You can subscribe to this list here.
2010 
_{Jan}

_{Feb}

_{Mar}

_{Apr}

_{May}
(1) 
_{Jun}
(8) 
_{Jul}
(16) 
_{Aug}
(6) 
_{Sep}

_{Oct}

_{Nov}

_{Dec}
(5) 

2011 
_{Jan}
(4) 
_{Feb}
(3) 
_{Mar}
(5) 
_{Apr}

_{May}
(24) 
_{Jun}

_{Jul}
(5) 
_{Aug}
(17) 
_{Sep}

_{Oct}
(6) 
_{Nov}
(9) 
_{Dec}
(8) 
2012 
_{Jan}
(5) 
_{Feb}
(14) 
_{Mar}
(25) 
_{Apr}
(7) 
_{May}
(15) 
_{Jun}
(12) 
_{Jul}
(22) 
_{Aug}
(4) 
_{Sep}
(10) 
_{Oct}
(10) 
_{Nov}
(19) 
_{Dec}
(17) 
2013 
_{Jan}
(8) 
_{Feb}
(10) 
_{Mar}
(16) 
_{Apr}
(3) 
_{May}
(16) 
_{Jun}
(26) 
_{Jul}

_{Aug}
(9) 
_{Sep}

_{Oct}
(8) 
_{Nov}
(17) 
_{Dec}
(2) 
2014 
_{Jan}
(37) 
_{Feb}
(15) 
_{Mar}
(6) 
_{Apr}
(9) 
_{May}
(11) 
_{Jun}
(11) 
_{Jul}
(9) 
_{Aug}
(9) 
_{Sep}
(19) 
_{Oct}
(4) 
_{Nov}
(22) 
_{Dec}
(21) 
2015 
_{Jan}

_{Feb}
(7) 
_{Mar}
(2) 
_{Apr}
(17) 
_{May}
(22) 
_{Jun}
(11) 
_{Jul}
(11) 
_{Aug}
(6) 
_{Sep}
(7) 
_{Oct}

_{Nov}
(5) 
_{Dec}

2016 
_{Jan}
(1) 
_{Feb}
(3) 
_{Mar}
(4) 
_{Apr}
(8) 
_{May}
(8) 
_{Jun}
(11) 
_{Jul}
(2) 
_{Aug}

_{Sep}

_{Oct}

_{Nov}

_{Dec}

S  M  T  W  T  F  S 




1

2

3

4

5

6

7

8

9

10

11

12

13

14

15
(3) 
16
(2) 
17
(4) 
18
(1) 
19

20

21

22

23

24

25

26

27

28
(1) 
29
(3) 
30

31
(2) 

From: Shantanu Agarwal <shantanuagr123@gm...>  20130531 08:30:28

Hi, Yeah this fixes the issue. Thank you. Regards, Shantanu On Fri, May 31, 2013 at 7:25 AM, Karl Rupp <rupp@...> wrote: > Hi Shantanu, > > I could finally look into the issue. It has to do with NVIDIA tricking the > compiler environment: NVCC pretends to be a Visual Studio compiler, yet it > seems to use a different hostcompiler (Clang?) in the background. Hence, > some forward declarations get disabled. > > Workaround: Replace line 337 in viennacl/forwards.h > #ifndef _MSC_VER > by > #if !defined(_MSC_VER)  defined(__CUDACC__) > > Thanks and best regards, > Karli > > > > > On 05/18/2013 04:21 AM, Shantanu Agarwal wrote: > >> Hi, >> >> I am using visual studio 2010 to build the examples provided with >> ViennaCL. When I try to build a cuda project, I get errors of this type >>  'namespace "viennacl::linalg" has no member "inner_prod_cpu" ' for >> several members. All projects not involving CUDA are building fine. >> Please help. >> >> Thanks >> >> Regards, >> Shantanu >> >> >> **** >>  >> AlienVault Unified Security Management (USM) platform delivers complete >> security visibility with the essential security capabilities. Easily and >> efficiently configure, manage, and operate all of your security controls >> from a single console and one unified framework. Download a free trial. >> http://p.sf.net/sfu/**alienvault_d2d <http://p.sf.net/sfu/alienvault_d2d>; >> >> >> >> ______________________________**_________________ >> ViennaCLsupport mailing list >> ViennaCLsupport@...**sourceforge.net<ViennaCLsupport@...> >> https://lists.sourceforge.net/**lists/listinfo/viennacl**support<https://lists.sourceforge.net/lists/listinfo/viennaclsupport>; >> >> > 
From: Karl Rupp <rupp@iu...>  20130531 01:55:30

