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: Never
b. Double precision: Frequently
c. More than double precision: Never
d. Complex single precision: Never
e. Complex double precision: Never
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
10,000s X 10,000s

4. Does your application come close to, or run out of memory on important problems?
No

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, Symmetric-multi-procs, Distributed-shared-memory, Distributed-memory, Hybrid-shared

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
PETSc

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

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

9. Do your applications solve linear algebra problems of the type?
Banded linear systems, General linear systems, Least-square problems, SVDs

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

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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?
C, Fortran 77, Fortran 90/95, R

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

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?
Yes

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

10. Is the ScaLAPACK procedure interface a barrier to more extensive use?
Yes

11. From which languages do you call ScaLAPACK routines?
C

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
A consistent interface for the QR routines with pivoting would be most useful. The public version of QR routines seem to fail for one-column matrices.

14. If you have installed ScaLAPACK yourself, how could the installation process be improved?
Definitely can be improved compared to GNU and R packages or even PETSc which requires more work than the other two.

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?
A clear and concise definition of the API would be most welcome. The explanations in the comments is commonly found to be confusing among would be user.

Targeted Environment Specifics

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

2. If your applications run in a shared-memory environment, which styles of parallelism do they employ?
Platform threading


Additional Information

7. Use DOE-lab resourcesYes



Response #2: 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: Never
b. Double precision: Very Frequently
c. More than double precision: Never
d. Complex single precision: Never
e. Complex double precision: Very Frequently
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
1,000s X 1,000s

4. Does your application come close to, or run out of memory on important problems?
No

5. Number of processors used for your application:

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

6. Which architectures do you use or intend to use in the next three years?
Distributed-memory

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

a. User defined matrix types:
b. Using optional arguments in the language interface:
c. Automatic memory allocation of the work space: Very useful
d. More complicated matrix data structures:

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

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

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 77, Fortran 90/95

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

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

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

Targeted Environment Specifics

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


Additional Information

7. Use DOE-lab resourcesYes



Response #3: 2005-08-09

General

1. Is dense linear algebra a performance bottleneck in your applications?
No

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

a. Single precision: Sometimes
b. Double precision: Very Frequently
c. More than double precision: Rarely
d. Complex single precision: Sometimes
e. Complex double precision: Frequently
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
1,000,000 or more X 1,000,000 or more

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: Less than 10
b. Distributed shared-memory: More than 10
c. Distributed memory: Less than 10

6. Which architectures do you use or intend to use in the next three years?
Sequential, Distributed-shared-memory, Distributed-memory

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
no

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: Not useful
c. Automatic memory allocation of the work space: Somewhat useful
d. More complicated matrix data structures: Not useful

9. Do your applications solve linear algebra problems of the type?
Symmetric eigenvalue, SVDs

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?
Netlib

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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

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

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

2. If you do not use ScaLAPACK, why?
Not solving linear algebra

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

Targeted Environment Specifics

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

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


Additional Information

7. Use DOE-lab resourcesYes



Response #4: 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: Very Frequently
c. More than double precision: Never
d. Complex single precision: Very Frequently
e. Complex double precision: Sometimes
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
1,000s X 1,000s

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, Distributed-memory, Hybrid-shared

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
fftw

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

a. User defined matrix types: Very 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: Somewhat useful

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, Banded linear systems, General linear systems, Generalized eigenvalue, Symmetric eigenvalue

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?
Roughly equal, given well optimized BLAS (e.g. Atlas)

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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?
C, C++, Fortran 77, Fortran 90/95, Python

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using LAPACK?
Numeric Python

13. How could the LAPACK interface be improved to feel more natural to your application and implementation language?
Better integration with Python

14. If you have installed LAPACK yourself, how could the installation process be improved?
It could not be improved

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

16. What information in the LAPACK guide is hard to find or is missing, if any?
N/A

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?
Netlib

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?
Yes

11. From which languages do you call ScaLAPACK routines?
Fortran 77

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using ScaLAPACK?
None, although I'd like some

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
Better interface to Blacs. The redist utils are a great start, but they're poorly documented, and frequently when the system administrators install scalapack they don't even know that they should also build redist.

