master] 2 commits: ci/cd: Added Clang 10 compile test

Tom Henderson-2
On 6/8/20 11:22 AM, Natale Patriciello wrote:

> Hello,
>
> the problem is that testing requires time, and we are often blocked by
> the timeout set on the worker machine (which is usually 3 hours).
>
> With that change, I hope to avoid problems such as:
>
> "There has been a timeout failure or the job got stuck. Check your timeout limits or try again"
>
> (example taken from https://gitlab.com/nsnam/ns-3-dev/-/jobs/574473753) by running test
> only in optimized mode. The timing problem has been quite present in the
> last months (you can see it from the history of the pipelines that
> almost every weekly schedule has failed).
>
> Some time ago Tom was running a dedicated nsnam worked without the 3h
> timelimit for the RUN_FOREVER tasks, but it would be a good thing if we
> can contribute more power. When the code will stabilize, near the
> release, I will try to do a run on my private branch to check timings again.
>
> Does that sound reasonable? We can even try it now, by removing the
> nsnam tag from the worker and see if everything stays under the 3h
> limit.

I have a TODO (apologies again to Natale) to hook two dedicated slaves
to our gitlab CI:

- Linux-based

- a macOS machine

I also have a Jenkins installation that runs EXTENSIVE (bi-weekly) and
TAKES_FOREVER (monthly):

https://ns-buildmaster.ee.washington.edu:8010/

I have added Fedora-32 to this and will start to test clang-10 daily
once the current build issues settle.

- Tom

Reply | Threaded
Open this post in threaded view
|

Re: [Ns-commits] [Git][nsnam/ns-3-dev][master] 2 commits: ci/cd: Added Clang 10 compile test

Barnes, Peter D.
In reply to this post by Natale Patriciello
Do we have any timings on the individual build steps?

For example, generating the full doxygen is very time consuming.  For continuous testing, however, the doxygen.warnings.report.sh script is much more useful, and an order of magnitude faster.

Peter

On Jun 8, 2020, at 11:22 AM, Natale Patriciello <[hidden email]<mailto:[hidden email]>> wrote:

Hello,

the problem is that testing requires time, and we are often blocked by
the timeout set on the worker machine (which is usually 3 hours).

With that change, I hope to avoid problems such as:

"There has been a timeout failure or the job got stuck. Check your timeout limits or try again"

(example taken from https://gitlab.com/nsnam/ns-3-dev/-/jobs/574473753) by running test
only in optimized mode. The timing problem has been quite present in the
last months (you can see it from the history of the pipelines that
almost every weekly schedule has failed).

Some time ago Tom was running a dedicated nsnam worked without the 3h
timelimit for the RUN_FOREVER tasks, but it would be a good thing if we
can contribute more power. When the code will stabilize, near the
release, I will try to do a run on my private branch to check timings again.

Does that sound reasonable? We can even try it now, by removing the
nsnam tag from the worker and see if everything stays under the 3h
limit.

Nat


On 08/06/20 at 06:02pm, Barnes, Peter D. wrote:
Hello Natale,

Could you please commit on this commit:

On Jun 8, 2020, at 1:09 AM, N. <[hidden email]<mailto:[hidden email]><mailto:[hidden email]>> wrote:
N. pushed to branch master at nsnam / ns-3-dev<https://gitlab.com/nsnam/ns-3-dev>
Commits:

 *   f8e8ade4<https://gitlab.com/nsnam/ns-3-dev/-/commit/f8e8ade4dfe0bd00b53bd3d8a7b82abc36e75d11>
by Natale Patriciello at 2020-06-08T10:09:32+02:00

ci/cd: Disabled debug/release tests, doing them only in optimized mode

Why not run tests in debug and release configurations?

We do allow code dependent on the build profile, via function-like macros NS_BUILD_x.  Such code in tests won’t get exercised after this commit.

P
_____________________________________________________________
Dr. Peter D. Barnes, Jr. CASC Division, B451 R2035
Lawrence Livermore National Laboratory Computation Directorate
7000 East Avenue, L-561 email:  [hidden email]<mailto:[hidden email]><mailto:[hidden email]>
P. O. Box 808 Voice:  (925) 422-3384
Livermore, California 94550




_____________________________________________________________
Dr. Peter D. Barnes, Jr. CASC Division, B451 R2035
Lawrence Livermore National Laboratory Computation Directorate
7000 East Avenue, L-561 email:  [hidden email]<mailto:[hidden email]>
P. O. Box 808 Voice:  (925) 422-3384
Livermore, California 94550