Hi Shantanu, I could finally look into the issue. It has to do with NVIDIA tricking the compiler environment: NVCC pretends to be a Visual Studio compiler, yet it seems to use a different hostcompiler (Clang?) in the background. Hence, some forward declarations get disabled. Workaround: Replace line 337 in viennacl/forwards.h #ifndef _MSC_VER by #if !defined(_MSC_VER)  defined(__CUDACC__) Thanks and best regards, Karli On 05/18/2013 04:21 AM, Shantanu Agarwal wrote: > Hi, > > I am using visual studio 2010 to build the examples provided with > ViennaCL. When I try to build a cuda project, I get errors of this type >  'namespace "viennacl::linalg" has no member "inner_prod_cpu" ' for > several members. All projects not involving CUDA are building fine. > Please help. > > Thanks > > Regards, > Shantanu > > >  > AlienVault Unified Security Management (USM) platform delivers complete > security visibility with the essential security capabilities. Easily and > efficiently configure, manage, and operate all of your security controls > from a single console and one unified framework. Download a free trial. > http://p.sf.net/sfu/alienvault_d2d > > > > _______________________________________________ > ViennaCLsupport mailing list > ViennaCLsupport@... > https://lists.sourceforge.net/lists/listinfo/viennaclsupport > 
From: Samar Vafai <samar.vafai@st...>  20130529 15:12:58

Thank you very much for your reply. Samar Quoting Karl Rupp <rupp@...>: > Dear Samar, > >> I need to solve a linear system of equations Ax=b >> by using the bicgstab iterative solver on GPU. >> The matrix is dense. The data type is cuDoublecomplex >> and the CUDA library is implemented. I would like to >> ask you if the Viennacl library supports the cuDoublecomplex >> data type? > > ViennaCL currently supports only real data types. > > >> Does the bicgstab iterative solver belonging to the library supports >> the dense matrices? > > Yes, the matrix type is generic and only needs to support the > matrixvector operation. One can even use "matrixfree" approaches > if required. > > >> and as much as I understand the library supports CUDA? > > CUDA is one of the computational backends, yes. > > >> In general, is it possible to call the function which is written for >> the sparse matrices for the dense matrices, instead? > > If you are talking about the BiCGStabfunction: yes, this is possible. > > >> In the case that I cannot implement the Viennacl library for my problem, >> do you have any suggestions? I mean is there any other libraries that I can >> use instead? > > As your matrix is dense, you are limited by the memory consumption > of the system matrix. Also, matrixvector products are inherently > memorybandwidth limited, so you won't get the high FLOP rates GPUs > offer. Instead, you might want to consider a dense LU factorization > as offered by e.g. MAGMA [1] also for complex data types. This way > you also don't need to worry about preconditioners. > > Best regards, > Karli > > [1] http://icl.cs.utk.edu/magma/ 
From: Karl Rupp <rupp@iu...>  20130529 13:03:59

Dear Samar, > I need to solve a linear system of equations Ax=b > by using the bicgstab iterative solver on GPU. > The matrix is dense. The data type is cuDoublecomplex > and the CUDA library is implemented. I would like to > ask you if the Viennacl library supports the cuDoublecomplex > data type? ViennaCL currently supports only real data types. > Does the bicgstab iterative solver belonging to the library supports > the dense matrices? Yes, the matrix type is generic and only needs to support the matrixvector operation. One can even use "matrixfree" approaches if required. > and as much as I understand the library supports CUDA? CUDA is one of the computational backends, yes. > In general, is it possible to call the function which is written for > the sparse matrices for the dense matrices, instead? If you are talking about the BiCGStabfunction: yes, this is possible. > In the case that I cannot implement the Viennacl library for my problem, > do you have any suggestions? I mean is there any other libraries that I can > use instead? As your matrix is dense, you are limited by the memory consumption of the system matrix. Also, matrixvector products are inherently memorybandwidth limited, so you won't get the high FLOP rates GPUs offer. Instead, you might want to consider a dense LU factorization as offered by e.g. MAGMA [1] also for complex data types. This way you also don't need to worry about preconditioners. Best regards, Karli [1] http://icl.cs.utk.edu/magma/ 
From: Samar Vafai <samar.vafai@st...>  20130529 10:12:11

