LAPACK and ScaLAPACK Survey Results - by response

[ view answers grouped by question ]



Response #1: 2005-08-09

General

1. Is dense linear algebra a performance bottleneck in your applications?
Yes

2. How often do your applications use the arithmetic precisions listed below:

a. Single precision: Rarely
b. Double precision: Sometimes
c. More than double precision: Rarely
d. Complex single precision: Sometimes
e. Complex double precision: Very Frequently
f. Complex, more than double precision: Rarely

3. What dense matrix sizes are most important or time-consuming for your application?
100,000s X 100,000s

4. Does your application come close to, or run out of memory on important problems?
Yes

5. Number of processors used for your application:

a. SMP: More than 10
b. Distributed shared-memory: More than 10
c. Distributed memory: More than 100

6. Which architectures do you use or intend to use in the next three years?
Sequential, Multi-core-thread, Vector-computers, Distributed-shared-memory, Distributed-memory, Hybrid-shared

8. Please rank how the following features would be useful to your current or planned applications?

a. User defined matrix types: Not useful
b. Using optional arguments in the language interface: Somewhat useful
c. Automatic memory allocation of the work space: Very useful
d. More complicated matrix data structures: Very useful

9. Do your applications solve linear algebra problems of the type?
General linear systems

LAPACK Usage

1. Do you use LAPACK (or a vendor version of LAPACK )?
Yes

4. If you use LAPACK, do you use a vendors version or one obtained directly from Netlib?
Vendor, Netlib

5. If you have used both a vendors version of LAPACK and Netlibs, how do the two versions compare?
vendor's version may only have LU solver, not the full library

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
No

8. Do your applications use a higher-level interface to LAPACK?
No

10. Is the LAPACK procedure interface a barrier to more extensive use?
No

11. From which languages do you call LAPACK routines?
Fortran 90/95

14. If you have installed LAPACK yourself, how could the installation process be improved?
would be nice if the testing and verification process is quicker

15. How frequently do you refer to the LAPACK Users Guide?
Frequently

16. What information in the LAPACK guide is hard to find or is missing, if any?
More examples would be nice (more examples for scalapack would be nice as well)

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
Yes

4. If you use ScaLAPACK, do you use a vendors version or one obtained directly from Netlib?
Vendor

6. Do your applications make direct ScaLAPACK calls?
Yes

7. Do your applications use libraries which depend on ScaLAPACK?
No

8. Do your applications use a higher-level interface to ScaLAPACK?
No

10. Is the ScaLAPACK procedure interface a barrier to more extensive use?
No

11. From which languages do you call ScaLAPACK routines?
Fortran 90/95

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
Would be nice if there are tools or example to help in setting up the matrix, distributing data, reading/writing the matrix

14. If you have installed ScaLAPACK yourself, how could the installation process be improved?
Installation can be complicated involving BLACS, PBLAS, MPI, interfacing C and Fortran 77/90 compilers. Common problem is 1 versus 2 underscores in symbol names. Perhaps there can be C interface stubs to accomodate both versions.

15. How frequently do you refer to the ScaLAPACK Users Guide?
Frequently

16. What information in the ScaLAPACK guide is hard to find or is missing, if any?
Details in the code such as how the pivot vector is distributed and used, layout of data in distributed band solver, details about the QR factorization.

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, Linux, Unicos

3. If your applications run in a distributed-memory environment, which styles of parallelism do they employ?
Message passing


Additional Information

5.Description of related activities
Compact storage for scalapack. Out of core extension for scalapack.
6.Additional Comments/Suggestions
Would be nice if the technology in HPL (look ahead, recursive options) can be merged back into PZGETRF/PZGETRS. Currently HPL is only in double in C so double complex is not available. HPL works on rhs but conceptually an extra pass over L should make it compatible with PxGETRF. General symmetric linear solver (LTL') (perhaps in packed storage as well) would be nice. Since performance of PBLAS is crucial for scalapack, it would be nice if there are parameters to tune PBLAS, especially on triangular solvers. More examples or tutorials (like PETSc) to help new users to use scalapack. Better interfaces for scalapack with other iterative linear solvers such as PETSc or Object oriented CCA technology.








Sat May 25 11:27:18 2013
0 seconds