Hi all,
(I've also tried to send this via NNTP to the mailing list. It may well
not work -- I'm still figuring things out.)
Karl Rupp r...@iue.tuwien.ac.at writes:
What's the common way of defining classes in the Python ecosystem?
Upper-case?
The official Style Guide[1] says that class names
Karl Rupp r...@iue.tuwien.ac.at writes:
Okay, so this shows ~150 us overhead for representing the types as in
pyviennacl. Compared with a pure CPU-implementation, this is certainly a
lot.
However, when comparing this with OpenCL, then the overhead is actually
not too bad and becomes
Hi Philippe,
Philippe Tillet phil.til...@gmail.com writes:
Well, it really depends on applications. 2**15 is actually still fairly
small, since it is not exactly big enough to be bandwidth-limited (as
opposed to kernel-launch overhead limited). I'd say considering the low
bandwidth of PCI-E
Hi everyone,
In the process of adding scheduler-matrix support to the wrapper, I got
a bit stuck on a bug that I caused by not initialising my not
initialising the matrix to which I was trying to assign a result. I
think it's sensible to throw a different exception from not
implemented in these
Toby St Clere Smithe m...@tsmithe.net
writes:
In the process of adding scheduler-matrix support to the wrapper, I got
a bit stuck on a bug that I caused by not initialising my not
initialising the matrix to which I was trying to assign a result. I
think it's sensible to throw a different
Karl Rupp r...@iue.tuwien.ac.at writes:
yes, I totally agree that we should have different types of exceptions.
Please commit it yourself directly, you should have push permissions. :-)
Right-ho, will do :) (I also managed to coerce Boost into handling char*
exceptions -- the trick was
Hi,
Last thing from me for the night, but I've just swapped to using async_
then fast_copy instead of plain old copy for vectors -- and both give me
a segfault using nvidia's libopencl. Is there anything debuggable in
this case, since the damn thing is closed source? The backtrace is:
#0
Hi Karl,
Karl Rupp r...@iue.tuwien.ac.at writes:
I consider Vector * Vector - Matrix to be surprising or at least
somewhat non-intuitive. Following your operations, the most reasonable
definition would be
Matrix * Matrix - Matrix (matrix product)
Vector * Vector - Vector
Hi all,
Karl Rupp r...@iue.tuwien.ac.at writes:
As I'm presumably unavailable on Friday and don't know about my
availability over the weekend, what about tomorrow, Thursday, at 19:00
UTC? Does this work for at least Toby and Phil? Feel free to propose
other time slots for tomorrow, I
Hi Karl,
Karl Rupp r...@iue.tuwien.ac.at writes:
Parallella will ship their first (OpenCL-enabled!) boards in October and
also offers a university partner program:
http://www.parallella.org/pup/
I'd forgotten about these things! Their roadmap does look intruiging;
I'm looking forward to
Karl Rupp r...@iue.tuwien.ac.at writes:
Even though high core counts look promising in terms of FLOPs/Watt, the
tricky part is to actually use them efficiently... :-)
I don't think we will see 64K cores by that time, though. Their inital
roadmap listed 1k cores for 2014 already, which is
Hello!
Alex Eftimiades alexeftimia...@gmail.com
writes:
I have been looking for an open source linear algebra library that
will work on a GPU for a while now, and it seems ViennaCl is one of
the few out there. It looks really good, but unfortunately for me, it
is a C++ library. I am
Hi all,
Just a quick update. I'm in the process of writing decent documentation
for my classes and functions. There are quite a few: sloccount tells me
that, including tests, I've got about 5500 source lines of code (not
including comments or docs), that it's worth $160,000, and that it
should
Hi Philippe,
Philippe Tillet phil.til...@gmail.com
writes:
Concerning version numbering, I would tend to think that PyViennaCL should
indeed be rebuilt/repackaged at least every minor release, so that it can
benefit performance improvements in the core. On the other hand, following
the same
Karl Rupp r...@iue.tuwien.ac.at writes:
Very nice! :-)
Thanks! :)
Are you sure about the lack of precision in NumPy? Do you get relative
errors in the 10^{-6} range? Please let me know if there are any issues
with slice and range (or any other operations) in whatever scenarios.