Dear Viennacl people, I need to solve a linear system of equations Ax=b by using the bicgstab iterative solver on GPU. The matrix is dense. The data type is cuDoublecomplex and the CUDA library is implemented. I would like to ask you if the Viennacl library supports the cuDoublecomplex data type? Does the bicgstab iterative solver belonging to the library supports the dense matrices? and as much as I understand the library supports CUDA? In general, is it possible to call the function which is written for the sparse matrices for the dense matrices, instead? In the case that I cannot implement the Viennacl library for my problem, do you have any suggestions? I mean is there any other libraries that I can use instead? Thank you in adavance. Look forward to hearing from you. My Best Regards, Samar 
From: Anh Tuan <slam_ono@ya...>  20130528 08:21:40

Hi all, May I know whether I can use lanczos alorithm in GPU or multiCPU cores? From the example code, It seems only run on only one CPU core for me. Thanks 
From: Shantanu Agarwal <shantanuagr123@gm...>  20130518 09:22:03

Hi, I am using visual studio 2010 to build the examples provided with ViennaCL. When I try to build a cuda project, I get errors of this type  'namespace "viennacl::linalg" has no member "inner_prod_cpu" ' for several members. All projects not involving CUDA are building fine. Please help. Thanks Regards, Shantanu 
From: Karl Rupp <rupp@iu...>  20130517 19:03:50

Hi Kirell, I don't quite understand what you mean by 'no longer need to use templates'? Do you want to precompile the Lanczos implementation? Best regards, Karli On 05/17/2013 09:55 AM, Kirell Benzi wrote: > Hello, > Ok, so I guess if I need Lanczos, I no longer need to use templates and put the implementation in the cpp for faster compilation time. Or do you suggest I still use templates to support float and double ? > > Cheers. >  > > Kirell Benzi > PhD student, Research assistant, > Laboratoire de traitement des signaux 2 > Institute of Electrical Engineering  IEL > Ecole Polytechnique Fédérale de Lausanne  EPFL > EPFL/STI/IEL/LTS2 > ELE 229 (Bâtiment ELE) > Station 11 > CH1015 Lausanne > Switzerland > Home page: http://people.epfl.ch/kirell.benzi > Lab page: http://lts2.epfl.ch > Tel: +41 21 69 37895 >  > > On 17 mai 2013, at 16:50, Karl Rupp <rupp@...> > wrote: > >> Hi Kirell, >> >> the restriction to double is intentional: Roundoff errors cause the Lanczos algorithm to easily report socalled 'ghost eigenvalues'. To avoid these spurious ghosts, one has to do full/partial reorthogonalization and the like already for 'double'. You would barely get any reasonable results with float. >> >> Best regards, >> Karli >> >> >> On 05/17/2013 09:37 AM, Kirell Benzi wrote: >>> Hello, >>> It is possible to use the Lanczos algorithm for largest eigenvalues with >>> a compressed_matrix<float>. It seems only possible using doubles. >>> Thank you. >>> >>> Cheers, >>>  >>> >>> Kirell Benzi >>> PhD student, Research assistant, >>> Laboratoire de traitement des signaux 2 >>> Institute of Electrical Engineering  IEL >>> Ecole Polytechnique Fédérale de Lausanne  EPFL >>> EPFL/STI/IEL/LTS2 >>> ELE 229 (Bâtiment ELE) >>> Station 11 >>> CH1015 Lausanne >>> Switzerland >>> Home page: http://people.epfl.ch/kirell.benzi >>> Lab page: http://lts2.epfl.ch <http://lts2.epfl.ch/>; >>> Tel: +41 21 69 37895 >>>  >>> >>> >>> >>>  >>> AlienVault Unified Security Management (USM) platform delivers complete >>> security visibility with the essential security capabilities. Easily and >>> efficiently configure, manage, and operate all of your security controls >>> from a single console and one unified framework. Download a free trial. >>> http://p.sf.net/sfu/alienvault_d2d >>> >>> >>> >>> _______________________________________________ >>> ViennaCLsupport mailing list >>> ViennaCLsupport@... >>> https://lists.sourceforge.net/lists/listinfo/viennaclsupport >>> >> > > 
From: Kirell Benzi <kirell.benzi@ep...>  20130517 14:55:53

