PARDISO Scalability

PARDISO Scalability

Hi,

I have a question about scalability of the PARDISO solver. I'm using Intel MKL PARDISO with my Finite Element code. I tested simple linear elastic problem with around 800 000 unknowns. The time results are:

 

threads
pardiso phase 22
pardiso phase 33
sum pardiso
speed up pardiso phase 22
speedup pardiso phase 33
speed up sum pardiso

1
1433.922
18.8455
1452.7675
1
1
1

2
720.3288
10.3364
730.6652
1.9906492702
1.8232169808
1.988280679

4
381.09201
7.709
388.80101
3.7626661341
2.4446101959
3.7365322173

6
285.75699
8.2753
294.03229
5.0179769881
2.2773192513
4.9408434019

8
234.9064
7.7265
242.6329
6.1042270453
2.4390733191
5.9875124107

10
225.3418
9.33
233.6718
6.3633200764
2.0198821008
6.2171280403

12
205.16
7.4491
212.6091
6.9892864106
2.5299029413
6.8330447756

 

I've tested it on a node with 2 processors with 6 cores Xeon X5650 with 2,66Ghz and 24 GB RAM DDR3 1333MHz .

I'm wondering if speed up around 7 is ok or I may get better speed up, when I play a little bit with input parameters for PARDISO?

Is there some articles or papers about scalability of the PARDISO solver? Could you pass me some links to them?

My input parameters are:

     iparm(1) = 1      
      iparm(3) = 0
      iparm(4) = 0
      iparm(5) = 0
      iparm(6) = 0

      iparm(7) = 0
      iparm(8) =10 
      iparm(9) = 0 
      iparm(10) = 8
      iparm(11) = 1
      iparm(12) = 0

      iparm(13) = 1
      iparm(14) = 0
      iparm(15) = 0
      iparm(16) = 0
      iparm(17) = 0
      iparm(18) = -1
      iparm(19) = 0 
      iparm(20) = 0
      iparm(21) = 1
      iparm(22) = 0
      iparm(23) = 0      
    iparm(24) = 0
      iparm(25) = 0 
      iparm(27) = 0 
      iparm(28) = 0
    iparm(30) = 0
    iparm(31) = 0
    iparm(35) = 0 
      iparm(60) =0

Thanks for all advices.

best regards,

Pawel J.

 

3 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

Sorry my table was unreadable. I've attached screenshot from Calc.

best regards,
Pawel J.

Pawel,

  The speedup of around 7 seems to be fine. Since the matrix size is large, you may try the Cluster PARDISO from MKL.  Please find details here.

https://software.intel.com/en-us/articles/intel-math-kernel-library-parallel-direct-sparse-solver-for-clusters

--Vipin

Leave a Comment

Please sign in to add a comment. Not a member? Join today