Well, for
Karl Rupp r...@iue.tuwien.ac.at writes:
copy() is available for stability, fast_copy() for performance. I don't
think we should provide a fast_copy() which silently throws away
performance just for the sake of stability. I expect that a user wants
to get some sort of error if the
Karl Rupp r...@iue.tuwien.ac.at writes:
That's actually the case already (at least mostly): copy() packs the
data in one or more suitably sized arrays and then calls fast_copy() :-)
Well, that shows me not to respond without looking at the code in
question!
Ciao,
Toby
Karl Rupp r...@iue.tuwien.ac.at writes:
This is now fixed. You should be able to instantiate the basic types
with char, uchar, short, ushort as well. So far I have explicitly tested
this only with vector, though.
Excellent. I'll try and get this enabled in the next couple of days, and
then
Hi all,
I don't have much to report, but I thought I should check in to show
that I'm still alive, since there's been a flurry of git activity
recently. I'm slowly getting on top of my workload, and should finally
be there around the beginning of December. I'll set aside a day then to
get
Hi all,
I've made a fair amount of headway getting a binary package built for
PyViennaCL on Ubuntu (and thus Debian). If anyone is using an i386
system (not amd64 yet!) on Ubuntu trusty (14.04), they can test it out,
using the PPA at [1].
But herein lies the problem. I've used the build tools
Toby St Clere Smithe m...@tsmithe.net
writes:
Thanks; I think 2 probably is most sensible, and you're right about the
default version point. I've managed (by turning off a couple of compiler
debug flags and optimisations that are probably unnecessary in a first
release) to get the amd64 peak
Hi Karl,
Karl Rupp r...@iue.tuwien.ac.at writes:
Nice, ~1GB should be fine, as this is also not that unusual with other
projects. I intend to refactor some of the ViennaCL internals in order
to reduce the memory consumption after 1.5.0 is out, which should also
reduce the burden on the
Toby St Clere Smithe m...@tsmithe.net
writes:
Yep, looks like the build was successful, so I'll go ahead and make sure
it's all working on older distributions now.
And so, after a little over a day backporting the package build system
from Python 3 to Python 2 (yay!), we have packages
Toby St Clere Smithe m...@tsmithe.net
writes:
So having split up my sources, I'm having trouble with shared
pointers. It seems that because of the way the sources have been split
into a number of functions, some pointer goes out of scope and gets
dereferenced twice.. I'm not sure the problem
Hey Philippe,
Philippe Tillet phil.til...@gmail.com
writes:
I wrote the shared_ptr some years ago (inspired by some code found on the
internet), so it's not impossible that a couple of bugs sneaked in. The
first thing to do is to use C++11 and to replace the whole content of
Hi all,
So I've uploaded PyViennaCL packages which don't have the shared_ptr
troubles that my previous ones did (in the end, I split what used to be
a single Python extension into many smaller extensions under one package
namespace). These packages seem mostly to work, and (for instance), pass
my
Hi Karl,
Karl Rupp r...@iue.tuwien.ac.at writes:
So I've uploaded PyViennaCL packages which don't have the shared_ptr
troubles that my previous ones did (in the end, I split what used to be
a single Python extension into many smaller extensions under one package
namespace). These packages
Toby St Clere Smithe m...@tsmithe.net
writes:
I'll dig around some more...
So it turned out that it wasn't anything to do with my previous error,
just that having split up the files, I had not put the OpenCL #define in
all the right places.
In fact, come to think of it, that may be the cause
Karl Rupp r...@iue.tuwien.ac.at writes:
Ah, I see. What about setting the defines globally through the CMake
build system? That should be far less error-prone and easier to configure...
Quite...! :)
--
Rapidly
Dear Karl,
Karl Rupp r...@iue.tuwien.ac.at writes:
* Installation: Works nicely from the PPA, no problems with my Linux
Mint Maya (based on Ubuntu 12.04, so this is expected).
Good! :)
* Found the PyViennaCL documentation, but it was *not* installed
automatically, but listed as
Hi Karl,
Karl Rupp r...@iue.tuwien.ac.at writes:
the Google Summer of Code [1] is approaching. It attracted some great
contributors in the past, most notably Philippe and Toby, and I hope
there's more to come. So, guys, please provide your project ideas.
I have a couple of inter-related
Toby St Clere Smithe m...@tsmithe.net writes:
I'm definitely aiming for a release in the next couple of weeks. This
weekend is my birthday, so I'm probably looking at some time around the
next weekend (14th-17th ish).