Hello, Ok, so I guess if I need Lanczos, I no longer need to use templates and put the implementation in the cpp for faster compilation time. Or do you suggest I still use templates to support float and double ? Cheers.  Kirell Benzi PhD student, Research assistant, Laboratoire de traitement des signaux 2 Institute of Electrical Engineering  IEL Ecole Polytechnique Fédérale de Lausanne  EPFL EPFL/STI/IEL/LTS2 ELE 229 (Bâtiment ELE) Station 11 CH1015 Lausanne Switzerland Home page: http://people.epfl.ch/kirell.benzi Lab page: http://lts2.epfl.ch Tel: +41 21 69 37895  On 17 mai 2013, at 16:50, Karl Rupp <rupp@...> wrote: > Hi Kirell, > > the restriction to double is intentional: Roundoff errors cause the Lanczos algorithm to easily report socalled 'ghost eigenvalues'. To avoid these spurious ghosts, one has to do full/partial reorthogonalization and the like already for 'double'. You would barely get any reasonable results with float. > > Best regards, > Karli > > > On 05/17/2013 09:37 AM, Kirell Benzi wrote: >> Hello, >> It is possible to use the Lanczos algorithm for largest eigenvalues with >> a compressed_matrix<float>. It seems only possible using doubles. >> Thank you. >> >> Cheers, >>  >> >> Kirell Benzi >> PhD student, Research assistant, >> Laboratoire de traitement des signaux 2 >> Institute of Electrical Engineering  IEL >> Ecole Polytechnique Fédérale de Lausanne  EPFL >> EPFL/STI/IEL/LTS2 >> ELE 229 (Bâtiment ELE) >> Station 11 >> CH1015 Lausanne >> Switzerland >> Home page: http://people.epfl.ch/kirell.benzi >> Lab page: http://lts2.epfl.ch <http://lts2.epfl.ch/>; >> Tel: +41 21 69 37895 >>  >> >> >> >>  >> AlienVault Unified Security Management (USM) platform delivers complete >> security visibility with the essential security capabilities. Easily and >> efficiently configure, manage, and operate all of your security controls >> from a single console and one unified framework. Download a free trial. >> http://p.sf.net/sfu/alienvault_d2d >> >> >> >> _______________________________________________ >> ViennaCLsupport mailing list >> ViennaCLsupport@... >> https://lists.sourceforge.net/lists/listinfo/viennaclsupport >> > 
From: Karl Rupp <rupp@iu...>  20130517 14:50:52

Hi Kirell, the restriction to double is intentional: Roundoff errors cause the Lanczos algorithm to easily report socalled 'ghost eigenvalues'. To avoid these spurious ghosts, one has to do full/partial reorthogonalization and the like already for 'double'. You would barely get any reasonable results with float. Best regards, Karli On 05/17/2013 09:37 AM, Kirell Benzi wrote: > Hello, > It is possible to use the Lanczos algorithm for largest eigenvalues with > a compressed_matrix<float>. It seems only possible using doubles. > Thank you. > > Cheers, >  > > Kirell Benzi > PhD student, Research assistant, > Laboratoire de traitement des signaux 2 > Institute of Electrical Engineering  IEL > Ecole Polytechnique Fédérale de Lausanne  EPFL > EPFL/STI/IEL/LTS2 > ELE 229 (Bâtiment ELE) > Station 11 > CH1015 Lausanne > Switzerland > Home page: http://people.epfl.ch/kirell.benzi > Lab page: http://lts2.epfl.ch <http://lts2.epfl.ch/>; > Tel: +41 21 69 37895 >  > > > >  > AlienVault Unified Security Management (USM) platform delivers complete > security visibility with the essential security capabilities. Easily and > efficiently configure, manage, and operate all of your security controls > from a single console and one unified framework. Download a free trial. > http://p.sf.net/sfu/alienvault_d2d > > > > _______________________________________________ > ViennaCLsupport mailing list > ViennaCLsupport@... > https://lists.sourceforge.net/lists/listinfo/viennaclsupport > 
From: Kirell Benzi <kirell.benzi@ep...>  20130517 14:38:17

