Hi Zhe,

On 2/2/20 18:12, Wang, Zhe wrote:
Hi Herve,

I made a few modifications to scruff package trying to reduce computing time 
for the examples and vignette compilation. But still the development build does 
not pass on the Windows host due to timeout error. Can you check what went 
wrong on Windows OS?

What "went wrong" is that for some reason the code in the vignette is too slow on Windows. Note that 'R CMD build scruff' takes about 1273 seconds on malbec2. Given that things tend to run slower on Windows compared to Linux (for example, 'R CMD build Rsubread' takes 259 seconds on Windows vs 96 seconds only on Linux), it is not unreasonable to imagine that 'R CMD build scruff' would take more than 2400 seconds on tokay2. Hence the TIMEOUT.

Did you take a look at Rubread's performance regression introduced on Jan 7? Did you report it to the Rsubread maintainer? When you choose to depend on someone's else package, it means that you are prepared to deal with the changes they make to their package. This includes letting them know when they introduce problems that affect you.

Cheers,
H.


Thank you,
Zhe

-----Original Message-----
From: Wang, Zhe
Sent: Thursday, January 9, 2020 8:16 AM
To: Pages, Herve <hpa...@fredhutch.org>
Subject: RE: [Bioc-devel] TIMEOUT error in building vignette in package scruff

Hi Herve,

Thanks a lot for helping out. I'll take a look at Rsubread's source code and 
see if there's anything I can do to fix this. If not I guess I'll contact their 
maintainer for a solution.

Zhe

-----Original Message-----
From: Pages, Herve [mailto:hpa...@fredhutch.org]
Sent: Wednesday, January 8, 2020 7:07 PM
To: Wang, Zhe <z...@bu.edu>; bioc-devel@r-project.org
Subject: Re: [Bioc-devel] TIMEOUT error in building vignette in package scruff

I can confirm that downgrading Rsubread to an older version eliminates the 
slowdown. Here are the timings I get on my laptop (Ubuntu 16.04.6
LTS) for 'R CMD build scruff' with different versions of Rsubread (all these 
timings use the same R-devel version and the latest master branch of scruff):

               Rsubread           |
    commit       date     version | time R CMD build scruff
    ------------------------------+------------------------
    3eb25843  Jan  7 2020   2.1.1 |   49 min 20 sec
    6438d4a9  Dec 17 2019   2.1.1 | > 10 min (*)
    3f716691  Dec 16 2019   2.1.1 | > 10 min (*)
    b3602eb6  Dec 10 2019   2.1.1 | > 10 min (*)
    ef5e4c21  Dec  9 2019   2.1.1 |   45 min 33 sec
    d0b1acfb  Oct 29 2019   2.1.0 |    3 min 47 sec
    ------------------------------+------------------------
              (*) command was interrupted before completion

So the culprit seems to be commit ef5e4c21d:

    hpages@spectre:~/Rsubread$ git log ef5e4c21d -n 1
    commit ef5e4c21d5a6d633aec0b2922dd9f230fed23463
    Author: Yang Liao <dbpzd...@gmail.com>
    Date:   Mon Dec 9 13:42:08 2019 +1100

        Updated Ambient-RNA detection algorithm

Hope this helps,
H.


On 1/8/20 10:30, Pages, Herve wrote:
Hi Zhe,

The code in the vignette seems to rely a lot on the Rsubread package
to perform some very computationally intensive operations. Could it be
that some recent changes to the Rsubread package are somehow related
to the sudden slowdown of these operations? That's something I suggest
you investigate.

Cheers,
H.


On 1/8/20 07:50, Wang, Zhe wrote:
Hi all,

I am the maintainer of package scruff. Recently I encountered a timeout error 
in creating vignette in the development branch of the package. I tried 
increasing the number of cores for parallelization from 2 to 3 but it still 
timed out. I did not make any changes to the R codes in the vignette before 
this error happened. I tested building the vignette locally and it compiles 
without error. Does anyone know how I should fix this? Should I comment out 
some of the command in the vignette to reduce computing time? Should I try 
increasing the core number a bit more?

Thanks,
Zhe


        [[alternative HTML version deleted]]

_______________________________________________
Bioc-devel@r-project.org mailing list
https://urldefense.proofpoint.com/v2/url?u=https-3A__stat.ethz.ch_mai
lman_listinfo_bioc-2Ddevel&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3X
eAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=stHSi8EN5PUgMEVus6KdL2z7-DfGz
ikck8btNROFUx0&s=dSXLn6y8wJtjg6ukcpOpGbSHfXqqsH97r9fuSJkakGo&e=



--
Hervé Pagès

Program in Computational Biology
Division of Public Health Sciences
Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N, M1-B514
P.O. Box 19024
Seattle, WA 98109-1024

E-mail: hpa...@fredhutch.org
Phone:  (206) 667-5791
Fax:    (206) 667-1319


--
Hervé Pagès

Program in Computational Biology
Division of Public Health Sciences
Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N, M1-B514
P.O. Box 19024
Seattle, WA 98109-1024

E-mail: hpa...@fredhutch.org
Phone:  (206) 667-5791
Fax:    (206) 667-1319

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

Reply via email to