Re: [Bioc-devel] benchdamic TIMEOUT in merida1 and kunpeng2

[Matteo Calgaro]

Dear Martin and Levi,
I truly appreciate your tips and the time you took to assist me. I'm going
to look further into your suggestions by trying to reduce the computational
time needed to build the vignette. I'll let you know if I manage to fix the
problem.

Matteo

Il giorno lun 17 lug 2023 alle ore 11:45 Levi Waldron <
lwaldron.resea...@gmail.com> ha scritto:

> It looks to me like benchdamic doesn't actually use the HMP16SData
> package, but a repackaged example dataset placed in the data/ directory. I
> do notice that the vignette takes very long to run, especially the chunks
> below. My guess is that you are seeing differences in speed or number of
> available cores on the build machines, one of which can't build the
> vignette within the 10-minute time limit (e.g. it takes more than 10
> minutes to build on my laptop). Can you reduce the number of rows or
> columns in your example dataset, or reduce permutations, to speed up the
> computation while still demonstrating how the package should be used?
>
> ```{r runMocks}
> bpparam <- BiocParallel::SerialParam()
> # Random grouping each time
> Stool_16S_mockDA <- runMocks(
> mocks = my_mocks,
> method_list = my_methods,
> object = ps_stool_16S,
> weights = zinbweights,
> verbose = FALSE,
> BPPARAM = bpparam)
> ```
>
> ```{r runSplits}
> # Set the parallel framework
> # Remember that ANCOMBC based methods are compatible only with
> SerialParam()
> bpparam <- BiocParallel::SerialParam()
>
> # Make sure the subject ID variable is a factor
> phyloseq::sample_data(ps_plaque_16S)[, "RSID"] <- as.factor(
> phyloseq::sample_data(ps_plaque_16S)[["RSID"]])
>
> Plaque_16S_splitsDA <- runSplits(
> split_list = my_splits,
> method_list = my_methods_noWeights,
> normalization_list = my_normalizations,
> object = ps_plaque_16S,
> min_counts = 0, min_samples = 2,
> verbose = TRUE,
> BPPARAM = bpparam)
> ```
>
> (returns lots of "Warning: Computed variance-covariance matrix problem:
> not a positive definite matrix (and positive semidefiniteness is not
> checked); returning NA matrixboundary (singular) fit: see
> help('isSingular')" warnings)
>
>
> --
>
> Levi Waldron
>
> Associate Professor
>
> Department of Epidemiology and Biostatistics
>
> CUNY Graduate School of Public Health and Health Policy
>
> Institute for Implementation Science in Population Health
>
> 55 W 125th St, New York NY 10035
>
> https://waldronlab.io
>
> Book a meeting with me
> 
>
> Join the microbiome Virtual International Forum:
> https://microbiome-vif.org
>

[[alternative HTML version deleted]]

___
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel

Re: [Bioc-devel] benchdamic TIMEOUT in merida1 and kunpeng2

[Levi Waldron]

It looks to me like benchdamic doesn't actually use the HMP16SData package,
but a repackaged example dataset placed in the data/ directory. I do notice
that the vignette takes very long to run, especially the chunks below. My
guess is that you are seeing differences in speed or number of available
cores on the build machines, one of which can't build the vignette within
the 10-minute time limit (e.g. it takes more than 10 minutes to build on my
laptop). Can you reduce the number of rows or columns in your example
dataset, or reduce permutations, to speed up the computation while still
demonstrating how the package should be used?

```{r runMocks}
bpparam <- BiocParallel::SerialParam()
# Random grouping each time
Stool_16S_mockDA <- runMocks(
mocks = my_mocks,
method_list = my_methods,
object = ps_stool_16S,
weights = zinbweights,
verbose = FALSE,
BPPARAM = bpparam)
```

```{r runSplits}
# Set the parallel framework
# Remember that ANCOMBC based methods are compatible only with SerialParam()
bpparam <- BiocParallel::SerialParam()

# Make sure the subject ID variable is a factor
phyloseq::sample_data(ps_plaque_16S)[, "RSID"] <- as.factor(
phyloseq::sample_data(ps_plaque_16S)[["RSID"]])

Plaque_16S_splitsDA <- runSplits(
split_list = my_splits,
method_list = my_methods_noWeights,
normalization_list = my_normalizations,
object = ps_plaque_16S,
min_counts = 0, min_samples = 2,
verbose = TRUE,
BPPARAM = bpparam)
```