Hello, It is possible to use the Lanczos algorithm for largest eigenvalues with a compressed_matrix<float>. It seems only possible using doubles. Thank you. Cheers,  Kirell Benzi PhD student, Research assistant, Laboratoire de traitement des signaux 2 Institute of Electrical Engineering  IEL Ecole Polytechnique Fédérale de Lausanne  EPFL EPFL/STI/IEL/LTS2 ELE 229 (Bâtiment ELE) Station 11 CH1015 Lausanne Switzerland Home page: http://people.epfl.ch/kirell.benzi Lab page: http://lts2.epfl.ch<http://lts2.epfl.ch/>; Tel: +41 21 69 37895  
From: Karl Rupp <rupp@iu...>  20130516 16:53:13

Hi Kirell, you're welcome :) I opened a ticket so that this functionality hopefully makes it into the next release: https://github.com/viennacl/viennacldev/issues/22 Best regards, Karli On 05/16/2013 09:55 AM, Kirell Benzi wrote: > Hello, > It is working, thank you very much ! > > Cheers, >  > > Kirell Benzi > PhD student, Research assistant, > Laboratoire de traitement des signaux 2 > Institute of Electrical Engineering  IEL > Ecole Polytechnique Fédérale de Lausanne  EPFL > EPFL/STI/IEL/LTS2 > ELE 229 (Bâtiment ELE) > Station 11 > CH1015 Lausanne > Switzerland > Home page: http://people.epfl.ch/kirell.benzi > Lab page: http://lts2.epfl.ch > Tel: +41 21 69 37895 >  > > On 15 mai 2013, at 17:56, Karl Rupp <rupp@...> wrote: > >> Hi Kirell, >> >> if you're using OpenCL, there's a way for promoting a vector to a matrix as follows: >> >> vector<double> x = prod(A, y); // this is your SpMV operation >> >> // create the proxy object. Adjust num_cols, num_rows is probably 1 >> matrix<double> x2(x.handle().opencl_handle(), num_rows, num_cols); >> >> your_range += x2; //here you add to your subrange >> >> >> This currently doesn't work with CUDA or the OpenMP backends, though. >> >> Best regards, >> Karli >> >> >> >> On 05/15/2013 10:48 AM, Kirell Benzi wrote: >>> Hi, >>> In append I mean setting the entries. >>> >>> something like this: >>> >>> matrix.at(row) += myVector; >>> >>> I have tried to create the vector as a matrix, but this vector is the result of of compressed_matrix * vector. The sparse matrix * matrix is not implemented is viennacl, so I need to transform my vector in a matrix. >>> >>> The other solution would be to have a std::vector< viennacl:::vector> > instead of a matrix. But it will be slow since for each operation I would have to switch back and forth with the cpu_ram. >>> >>> Do you have any advices ? >>> >>> ++ >>> >>>  >>> >>> Kirell Benzi >>> PhD student, Research assistant, >>> Laboratoire de traitement des signaux 2 >>> Institute of Electrical Engineering  IEL >>> Ecole Polytechnique Fédérale de Lausanne  EPFL >>> EPFL/STI/IEL/LTS2 >>> ELE 229 (Bâtiment ELE) >>> Station 11 >>> CH1015 Lausanne >>> Switzerland >>> Home page: http://people.epfl.ch/kirell.benzi >>> Lab page: http://lts2.epfl.ch >>> Tel: +41 21 69 37895 >>>  >>> >>> On 15 mai 2013, at 17:40, Karl Rupp <rupp@...> >>> wrote: >>> >>>> Hi Kirell, >>>> >>>>> I'm new to viennacl. >>>>> I want to add a row vector to a specific row of in matrix. >>>>> I have created a matrix_range in a for loop to specify the row I want. >>>>> Unfortunately, I cannot append a vector to a matrix range. >>>>> >>>>> I have tried to transform my vector to a matrix, but it seems impossible. >>>> >>>> what exactly do you mean by 'append'? Should a NxNmatrix become a (N+1)xNmatrix, or do you just want to set the entries? >>>> >>>> In the former case, you need to resize() the matrix and refill it. In the latter case you currently need to create the 'vector' as a 1xN or 1xMmatrix and then assign to the matrixrange. We don't yet have the generic row() and column() functions ready to directly assign vectors to a matrix row/column. >>>> >>>> Best regards, >>>> Karli >>> >>> >> > > 
From: Kirell Benzi <kirell.benzi@ep...>  20130516 14:56:17