14. If you have installed ScaLAPACK yourself, how could the installation process be improved?
It's pretty good, although I recall having to go through a lot of configuration tweaking to make it work. Autoconf might help here.

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

16. What information in the ScaLAPACK guide is hard to find or is missing, if any?
redist!! Plus, information *about* redistributing matrices with/without the redist utilities. Scalapack is of no use to anyone if people can't get their matrices distributed on the computers.

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
Linux, Mac OS X

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
Working on density matrix techniques as eigensolver replacements in quantum chemistry. This relies heavily on (sca)lapack.
6.Additional Comments/Suggestions
Lapack is one of the best things that has ever happened to me. Everyone of the lapack workers that I've interacted with has bent over backwards to be helpful. This survey is yet another example of how the lapack people care about being useful to their community. If you ever need someone to write glowing letters of support for grant applications (particularly someone from a US National Lab other than ORNL), don't hesitate to contact me at the address above.
7. Use DOE-lab resourcesYes



Response #5: 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: Very Frequently
c. More than double precision: Never
d. Complex single precision: Never
e. Complex double precision: Never
f. Complex, more than double precision: Never

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

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 100
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, Symmetric-multi-procs, Vector-computers, Distributed-shared-memory, Distributed-memory, Hybrid-shared

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
scilib

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: Not useful
c. Automatic memory allocation of the work space: Very useful
d. More complicated matrix data structures: Somewhat useful

9. Do your applications solve linear algebra problems of the type?
General linear systems, Generalized eigenvalue, Symmetric eigenvalue

LAPACK Usage

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

2. If you do not use LAPACK, why?
Cost of learning

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?
Netlib's version provides the source code, very helpful in porting. I use vendor's LAPACK functions whenever possible.

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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 77

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

2. If you do not use ScaLAPACK, why?
Cost of learning

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, IRIX, Linux, Unicos, Windows (other)


Additional Information

7. Use DOE-lab resourcesYes



Response #6: 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: Never
b. Double precision: Very Frequently
c. More than double precision: Never
d. Complex single precision: Never
e. Complex double precision: Sometimes
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
1,000s X 1,000s

4. Does your application come close to, or run out of memory on important problems?
No

5. Number of processors used for your application:

a. SMP: More than 10
b. Distributed shared-memory: More than 100
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, Symmetric-multi-procs, Vector-computers, Distributed-shared-memory, Distributed-memory, Hybrid-shared

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
Peigs

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

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

9. Do your applications solve linear algebra problems of the type?
Generalized eigenvalue, Symmetric eigenvalue

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

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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?
C, Fortran 77

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

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?
Netlib

6. Do your applications make direct ScaLAPACK calls?
No

7. Do your applications use libraries which depend on ScaLAPACK?
Yes

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

11. From which languages do you call ScaLAPACK routines?
Fortran 77

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using ScaLAPACK?
Global Arrays

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
64-bit integer arguments are needed (Fortran Integer*8 or C/C++ long long).

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, HP/UX, IRIX, Linux, Solaris, Tru64, Unicos

2a. Please specify any particular libraries of frameworks used?
Global Arrays (that uses System V shared memory)

3a. Please specify any particular libraries of frameworks used?
Global Arrays


Additional Information

5.Description of related activities
Development of the computational chemistry software NWChem http://www.emsl.pnl.gov/docs/nwchem
7. Use DOE-lab resourcesYes



Response #7: 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: Sometimes
b. Double precision: Very Frequently
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?
1,000s X 1,000s

4. Does your application come close to, or run out of memory on important problems?
No

5. Number of processors used for your application:

a. SMP: Less than 10
b. Distributed shared-memory: More than 100
c. Distributed memory: More than 10,000

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

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

a. User defined matrix types: Very 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: Somewhat useful

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, Banded linear systems, General linear systems, Generalized eigenvalue

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?
I usually try to use the vendor's version first and if it is not available, costs too much, or for some other reason too much work, use Netlib's version next.

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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?
C, C++, Fortran 77, Fortran 90/95

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using LAPACK?
We have built a C++ interface for BLAS and LAPACK that uses the ability to overload arguments and removes the need to encode precision into the name of the method/subroutine. We have also defined the simples possible matrix concept and use generic linear algebar operations that interface to LAPACK and BLAS.