(returns lots of "Warning: Computed variance-covariance matrix problem: not
a positive definite matrix (and positive semidefiniteness is not
checked); returning
NA matrixboundary (singular) fit: see help('isSingular')" warnings)


-- 

Levi Waldron

Associate Professor

Department of Epidemiology and Biostatistics

CUNY Graduate School of Public Health and Health Policy

Institute for Implementation Science in Population Health

55 W 125th St, New York NY 10035

https://waldronlab.io

Book a meeting with me


Join the microbiome Virtual International Forum: https://microbiome-vif.org

[[alternative HTML version deleted]]

___
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel

Re: [Bioc-devel] benchdamic TIMEOUT in merida1 and kunpeng2

[Martin Grigorov]

HI,

I don't have a good explanation why it was timing out on kunpeng2 but I
think it will pass from now on!

Here is what I did:

1) I tried to build benchdamic on kunpeng2:

$ R CMD build --keep-empty-dirs --no-resave-data benchdamic
* checking for file ‘benchdamic/DESCRIPTION’ ... OK
* preparing ‘benchdamic’:
* checking DESCRIPTION meta-information ... OK
* installing the package to build vignettes
* creating vignettes ...

and it hung here for several minutes...

2) I checked the single vignette in the package and I noticed that it
uses/depends on data package HMP16SData

3) in a new terminal session I started R and installed the data package:

BiocManager::install("HMP16SData")
Bioconductor version 3.18 (BiocManager 1.30.21), R 4.3.1 (2023-06-16)
Installing package(s) 'HMP16SData'
trying URL '
https://bioconductor.org/packages/3.18/data/experiment/src/contrib/HMP16SData_1.21.0.tar.gz
'
Content type 'application/x-gzip' length 17066704 bytes (16.3 MB)
==
downloaded 16.3 MB

* installing *source* package ‘HMP16SData’ ...
** using staged installation
** R
** data
*** moving datasets to lazyload DB
** inst
** byte-compile and prepare package for lazy loading
** help
*** installing help indices
** building package indices
** installing vignettes
** testing if installed package can be loaded from temporary location
** testing if installed package can be loaded from final location
** testing if installed package keeps a record of temporary installation
path
* DONE (HMP16SData)

Since it didn't stop early with a message like "The package is already
installed with the last version" it seems this data package was not
installed until now. Should this data package be listed as
dependency/suggest in the DESCRIPTION file ?1

4) and now the build in the first terminal session finished successfully:
...
* checking for LF line-endings in source and make files and shell scripts
* checking for empty or unneeded directories
* building ‘benchdamic_1.7.3.tar.gz’

Since the data package is installed I think from now on
the benchdamic package will not timeout on kunpeng2 until the next update
of the R installation/version when all installed packages in site-library
will be lost and they will have to be downloaded again (automatically or
manually, as I did for HMP16SData)

Regards,
Martin

On Sun, Jul 16, 2023 at 1:19 PM Matteo Calgaro  wrote:

> Dear all,
> my package benchdamic has been giving TIMEOUT in the build results on
> merida1
> 
> (release macOS) and kunpeng2
> 
> (devel Linux).
>
> The devel and release versions of benchdamic are equal due to some updates
> I had to do in both versions to avoid errors. Interestingly, the package
> builds well in the macOS platform in devel (Iconway) while it gives TIMEOUT
> in release. I held off reporting the issue until I had implemented the new
> features, even though the TIMEOUT state has been present since the last
> Bioconductor release.
> The build report only seems to indicate that it times out while building
> vignettes. How should I solve this?
>
> Thanks a lot in advance.
> Best regards.
> Matteo Calgaro
>
> [[alternative HTML version deleted]]
>
> ___
> Bioc-devel@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>

[[alternative HTML version deleted]]

___
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel

[Bioc-devel] benchdamic TIMEOUT in merida1 and kunpeng2

[Matteo Calgaro]

Dear all,
my package benchdamic has been giving TIMEOUT in the build results on
merida1

(release macOS) and kunpeng2

(devel Linux).

The devel and release versions of benchdamic are equal due to some updates
I had to do in both versions to avoid errors. Interestingly, the package
builds well in the macOS platform in devel (Iconway) while it gives TIMEOUT
in release. I held off reporting the issue until I had implemented the new
features, even though the TIMEOUT state has been present since the last
Bioconductor release.
The build report only seems to indicate that it times out while building
vignettes. How should I solve this?

Thanks a lot in advance.
Best regards.
Matteo Calgaro

[[alternative HTML version deleted]]

___
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel