Re: [petsc-users] Shell matrix MatCreateVecs

> On Aug 24, 2017, at 7:42 PM, Jed Brown wrote: > > "Jose E. Roman" writes: > >>> El 24 ago 2017, a las 22:51, Greg Meyer escribió: >>> >>> Hi, >>> >>> I have written a shell matrix for non-standard vectors (CUDA to be >>> specific) that works great. MatMult and MatNorm perform as they sho

Re: [petsc-users] Shell matrix MatCreateVecs

"Jose E. Roman" writes: >> El 24 ago 2017, a las 22:51, Greg Meyer escribió: >> >> Hi, >> >> I have written a shell matrix for non-standard vectors (CUDA to be specific) >> that works great. MatMult and MatNorm perform as they should. However, when >> I try to use it in a SLEPc algorithm, it

[petsc-users] Logo?

Hey folks, Does PETSc have any sort of official logo? I often like to include these in my acknowledgment slides but can't find any for PETSc. I'd appreciate if you could point me to if there is one. Mohammad

Re: [petsc-users] Shell matrix MatCreateVecs

> El 24 ago 2017, a las 22:51, Greg Meyer escribió: > > Hi, > > I have written a shell matrix for non-standard vectors (CUDA to be specific) > that works great. MatMult and MatNorm perform as they should. However, when I > try to use it in a SLEPc algorithm, it breaks because MatCreateVecs yi

[petsc-users] Shell matrix MatCreateVecs

Hi, I have written a shell matrix for non-standard vectors (CUDA to be specific) that works great. MatMult and MatNorm perform as they should. However, when I try to use it in a SLEPc algorithm, it breaks because MatCreateVecs yields standard vectors even if I set "-vec_type cuda". Looking through

Re: [petsc-users] MatMult performance on baij vs aij

Hi, For the same sparse matrix, will the storage format (baij or aij) have a noticeable performance effect on MatMult give the baij may have a better locality? The gains are mostly due to the reduced number of indices loaded from memory. For normal aij, each floating point number (typicall

Re: [petsc-users] MatMult performance on baij vs aij

Depending on the block size and the particular hardware and compiler BAIJ can give a noticeable improvement. Expect something between 1 and 2 times faster. Barry > On Aug 24, 2017, at 10:28 AM, Xiangdong wrote: > > Hello everyone, > > For the same sparse matrix, will the storage forma

[petsc-users] MatMult performance on baij vs aij

Hello everyone, For the same sparse matrix, will the storage format (baij or aij) have a noticeable performance effect on MatMult give the baij may have a better locality? Thank you. Best, Xiangdong

Re: [petsc-users] Petsc ILU PC Change between 3.6.4 and 3.7.x?

"Klaij, Christiaan" writes: > Matt, > > Thanks, I can understand the lower condition number of P A, but > what about r? Doesn't that change to P r and if so why can we > assume that ||r|| and ||P r|| have the same order? Matt's equation was of course wrong in a literal sense, but is based on the

Re: [petsc-users] Petsc with OpenMP

Hi Danyang, Thank you for the information. Khai From: Danyang Su [danyang...@gmail.com] Sent: Wednesday, August 23, 2017 12:52 PM To: Khai Pham; PETSc Subject: Re: [petsc-users] Petsc with OpenMP Hi Khai, I finally implemented LIS solver (linear algebra library

Re: [petsc-users] Petsc ILU PC Change between 3.6.4 and 3.7.x?

On Thu, Aug 24, 2017 at 2:29 AM, Klaij, Christiaan wrote: > Matt, > > Thanks, I can understand the lower condition number of P A, but > what about r? Doesn't that change to P r and if so why can we > assume that ||r|| and ||P r|| have the same order? > r and Pr are the things we can control. We m