13. How could the LAPACK interface be improved to feel more natural to your application and implementation language?
Short of doing a full thing with a generic linear algebra library (modeled after blitz++ or MTL), I favor the pragmatic and simple approach I described in item 12.

14. If you have installed LAPACK yourself, how could the installation process be improved?
This is a minor issue: complete the LAPACK implementation in ATLAS so that optimized BLAS and LAPACK can be installed in one package.

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

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, BSD, Linux, Solaris, Tru64, Unicos, Windows (cygwin), Mac OS X

2. If your applications run in a shared-memory environment, which styles of parallelism do they employ?
Multiple

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


Additional Information

6.Additional Comments/Suggestions
I am currently not using ScaLAPACK but I probably should. I would like to spend some time thinking about how to best integrate with C++. I also would like to discuss is the way to use BLAS and LAPACK from C++ can be standardised. Several people have asked me for help, and I'm offering the interfaces we developed in the psimag toolkit. It would be nice if we could decide on properly suporting one model.
7. Use DOE-lab resourcesYes



Response #8: 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: Very Frequently
b. Double precision: Very Frequently
c. More than double precision: Never
d. Complex single precision: Very Frequently
e. Complex double precision: Very Frequently
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
1,000s X 1,000s

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:
b. Distributed shared-memory:
c. Distributed memory: More than 10

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

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: Not useful
c. Automatic memory allocation of the work space: Not useful
d. More complicated matrix data structures: Very useful

9. Do your applications solve linear algebra problems of the type?
Generalized eigenvalue, Symmetric eigenvalue, SVDs

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?
Mostly interchangeable. However, on occasion, the flexibility of configuring through ATLAS has been a key enabler.

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 77

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using LAPACK?
None.

13. How could the LAPACK interface be improved to feel more natural to your application and implementation language?
I actually find the procedural interface as it stands quite natural, but then I'm one of the f77 dinosaurs.

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

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 77

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using ScaLAPACK?
None.

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
A more extensive, and *universal*, pblas library. Portability of code is of paramount concern (for maintainability), and many of the support routines are not universal, and therefore the underlying distributed data structures need to be deciphered and the necessary functions hand-coded. Very ppu (personal processing unit) time-consuming.

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

16. What information in the ScaLAPACK guide is hard to find or is missing, if any?
Google . Works quite well, actually.

Targeted Environment Specifics

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

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


Additional Information

6.Additional Comments/Suggestions
I hope that, whatever changes are made to "modernize" lapack, that they be made in the manner to augment the current basic procedural functionality rather than replace it.
7. Use DOE-lab resourcesYes



Response #9: 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: Sometimes
b. Double precision: Very Frequently
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?
10,000s X 10,000s

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: Less than 10
b. Distributed shared-memory: More than 10
c. Distributed memory: More than 10

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

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
IBM ESSL

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

a. User defined matrix types: Somewhat 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: Not useful

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

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?
Netlib

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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?
C, Fortran 77, Fortran 90/95

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

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?
Netlib

6. Do your applications make direct ScaLAPACK calls?
Yes

7. Do your applications use libraries which depend on ScaLAPACK?
Yes

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?
C

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, IRIX, Linux, Windows (cygwin)

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


Additional Information

7. Use DOE-lab resourcesYes



Response #10: 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: Very Frequently
c. More than double precision: Frequently
d. Complex single precision: Rarely
e. Complex double precision: Very Frequently
f. Complex, more than double precision: Frequently

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:
b. Distributed shared-memory: More than 10
c. Distributed memory: More than 10

6. Which architectures do you use or intend to use in the next three years?
Sequential, Distributed-memory

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

a. User defined matrix types: Somewhat 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: Somewhat useful

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

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

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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

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

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?
Yes

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

10. Is the ScaLAPACK procedure interface a barrier to more extensive use?
Yes

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

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

Targeted Environment Specifics

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

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


Additional Information

7. Use DOE-lab resourcesYes



Response #11: 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: Never
b. Double precision: Frequently
c. More than double precision: Never
d. Complex single precision: Never
e. Complex double precision: Very Frequently
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
10,000s X 10,000s

4. Does your application come close to, or run out of memory on important problems?
No

5. Number of processors used for your application:

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

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

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
ARPACK and PARPACK

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: Somewhat useful

9. Do your applications solve linear algebra problems of the type?
General linear systems, Generalized eigenvalue, Non-symmetric eigenvalue

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?
No real differences.