My list of features to complete in PyViennaCL (as separate from ViennaCL
Just another brief thought before I go to bed -- Karl, I'll respond to
your response in the morning ! :)
Karl Rupp r...@iue.tuwien.ac.at writes:
The first project I have in mind is the benchmarking GUI we brainstormed
about in IRC. It's probably a good idea to push out a first working
Hey all,
So I've fixed all the below stuff (but I've not yet uploaded a new
package to the PPA). I just want a small piece of advice (below).
Karl Rupp r...@iue.tuwien.ac.at writes:
* Okay, so let's try to print individual elements of A and B:
A(1,1);# error!
A.item(1,1); #
Hey, Philippe,
Philippe Tillet phil.til...@gmail.com
writes:
I was thinking about another nice-to-have feature. I'd quite like to
make it possible (if it is in fact possible..) to play with prototyping
implementations of other algorithms for ViennaCL in PyViennaCl using
PyOpenCL and PyCUDA;
Hi Karl,
Karl Rupp r...@iue.tuwien.ac.at writes:
I'm on it. A couple of notes on the setup:
* The manual checkouts of external/boost_numpy and
external/viennacl-dev don't work for me (git 1.7.9.5). Is this supposed
to be automatic? If not, there should be instructions in the README.
Hmm,
, though...
Do you have a link to a diff I can read? Also, I'm still just using
ViennaCL 1.5.1: did the change happen there? In any case, I might need
to change things later..
Cheers,
Toby
Philippe
2014-02-18 17:29 GMT+00:00 Toby St Clere Smithe m...@tsmithe.net:
Hey,
Karl Rupp r
Hi all,
The time has finally (almost) come. I'd like to make a 1.0.0 release
tomorrow (or rather, later today), but first of all, there are three
things I'd like to happen.
Firstly, I need to move pyviennacl into its own repository under
viennacl-dev on GitHub. I don't know how to do that, but
Hi Karl,
Karl Rupp r...@iue.tuwien.ac.at writes:
I just created https://github.com/viennacl/pyviennacl-dev and added you
as collaborator. Simply push the relevant code using e.g.
git remote add origin URL
git push -u origin branchname
You might want to either strip out old ViennaCL
Hi,
Karl Rupp r...@iue.tuwien.ac.at writes:
For now I have to generate the HTML-page myself, unfortunately it
doesn't yet scale beyond the machines of the institute. Let me know if
there is any special content to be included, otherwise I'll just extract
the highlights from the Sphinx
Toby St Clere Smithe m...@tsmithe.net
writes:
OK, so you can use 'tsmithe-sf' as my account name. I thought I had
'tsmithe', but the account recovery system seems to be broken and I can
only remember my username and registered e-mail address..
The recovery e-mail came through eventually, so
Hi Karl,
Karl Rupp r...@iue.tuwien.ac.at writes:
alright, I added you to the developer group. I don't know whether this
is sufficient for write permissions at the project web, but just give it
a try. User permission management is quite coarse-grained on sf.net.
Great! See
Hi Evan,
Evan Bollig bol...@gmail.com writes:
Toby, another suggestion for you: publish an ipython notebook to help
users step through a tutorial.
You could also post it to http://nbviewer.ipython.org/
I've got the example / tutorial code published at [1]. I should think
about automatically
Hey,
Karl Rupp r...@iue.tuwien.ac.at writes:
http://viennacl.sourceforge.net/pyviennacl.html
Looks good! Evan's suggestion makes me think that the example code
should be made really obvious. Could you put another link there like
[PyViennaCL Examples] pointing at
Hi again,
Toby St Clere Smithe m...@tsmithe.net
writes:
Karl Rupp r...@iue.tuwien.ac.at writes:
http://viennacl.sourceforge.net/pyviennacl.html
Looks good! Evan's suggestion makes me think that the example code
should be made really obvious. Could you put another link there like
[PyViennaCL
Another late night e-mail from tsmithe:
Karl Rupp r...@iue.tuwien.ac.at writes:
Hey,
Looks good! Evan's suggestion makes me think that the example code
should be made really obvious. Could you put another link there like
[PyViennaCL Examples] pointing at
Hi all,
I don't have much expertise with MSVC, but I'm trying to build
pyviennacl, and I've got a couple of weird bugs. I've looked at the
source, but really have no idea what it's complaining about. Why can't
it resolve the type ambiguity here?