Hello, It is working, thank you very much ! Cheers,  Kirell Benzi PhD student, Research assistant, Laboratoire de traitement des signaux 2 Institute of Electrical Engineering  IEL Ecole Polytechnique Fédérale de Lausanne  EPFL EPFL/STI/IEL/LTS2 ELE 229 (Bâtiment ELE) Station 11 CH1015 Lausanne Switzerland Home page: http://people.epfl.ch/kirell.benzi Lab page: http://lts2.epfl.ch Tel: +41 21 69 37895  On 15 mai 2013, at 17:56, Karl Rupp <rupp@...> wrote: > Hi Kirell, > > if you're using OpenCL, there's a way for promoting a vector to a matrix as follows: > > vector<double> x = prod(A, y); // this is your SpMV operation > > // create the proxy object. Adjust num_cols, num_rows is probably 1 > matrix<double> x2(x.handle().opencl_handle(), num_rows, num_cols); > > your_range += x2; //here you add to your subrange > > > This currently doesn't work with CUDA or the OpenMP backends, though. > > Best regards, > Karli > > > > On 05/15/2013 10:48 AM, Kirell Benzi wrote: >> Hi, >> In append I mean setting the entries. >> >> something like this: >> >> matrix.at(row) += myVector; >> >> I have tried to create the vector as a matrix, but this vector is the result of of compressed_matrix * vector. The sparse matrix * matrix is not implemented is viennacl, so I need to transform my vector in a matrix. >> >> The other solution would be to have a std::vector< viennacl:::vector> > instead of a matrix. But it will be slow since for each operation I would have to switch back and forth with the cpu_ram. >> >> Do you have any advices ? >> >> ++ >> >>  >> >> Kirell Benzi >> PhD student, Research assistant, >> Laboratoire de traitement des signaux 2 >> Institute of Electrical Engineering  IEL >> Ecole Polytechnique Fédérale de Lausanne  EPFL >> EPFL/STI/IEL/LTS2 >> ELE 229 (Bâtiment ELE) >> Station 11 >> CH1015 Lausanne >> Switzerland >> Home page: http://people.epfl.ch/kirell.benzi >> Lab page: http://lts2.epfl.ch >> Tel: +41 21 69 37895 >>  >> >> On 15 mai 2013, at 17:40, Karl Rupp <rupp@...> >> wrote: >> >>> Hi Kirell, >>> >>>> I'm new to viennacl. >>>> I want to add a row vector to a specific row of in matrix. >>>> I have created a matrix_range in a for loop to specify the row I want. >>>> Unfortunately, I cannot append a vector to a matrix range. >>>> >>>> I have tried to transform my vector to a matrix, but it seems impossible. >>> >>> what exactly do you mean by 'append'? Should a NxNmatrix become a (N+1)xNmatrix, or do you just want to set the entries? >>> >>> In the former case, you need to resize() the matrix and refill it. In the latter case you currently need to create the 'vector' as a 1xN or 1xMmatrix and then assign to the matrixrange. We don't yet have the generic row() and column() functions ready to directly assign vectors to a matrix row/column. >>> >>> Best regards, >>> Karli >> >> > 
From: Karl Rupp <rupp@iu...>  20130515 15:56:48