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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 77, Fortran 90/95

14. If you have installed LAPACK yourself, how could the installation process be improved?
More automatic configuration, and some basic tuning.

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

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

2. If you do not use ScaLAPACK, why?
Cost of learning

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

Targeted Environment Specifics

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

2. If your applications run in a shared-memory environment, which styles of parallelism do they employ?
Platform threading

2a. Please specify any particular libraries of frameworks used?
IBM smp essl

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

3a. Please specify any particular libraries of frameworks used?
petsc


Additional Information

7. Use DOE-lab resourcesYes



Response #12: 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: Never
b. Double precision: Very Frequently
c. More than double precision: Never
d. Complex single precision: Never
e. Complex double precision: Very Frequently
f. Complex, more than double precision: Never

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

6. Which architectures do you use or intend to use in the next three years?
Distributed-shared-memory, Distributed-memory, Hybrid-shared

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
ESSL

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

a. User defined matrix types: Very 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: Not useful

9. Do your applications solve linear algebra problems of the type?
Symmetric eigenvalue

LAPACK Usage

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

3. If using another package, which one(s)?
PARPACK

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

5. If you have used both a vendors version of LAPACK and Netlibs, how do the two versions compare?
vendor version is substantially faster.

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?
C++, Fortran 90/95

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

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?
C++

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
It is very inconvenient to prepare the data for ScaLAPACK. Should be more natural, at least as natural as lapack.

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

Targeted Environment Specifics

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

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
electronic structure calculation
7. Use DOE-lab resourcesYes



Response #13: 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: Very Frequently
c. More than double precision: Rarely
d. Complex single precision: Rarely
e. Complex double precision: Sometimes
f. Complex, more than double precision: Rarely

3. What dense matrix sizes are most important or time-consuming for your application?
10,000s

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: Less than 10
b. Distributed shared-memory: More than 100
c. Distributed memory: More than 1,000

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

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
linpack

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

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

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, Banded linear systems, General linear systems, Generalized eigenvalue, Least-square problems, SVDs

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?
well.

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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?
C, C++, Fortran 77, Fortran 90/95, tcl

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

ScaLAPACK Usage

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, BSD, IRIX, Linux, Solaris, Unicos, Windows (other)

2. If your applications run in a shared-memory environment, which styles of parallelism do they employ?
Multiple


Additional Information

6.Additional Comments/Suggestions
general question 8 is a bit unclearly stated.
7. Use DOE-lab resourcesYes



Response #14: 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: Frequently
b. Double precision: Frequently
c. More than double precision:
d. Complex single precision:
e. Complex double precision: Frequently
f. Complex, more than double precision:

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?
No

5. Number of processors used for your application:

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

6. Which architectures do you use or intend to use in the next three years?
Distributed-memory

9. Do your applications solve linear algebra problems of the type?
Generalized eigenvalue, Non-symmetric eigenvalue

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?
Netlib

5. If you have used both a vendors version of LAPACK and Netlibs, how do the two versions compare?
Vendor faster but Vendor is incomplete or undefined API (Lapack 1? 2? 3?)

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

8. Do your applications use a higher-level interface to LAPACK?
Yes

9. If you answered yes above, which higher-level interfaces do you use?
R

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

11. From which languages do you call LAPACK routines?
C, C++, Fortran 77, Fortran 90/95

14. If you have installed LAPACK yourself, how could the installation process be improved?
Place in autoconfig construct as a package in itself Place in autoconfig construct with scalapack, lapack, blas ./configure --with[out]-(scalapack,lapack,blas) --with-blas=ibm_essl

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

ScaLAPACK Usage

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

Targeted Environment Specifics


Additional Information

7. Use DOE-lab resourcesYes



Response #15: 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: Never
b. Double precision: Frequently
c. More than double precision: Very Frequently
d. Complex single precision:
e. Complex double precision:
f. Complex, more than double precision:

3. What dense matrix sizes are most important or time-consuming for your application?
10,000s X 10,000s

4. Does your application come close to, or run out of memory on important problems?
No

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 10

6. Which architectures do you use or intend to use in the next three years?
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: Not useful
c. Automatic memory allocation of the work space: Not useful
d. More complicated matrix data structures: Not useful

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, Banded linear systems, 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

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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

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

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

2. If you do not use ScaLAPACK, why?
Use another pkg

3. If using another package, which one(s)?
SuperLU

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

Targeted Environment Specifics

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

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


Additional Information

7. Use DOE-lab resourcesYes



Response #16: 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: Rarely
c. More than double precision: Rarely
d. Complex single precision: Rarely
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 100
c. Distributed memory: More than 1,000

6. Which architectures do you use or intend to use in the next three years?
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: Somewhat useful
b. Using optional arguments in the language interface: Not useful
c. Automatic memory allocation of the work space: Very useful
d. More complicated matrix data structures: Not 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?
Presume vendor version provides better performance.

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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

11. From which languages do you call LAPACK routines?
Fortran 77

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

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 77

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, Linux, Unicos, xt3 catamount

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

3a. Please specify any particular libraries of frameworks used?
Direct calls to the BLACS


Additional Information

7. Use DOE-lab resourcesYes



Response #17: 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: Never
b. Double precision: Very Frequently
c. More than double precision: Never
d. Complex single precision: Never
e. Complex double precision: Rarely
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
1,000s X 1,000s

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: Less than 10
b. Distributed shared-memory: Less than 10
c. Distributed memory: More than 10

6. Which architectures do you use or intend to use in the next three years?
Sequential, Distributed-memory

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
PLAPACK

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: Not useful
d. More complicated matrix data structures: Very useful

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, Symmetric eigenvalue

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?
I use MKL on Windows machines and ATLAS on Linux machines. Both perform better than reference implementations.

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

8. Do your applications use a higher-level interface to LAPACK?
Yes

9. If you answered yes above, which higher-level interfaces do you use?
Matlab, PETSc

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

11. From which languages do you call LAPACK routines?
C, C++, Matlab

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using LAPACK?
I typically write "wrapper functions" that call LAPACK while hiding many of its arguments from the nonlinear solver or optimization solver. When using C++, these functions inherit the interface from an abstract matrix class.

13. How could the LAPACK interface be improved to feel more natural to your application and implementation language?
Some examples that call LAPACK from C would be helpful.

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

16. What information in the LAPACK guide is hard to find or is missing, if any?
Performance information is difficult to find. For instance, it took me a while to confirm my observation that the Cholesky factorization using packed symmetric format were much slower than Cholesky factorizations using full dense format.

ScaLAPACK Usage

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

3. If using another package, which one(s)?
PLAPACK

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

6. Do your applications make direct ScaLAPACK calls?
Yes

7. Do your applications use libraries which depend on ScaLAPACK?
Yes

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?
C

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using ScaLAPACK?
I typically write "wrapper functions" that call LAPACK while hiding many of its arguments from the nonlinear solver or optimization solver.

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
Some examples that call ScaLAPACK from C would be helpful.

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
Linux, Windows (cygwin), Windows (other)

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

3a. Please specify any particular libraries of frameworks used?
MPI


Additional Information

5.Description of related activities
Numical optimization solvers that I develop and distribute target applications nonlinear and semidefinite programming. I use LAPACK to factor dense matrices that are usually the Schur complements of much larger indefinite matrices. Parallel versions of these solvers use ScaLAPACK. Users of our software must also link to these packages.
6.Additional Comments/Suggestions
The most frequent difficulty that my users encounter while linking their application to my optimization solvers and LAPACK involves language interoperability. Despite our best efforts to help them configure things properly, calling the Fortran routines from C or C++ continues to frustrate users who are uninterested in these subtleties of computer science. I would also like to see the BLAS1 include y = alpha y + x, w = alpha x + beta y, and BLAS2 include alphs = x' A x where x is a column vector and A is a symmetric matrix.
7. Use DOE-lab resourcesYes



Response #18: 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: Sometimes
b. Double precision: Very Frequently
c. More than double precision: Never
d. Complex single precision: Sometimes
e. Complex double precision: Very Frequently
f. Complex, more than double precision: Never

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 1,000
b. Distributed shared-memory: More than 1,000
c. Distributed memory: More than 1,000

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

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
no

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

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

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on 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 77, Fortran 90/95

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

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?
Netlib

6. Do your applications make direct ScaLAPACK calls?
Yes

7. Do your applications use libraries which depend on ScaLAPACK?
Yes

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 77

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

Targeted Environment Specifics

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

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


Additional Information

7. Use DOE-lab resourcesYes



Response #19: 2005-08-10

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: Never
b. Double precision: Very Frequently
c. More than double precision: Rarely
d. Complex single precision: Never
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?
1,000,000 or more

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 100
b. Distributed shared-memory: More than 100
c. Distributed memory: More than 100

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

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
ESSL, NAG

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

a. User defined matrix types: Somewhat 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, Symmetric eigenvalue

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

5. If you have used both a vendors version of LAPACK and Netlibs, how do the two versions compare?
Prefer vendor's version.

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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?
C, C++

13. How could the LAPACK interface be improved to feel more natural to your application and implementation language?
Definitely oppose to any f90 interface. f77/C as is makes it very easy to use with any applications.

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

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?
C

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
The same: f77/C interfaces are sufficient.

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, Linux, Windows (cygwin), MAC OS X

2. If your applications run in a shared-memory environment, which styles of parallelism do they employ?
Multiple

2a. Please specify any particular libraries of frameworks used?
OpenMP

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

3a. Please specify any particular libraries of frameworks used?
MPI


Additional Information

5.Description of related activities
Condensed matter theory; scientific code developments
7. Use DOE-lab resourcesYes



Response #20: 2005-08-11

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

3. What dense matrix sizes are most important or time-consuming for your application?
1,000,000 or more

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 100
b. Distributed shared-memory: More than 100
c. Distributed memory: More than 100

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

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
Yes

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

a. User defined matrix types: Very useful
b. Using optional arguments in the language interface: Very useful
c. Automatic memory allocation of the work space:
d. More complicated matrix data structures: Somewhat useful

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, General linear systems, Generalized eigenvalue

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?
Netlib

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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?
C, Fortran 90/95

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

16. What information in the LAPACK guide is hard to find or is missing, if any?
I would like to see a list of NAG routines which have corresponding LAPACK equivalents.

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?
Netlib

6. Do your applications make direct ScaLAPACK calls?
Yes

7. Do your applications use libraries which depend on ScaLAPACK?
Yes

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?
C

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

Targeted Environment Specifics

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

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


Additional Information

7. Use DOE-lab resourcesYes



Response #21: 2005-08-15

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: Never
b. Double precision: Very Frequently
c. More than double precision: Never
d. Complex single precision: Never
e. Complex double precision: Sometimes
f. Complex, more than double precision: Never

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:
b. Distributed shared-memory:
c. Distributed memory: More than 1,000

6. Which architectures do you use or intend to use in the next three years?
Distributed-memory

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
PLAPACK

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: Very useful
c. Automatic memory allocation of the work space: Very useful
d. More complicated matrix data structures: Somewhat useful

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, 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?
Vender is much faster

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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?
C, Fortran 90/95

13. How could the LAPACK interface be improved to feel more natural to your application and implementation language?
Heavy use of C++

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

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?
Netlib

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?
C

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
Heavy use of C++

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, IRIX, Linux, Mac OS X, Solaris

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
BEM application developer
7. Use DOE-lab resourcesYes



Response #22: 2005-08-15

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: Very Frequently
c. More than double precision: Rarely
d. Complex single precision: Rarely
e. Complex double precision: Rarely
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
10,000s X 10,000s

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 1,000

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

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
PLAPACK, Pliris

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

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

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, General linear systems, Generalized eigenvalue, Symmetric eigenvalue, SVDs

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 versions are *usually* faster, but not always.

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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?
C, C++

13. How could the LAPACK interface be improved to feel more natural to your application and implementation language?
We have C++ application codes. We would like to have matrix 'objects' which would bundle size information, and possibly have various format/layout options (e.g. packed symmetric, etc.).

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

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?
Netlib

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?
C++

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using ScaLAPACK?
Getting the matrix into a 2-D block cyclic distribution is non-trivial.