OK, so after re-doing the build system, let's try again. Unfortunately,
the Windows terminal window didn't save very much history, so I only
have this much of the error output. I think it's enough to be
diagnostic; if not, I'll run it again, redirecting the output to a
file..
Once I've got this
Hi Karl,
Karl Rupp r...@iue.tuwien.ac.at writes:
alright, this looks like the issue is with dense matrices being passed
to BiCGStab. Does the build work if you disable the dense matrices for
the iterative solvers? If so, then I think you can temporarily fix this
within PyViennaCL and we
Hi all,
So with the new build system it is now possible to build PyViennaCL on
Windows. I only have Windows in a virtual machine, so it would be good
to know if it works on hardware with a proper OpenCL implementation. If
any of you are interested, read on!
You'd need to grab the sources using
Hi,
Andreas Kloeckner
li...@informa.tiker.net writes:
Karl Rupp r...@iue.tuwien.ac.at writes:
alright, some progress here: Installed Python, the Windows SDK, and
NumPy as described. Then things got complicated:
- Had to install setuptools. This should be mentioned in the README.
-
Hi,
Karl Rupp r...@iue.tuwien.ac.at writes:
Right -- so I've assumed that you have a working python and git
installation; it seemed fair to assume that, if you want to install from
source and want to be using Python, you'd have both set up.
I don't think we should assume that something is
Hi Karli,
Karl Rupp r...@iue.tuwien.ac.at writes:
One thing to watch out for here is ownership semantics. If ViennaCL
insists on owning storage, then you may be forced to copy to preserve
correctness.
This is something I will also have to be careful about -- thanks for
reminding me
Hey,
Karl Rupp r...@iue.tuwien.ac.at writes:
Yes, that worked, see my comment here:
https://github.com/viennacl/pyviennacl-dev/issues/2
Installation now succeeded. :-)
Did you change anything else since then?
Nope! I assume all is well, then :)
Cheers,
Toby
matrix_slicematrix_base , and
leave taking slices of slices unsupported. I'm slightly uncomfortable
about this, not least because it leaves a regression in my test suite,
but it'll enable me to continue with the more urgent work.
Any ideas?
Best,
Toby
--
Toby St Clere Smithe
http://tsmithe.net
Toby St Clere Smithe m...@tsmithe.net
writes:
I reiterate my call for an IRC meeting, but, following Namik's lead,
I'll post a brief status update about PyViennaCL 2014.
brief...
--
Open source business process
Karl Rupp r...@iue.tuwien.ac.at writes:
fine with me. My schedule is very much in flux in the next ~10 days or
so, so I might be unavailable on short notice. Rather than having one
big IRC meeting with all topics crushed together, I suggest we have a
couple of smaller topic-oriented
-devel
--
Toby St Clere Smithe
http://tsmithe.net
--
Open source business process management suite built on Java and Eclipse
Turn processes into business applications with Bonita BPM Community Edition
Quickly connect people
Hi Philippe,
Philippe Tillet phil.til...@gmail.com
writes:
I've already told this on IRC. The GEMV uses very conservative profiles
with very few threads. Now that I have ported a simple version of GEMM
(when only full matrices are used), I'll re-bind the generator into
pyviennacl and will try
Hi all,
I'm rewriting my test suite, and am currently getting exceptions from
the scheduler when trying to do something like this pseudo-code:
A = matrix
B = unary_sqrt(A)
C = A + B
The relevant exception is line 252 of scheduler/execute_axbx.hpp (I'm at
commit ca0c4d), and my statement
place.
Philippe
2014-07-24 18:28 GMT+02:00 Toby St Clere Smithe m...@tsmithe.net:
Hi all,
I'm rewriting my test suite, and am currently getting exceptions from
the scheduler when trying to do something like this pseudo-code:
A = matrix
B = unary_sqrt(A)
C = A + B
The relevant
,
Toby
--
Toby St Clere Smithe
http://tsmithe.net
--
Infragistics Professional
Build stunning WinForms apps today!
Reboot your WinForms applications with our WinForms controls.
Build a bridge from your legacy apps
will be pushed (hopefully) tonight.
