Search found 13 matches

by ReL
Tue Jul 31, 2012 8:16 am
Forum: User discussion
Topic: Minor bug in magma_?tritri[_gpu]
Replies: 1
Views: 1292

Minor bug in magma_?tritri[_gpu]

Hello, I have just noticed something while reading the source code of magma_?trtri: #define A(i, j) (a +(j)*lda + (i)) ... /* Check for singularity if non-unit */ if (nounit) { for (*info=0; *info < n; *info=*info+1) { if(A(*info,*info)==0) return *info; } *info=0; } It seems you are comparing the p...
by ReL
Tue Jul 10, 2012 5:41 am
Forum: User discussion
Topic: magma_dgeqrf documentation
Replies: 0
Views: 1182

magma_dgeqrf documentation

Hello again, It looks like the documentation magma_dgeqrf functions is not up-to-date: A (input/output) DOUBLE_PRECISION array on the GPU, dimension (LDDA,N) On entry, the M-by-N matrix A. On exit, the elements on and above the diagonal of the array contain the min(M,N)-by-N upper trapezoidal matrix...
by ReL
Tue Jul 10, 2012 2:45 am
Forum: User discussion
Topic: [SOLVED] magma_dsyevd wrong optimal WORK array size ?
Replies: 2
Views: 1032

Re: magma_dsyevd wrong optimal WORK array size ?

Thanks for the fast answer!
by ReL
Mon Jul 09, 2012 10:28 am
Forum: User discussion
Topic: [SOLVED] magma_dsyevd wrong optimal WORK array size ?
Replies: 2
Views: 1032

[SOLVED] magma_dsyevd wrong optimal WORK array size ?

Hello,

When using magma_dsyevd_gpu I run into errors if I use the optimal WORK array size computed by the function in query mode.

If I increase the array size of N elements, it seems to fix the issue. I would guess that the function isn't returning the correct optimal size for workspace.

Rémi
by ReL
Mon Jul 09, 2012 10:22 am
Forum: User discussion
Topic: Incorrect memory accesses in MAGMA implementation of dtrsm
Replies: 2
Views: 1324

Re: Incorrect memory accesses in MAGMA implementation of dtr

Hello,

Just to let you know that I tested that code against MAGMA 1.2.1 and magmablas_dtrsm still seems to be failing.

Regards,
Rémi
by ReL
Mon Jul 09, 2012 10:02 am
Forum: User discussion
Topic: [SOLVED] errors when using magma_dgels_gpu
Replies: 2
Views: 1448

Re: errors when using magma_dgels_gpu

Hello Stan,

Indeed the problem is fixed in MAGMA 1.2.1. Thanks for looking into that.

Rémi
by ReL
Fri Jun 22, 2012 4:07 am
Forum: User discussion
Topic: [SOLVED] magma_dprint(_gpu) and strange CuRand behavior
Replies: 2
Views: 1672

Re: magma_dprint(_gpu) and strange CuRand behavior

Thanks for the detailed answer.

I had a look at CURAND documention and it seems that this behavior is intended. I should have checked that before... Thanks again for the workaround.
by ReL
Wed Jun 20, 2012 10:42 am
Forum: User discussion
Topic: [SOLVED] errors when using magma_dgels_gpu
Replies: 2
Views: 1448

[SOLVED] errors when using magma_dgels_gpu

Hello, I have noticed some errors when using magma_dgels_gpu: CUBLAS error: memory mapping error (11) in magma_dsetmatrix at dset_get.cpp:98 testMagmaDGELS: dset_get.cpp:160: void magma_dcopymatrix(int, int, const double *, int, double *, int): Assertion `status == cudaSuccess' failed. magma_dgels3_...
by ReL
Wed Jun 20, 2012 9:02 am
Forum: User discussion
Topic: [SOLVED] magma_dprint(_gpu) and strange CuRand behavior
Replies: 2
Views: 1672

[SOLVED] magma_dprint(_gpu) and strange CuRand behavior

Hello, I have noticed a really strange behavior when the function magma_dprint(_gpu) is used along with NVIDIA CuRand library (random number generation on GPU). I create a uniform random number generator with a fixed seed, I use it to create 4 random vectors with 5 elements each and print them using...
by ReL
Tue Jun 12, 2012 5:32 am
Forum: User discussion
Topic: Incorrect memory accesses in MAGMA implementation of dtrsm
Replies: 2
Views: 1324

Re: Incorrect memory accesses in MAGMA implementation of dtr

Hello,

I have attached a short code sample that enables to reproduce the problem directly using magmablas_dtrsm.
testDtrsm.cpp
Code sample
(1.68 KiB) Downloaded 138 times
I hope it will help debugging that issue.

Rémi