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
Some object-orientation in terms of matrix types (maybe bundling the array descriptors with the matrix, etc.) would be useful from a C++ application's perspective.

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, IRIX, Linux, Solaris, Tru64

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
We use ScaLAPACK in a boundary element method application. We also have finite element method applications which use sparse solvers (both iterative and direct) which in turn use BLAS and LAPACK heavily.
7. Use DOE-lab resourcesYes



Response #23: 2005-08-15

General

1. Is dense linear algebra a performance bottleneck in your applications?
No

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

a. Single precision: Never
b. Double precision: Frequently
c. More than double precision: Never
d. Complex single precision: Never
e. Complex double precision: Never
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
1,000s

4. Does your application come close to, or run out of memory on important problems?
No

5. Number of processors used for your application:

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

6. Which architectures do you use or intend to use in the next three years?
Sequential, Distributed-memory

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

a. User defined matrix types: Somewhat 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: Somewhat useful

9. Do your applications solve linear algebra problems of the type?
Banded linear systems, 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?
Netlib

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?
Yes

9. If you answered yes above, which higher-level interfaces do you use?
MATLAB

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

11. From which languages do you call LAPACK routines?
C, C++

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using LAPACK?
Wrappers that handle memory allocation.

13. How could the LAPACK interface be improved to feel more natural to your application and implementation language?
Automated memory allocation.

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

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

2. If you do not use ScaLAPACK, why?
Not solving linear algebra

6. Do your applications make direct ScaLAPACK calls?
No

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

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

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

Targeted Environment Specifics

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

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


Additional Information

7. Use DOE-lab resourcesYes



Response #24: 2005-08-23

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

3. What dense matrix sizes are most important or time-consuming for your application?
1,000s X 1,000s

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: Less than 10
b. Distributed shared-memory: Less than 10
c. Distributed memory: Less than 10

6. Which architectures do you use or intend to use in the next three years?
Sequential, Multi-core-thread, 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: Somewhat 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?
Generalized eigenvalue, Symmetric eigenvalue, Least-square problems

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, Other

5. If you have used both a vendors version of LAPACK and Netlibs, how do the two versions compare?
Vendor's version (MLIB) optimized much better and thus faster

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

8. Do your applications use a higher-level interface to LAPACK?
Yes

9. If you answered yes above, which higher-level interfaces do you use?
Matlab

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

11. From which languages do you call LAPACK routines?
Fortran 77, Fortran 90/95, Matlab

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

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

2. If you do not use ScaLAPACK, why?
Cost of learning

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
HP/UX, Linux, Tru64, Windows (cygwin)

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
COndensed matter physics, high-temperature superconductivity
7. Use DOE-lab resourcesYes



Response #25: 2005-08-23

General

1. Is dense linear algebra a performance bottleneck in your applications?
No

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

a. Single precision: Rarely
b. Double precision: Very Frequently
c. More than double precision: Sometimes
d. Complex single precision: Never
e. Complex double precision: Never
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
1,000s X 1,000s

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:
b. Distributed shared-memory:
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, Symmetric-multi-procs, 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: Somewhat useful
b. Using optional arguments in the language interface: Not useful
c. Automatic memory allocation of the work space: Very useful
d. More complicated matrix data structures: Not useful

LAPACK Usage

1. Do you use LAPACK (or a vendor version of LAPACK )?
No

2. If you do not use LAPACK, why?
Not solving linear algebra

6. Do your applications make direct LAPACK calls?
No

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

15. How frequently do you refer to the LAPACK Users Guide?
Never

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

2. If you do not use ScaLAPACK, why?
Not solving linear algebra

6. Do your applications make direct ScaLAPACK calls?
No

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

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, BSD, HP/UX, IRIX, Linux, Mac OS X, Solaris

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


Additional Information

7. Use DOE-lab resourcesYes



Response #26: 2005-08-25

General

1. Is dense linear algebra a performance bottleneck in your applications?
No

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

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

3. What dense matrix sizes are most important or time-consuming for your application?
10,000s X 10,000s

4. Does your application come close to, or run out of memory on important problems?
No

5. Number of processors used for your application:

a. SMP: Less than 10
b. Distributed shared-memory: More than 10
c. Distributed memory: More than 1,000

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

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

a. User defined matrix types: Very useful
b. Using optional arguments in the language interface: Very useful
c. Automatic memory allocation of the work space: Somewhat useful
d. More complicated matrix data structures: Very useful

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, Banded linear systems, General linear systems, Generalized eigenvalue, Symmetric eigenvalue

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 releases seem only marginally optimized over reasonable compilation of netlibs code. ATLAS libraries are critical for any real performance bottle neck.