Hi Kirell, if you're using OpenCL, there's a way for promoting a vector to a matrix as follows: vector<double> x = prod(A, y); // this is your SpMV operation // create the proxy object. Adjust num_cols, num_rows is probably 1 matrix<double> x2(x.handle().opencl_handle(), num_rows, num_cols); your_range += x2; //here you add to your subrange This currently doesn't work with CUDA or the OpenMP backends, though. Best regards, Karli On 05/15/2013 10:48 AM, Kirell Benzi wrote: > Hi, > In append I mean setting the entries. > > something like this: > > matrix.at(row) += myVector; > > I have tried to create the vector as a matrix, but this vector is the result of of compressed_matrix * vector. The sparse matrix * matrix is not implemented is viennacl, so I need to transform my vector in a matrix. > > The other solution would be to have a std::vector< viennacl:::vector> > instead of a matrix. But it will be slow since for each operation I would have to switch back and forth with the cpu_ram. > > Do you have any advices ? > > ++ > >  > > Kirell Benzi > PhD student, Research assistant, > Laboratoire de traitement des signaux 2 > Institute of Electrical Engineering  IEL > Ecole Polytechnique Fédérale de Lausanne  EPFL > EPFL/STI/IEL/LTS2 > ELE 229 (Bâtiment ELE) > Station 11 > CH1015 Lausanne > Switzerland > Home page: http://people.epfl.ch/kirell.benzi > Lab page: http://lts2.epfl.ch > Tel: +41 21 69 37895 >  > > On 15 mai 2013, at 17:40, Karl Rupp <rupp@...> > wrote: > >> Hi Kirell, >> >>> I'm new to viennacl. >>> I want to add a row vector to a specific row of in matrix. >>> I have created a matrix_range in a for loop to specify the row I want. >>> Unfortunately, I cannot append a vector to a matrix range. >>> >>> I have tried to transform my vector to a matrix, but it seems impossible. >> >> what exactly do you mean by 'append'? Should a NxNmatrix become a (N+1)xNmatrix, or do you just want to set the entries? >> >> In the former case, you need to resize() the matrix and refill it. In the latter case you currently need to create the 'vector' as a 1xN or 1xMmatrix and then assign to the matrixrange. We don't yet have the generic row() and column() functions ready to directly assign vectors to a matrix row/column. >> >> Best regards, >> Karli > > 
From: Karl Rupp <rupp@iu...>  20130515 15:40:55

Hi Kirell, > I'm new to viennacl. > I want to add a row vector to a specific row of in matrix. > I have created a matrix_range in a for loop to specify the row I want. > Unfortunately, I cannot append a vector to a matrix range. > > I have tried to transform my vector to a matrix, but it seems impossible. what exactly do you mean by 'append'? Should a NxNmatrix become a (N+1)xNmatrix, or do you just want to set the entries? In the former case, you need to resize() the matrix and refill it. In the latter case you currently need to create the 'vector' as a 1xN or 1xMmatrix and then assign to the matrixrange. We don't yet have the generic row() and column() functions ready to directly assign vectors to a matrix row/column. Best regards, Karli 
From: Kirell Benzi <kirell.benzi@ep...>  20130515 15:03:45

Hello, I'm new to viennacl. I want to add a row vector to a specific row of in matrix. I have created a matrix_range in a for loop to specify the row I want. Unfortunately, I cannot append a vector to a matrix range. I have tried to transform my vector to a matrix, but it seems impossible. Any hints ? Thank you, Cheers  Kirell Benzi PhD student, Research assistant, Laboratoire de traitement des signaux 2 Institute of Electrical Engineering  IEL Ecole Polytechnique Fédérale de Lausanne  EPFL EPFL/STI/IEL/LTS2 ELE 229 (Bâtiment ELE) Station 11 CH1015 Lausanne Switzerland Home page: http://people.epfl.ch/kirell.benzi Lab page: http://lts2.epfl.ch<http://lts2.epfl.ch/>; Tel: +41 21 69 37895  