OK, I've done that. I wrote such a routine in Python; it's only 25
(verbose) lines.
Cheers,
Toby
--
Toby St Clere Smithe
http://tsmithe.net
--
Infragistics Professional
Build stunning
find
one for the iterative solvers!
Cheers,
Toby
--
Toby St Clere Smithe
http://tsmithe.net
--
Infragistics Professional
Build stunning WinForms apps today!
Reboot your WinForms applications with our WinForms controls
()) (conf.iters_ !=
conf.check_after_steps() + 1))
break;
// Stagnation check
Cheers,
Toby
--
Toby St Clere Smithe
http://tsmithe.net
--
Infragistics Professional
Build stunning WinForms apps
’ is not a member of ‘viennacl::linalg’
viennacl::linalg::pipelined_bicgstab_vector_update(result, alpha, p,
omega, s,
^
Cheers,
Toby
--
Toby St Clere Smithe
http://tsmithe.net
--
Want fast and easy
!
Cheers,
Toby
--
Toby St Clere Smithe
http://tsmithe.net
--
___
ViennaCL-devel mailing list
ViennaCL-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo
Hi all,
Before the release happens, I thought I should run the PyViennaCL
tests. I'm getting a segfault coming out of OpenCL (both nVidia's and
Beignet) when running the direct solvers:
test_direct_solvers.py:38: test_matrix_trans_lower_A_matrix_B_C_float32
Program received signal SIGSEGV,
Corrections / more info:
Toby St Clere Smithe m...@tsmithe.net
writes:
test_direct_solvers.py:38: test_matrix_trans_lower_A_matrix_B_C_float32
test_direct_solvers.py:38:
test_matrix_slice_unit_upper_A_matrix_slice_B_C_float32
These mean:
1. matrix A was a lower triangular matrix
seems suspicious to me. But I'm also
running the tests on nVidia again (with debugging on and a clean kernel
cache this time), and so far, no segfault..
[1] http://paste.ubuntu.com/8815933/
[2] http://paste.ubuntu.com/8815949/
[3] http://paste.ubuntu.com/8816098/
Best,
Toby
Toby St Clere Smithe
AM, Toby St Clere Smithe wrote:
I deleted my kernel cache and rebuilt with debugging info on -- same
behaviour, and nothing obvious in the output. See [1]. I've also got a
more detailed traceback at [2]. I also investigated the arguments and
local variables in the add_program call[3
--
___
ViennaCL-devel mailing list
ViennaCL-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/viennacl-devel
--
Toby St Clere Smithe
http://tsmithe.net
we
just keep them separate for contexts associated with different
platforms?
Best,
Toby
2014-11-04 16:32 GMT-05:00 Toby St Clere Smithe m...@tsmithe.net:
Hej Philippe,
Philippe Tillet phil.til...@gmail.com
writes:
Sorry for the late answer. I've been extremely busy with my stats
OK, so 2 more weird problems. The rest of my tests pass (but I still
don't quite have complete test coverage).
1. Cache issues continue:
On both nVidia and Beignet, I get some error and a segfault like this on
sparse gemv:
Build Status = -2 ( Err = -11 )
The segfault happens when calling (in
Toby St Clere Smithe m...@tsmithe.net
writes:
The segfault happens when calling (in ocl/context.hpp):
443 err = clGetProgramBuildInfo(temp, devices_[0].id(),
CL_PROGRAM_BUILD_LOG, 0, NULL, ret_val_size);
Oh, and the segfault happens in nVidia's OpenCL when it calls strlen
somewhere
://paste.ubuntu.com/8874078/
The rest of my tests pass just fine.
Best,
Toby
--
Toby St Clere Smithe
http://tsmithe.net
--
___
ViennaCL-devel mailing list
ViennaCL-devel
into the GMRES issue...
OK.
Cheers,
T
--
Toby St Clere Smithe
http://tsmithe.net
--
___
ViennaCL-devel mailing list
ViennaCL-devel@lists.sourceforge.net
https
corrective actions from your mobile device.
http://pubads.g.doubleclick.net/gampad/clk?id=154624111iu=/4140/ostg.clktrk
--
Toby St Clere Smithe
http://tsmithe.net
--
Comprehensive Server Monitoring with Site24x7.
Monitor 10
78 matches
Mail list logo