These are chat archives for elemental/chat

13th
Feb 2017
Aidan Dang
@AidanGG
Feb 13 2017 15:58
Hi all, I've got an example of a matrix that's giving me problems, as a .bin file from El::Write. It's about 63MB. What's the easiest way for me to send it?
Jack Poulson
@poulson
Feb 13 2017 16:02
is there a website you can post it to and link to?
Aidan Dang
@AidanGG
Feb 13 2017 16:03
I'll chuck it on my google drive
Jack Poulson
@poulson
Feb 13 2017 16:03
is there any context for this matrix?
Aidan Dang
@AidanGG
Feb 13 2017 16:04
It was generated as part of a tensor network calculation
Jack Poulson
@poulson
Feb 13 2017 16:04
was it an intermediate step in an alternating algorithm?
Aidan Dang
@AidanGG
Feb 13 2017 16:05
I'm using the svd to cut a matrix into two, then using the svd to cut one of those, and so on
Jack Poulson
@poulson
Feb 13 2017 16:06
OK
Aidan Dang
@AidanGG
Feb 13 2017 16:06
I'm fairly certain the matrix I have doesn't have any undefineds/NaNs
Jack Poulson
@poulson
Feb 13 2017 16:07
it would be good to see if Octave can safely compute the SVD of that matrix as well
my guess is that it would take about 10 minutes
Aidan Dang
@AidanGG
Feb 13 2017 16:07
I think scalapack does
Jack Poulson
@poulson
Feb 13 2017 16:08
assuming roughly a 3 thousand by 3 thousand matrix
ScaLAPACK uses the QR algorithm
have you tested Elemental's QR algorithm on it?
Aidan Dang
@AidanGG
Feb 13 2017 16:09
is that through bidiagSVDctrl.useQR?
Jack Poulson
@poulson
Feb 13 2017 16:09
yes, that is the best way
Aidan Dang
@AidanGG
Feb 13 2017 16:10
Ok, it still doesn't work
Jack Poulson
@poulson
Feb 13 2017 16:10
what is the error/exception?
Aidan Dang
@AidanGG
Feb 13 2017 16:10
Did not converge all singular values
Jack Poulson
@poulson
Feb 13 2017 16:11
one can increase the maximum number of iterations to see if it was almost converging
Aidan Dang
@AidanGG
Feb 13 2017 16:12
Is that qrCtrl.maxIterPerVal?
Jack Poulson
@poulson
Feb 13 2017 16:13
bidiagSVDCtrl.qrCtrl.maxIterPerVal = 20 could be tried
yes
Also, did you verify that ScaLAPACK returned with INFO=0?
Aidan Dang
@AidanGG
Feb 13 2017 16:15
Oh, I haven't, but the final results were sensible, so I assumed it was ok
Jack Poulson
@poulson
Feb 13 2017 16:17
I would recommend checking for the sake of consistency, as it could very well be that there is one small troublesome singular value
Jack Poulson
@poulson
Feb 13 2017 16:17
you can also tell Elemental not to worry if the singular values don't all converge
Aidan Dang
@AidanGG
Feb 13 2017 16:17
those are my distmatrix.bin and test script
I've tried setting qrctrl.demandconverged to false, but I don't think that worked either
Jack Poulson
@poulson
Feb 13 2017 16:18
it still complained about not converging?
I haven't tested that option explicitly but tried to set it up
I'm downloading it now
I have about ten minutes to make/drink coffee, figure this out, and run to work but can hopefully find something
Aidan Dang
@AidanGG
Feb 13 2017 16:21
Please don't stress too much about it, you've been a great help.
If I set demandconverged to false, it appears to hang
Jack Poulson
@poulson
Feb 13 2017 16:25
something seems to be wrong with that file:
poulson@poulson-ASUS:~/Source/Internal/Elemental/build-gcc$ ./bin/sandbox-test 
terminate called after throwing an instance of 'std::runtime_error'
  what():  Expected file to be 8 bytes but found 63406096

Aborted (core dumped)
how did you create it?
Aidan Dang
@AidanGG
Feb 13 2017 16:25
El::Write
I'm on 64 bit ints
Jack Poulson
@poulson
Feb 13 2017 16:27
ah, I wonder if that's it
also, have you checked that if you download that .bin file from your Google Drive that you can run your test program with it?
Aidan Dang
@AidanGG
Feb 13 2017 16:28
I'll check it out.
Jack Poulson
@poulson
Feb 13 2017 16:28
the metadata at the beginning of the file is an El::Int, and that size is variable
I think that is the problem with me trying to run with the 32-bit version
Aidan Dang
@AidanGG
Feb 13 2017 16:29
I can get you a version with 32bit ints
after I get some sleep, because it's 3:30am here in Melbourne. Thanks so much for your support.
Jack Poulson
@poulson
Feb 13 2017 16:31
no worries, I can build a 64-bit version and debug
get some sleep and I will fix it by tonight