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

8. Do your applications use a higher-level interface to LAPACK?
Yes

9. If you answered yes above, which higher-level interfaces do you use?
python/scipy

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

11. From which languages do you call LAPACK routines?
C, Fortran 77, Fortran 90/95, Python

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

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

2. If you do not use ScaLAPACK, why?
Not solving linear algebra

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
Linux, Mac OS X, Tru64, Windows (cygwin)

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
I work on the development and implementation of multilevel iterative algortihms on various parallel machines. The underlying linear systems are large and sparse, and so LAPACK routines are typically used for various local (serial) computations in a distributed memory environment. Increasingly we are using shared memory nodes within large clusters, however, hybrid memory models have yet to pay off. This may change as OpenMP and other threading capabilities improve.
6.Additional Comments/Suggestions
I think one of the frustrating parts of the LAPACK libraries is the build and patch system. I think moving to a modern revision control system (such as subversion), along with a good set of open-source development tools accessible through the web, introducing supported language bindings (e.g., python), and finally a more reasonable build system is critical to the next generation of users.
7. Use DOE-lab resourcesYes



Response #27: 2005-08-26

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

3. What dense matrix sizes are most important or time-consuming for your application?
1,000s X 1,000s

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

6. Which architectures do you use or intend to use in the next three years?
Distributed-memory

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

a. User defined matrix types:
b. Using optional arguments in the language interface:
c. Automatic memory allocation of the work space: Very useful
d. More complicated matrix data structures:

9. Do your applications solve linear algebra problems of the type?
General linear systems, Least-square problems, SVDs

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

5. If you have used both a vendors version of LAPACK and Netlibs, how do the two versions compare?
I was never able to get the NEtlib version to pass all it's own tests. I have not submitted the vendor package to the same tests!

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 77, Fortran 90/95

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

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

3. If using another package, which one(s)?
LAPACK - SCALAPACK does not support SVD

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

Targeted Environment Specifics


Additional Information

5.Description of related activities
Use D H Bailey's MP and Yozo's QD - for extra precison.
6.Additional Comments/Suggestions
1. A quad-precision version of LAPACK would be very useful. 2. After using the matrix solvers in LAPACK for many years, I am convinced there is a deep problem that exhibits itself in the SVD routines.
7. Use DOE-lab resourcesYes



Response #28: 2005-08-29

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: Sometimes
b. Double precision: Very Frequently
c. More than double precision: Rarely
d. Complex single precision: Frequently
e. Complex double precision: Sometimes
f. Complex, more than double precision: Never

3. What dense matrix sizes are most important or time-consuming for your application?
10,000s X 100,000s

4. Does your application come close to, or run out of memory on important problems?
No

5. Number of processors used for your application:

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

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

a. User defined matrix types: Somewhat useful
b. Using optional arguments in the language interface: Somewhat useful
c. Automatic memory allocation of the work space: Somewhat useful
d. More complicated matrix data structures: Very useful

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, Banded linear systems, General linear systems, Generalized eigenvalue, Symmetric eigenvalue, Non-symmetric eigenvalue, SVDs, singular 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?
Netlib

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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 77

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using LAPACK?
matrix plotters (ascii, graphical, etc)

13. How could the LAPACK interface be improved to feel more natural to your application and implementation language?
possibility to have more info about lower level steps in higher level routines

14. If you have installed LAPACK yourself, how could the installation process be improved?
ok as is

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

ScaLAPACK Usage

1. Do you use ScaLAPACK (or a vendor version of ScaLAPACK )?
No

3. If using another package, which one(s)?
symmlq

6. Do your applications make direct ScaLAPACK calls?
No

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

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

14. If you have installed ScaLAPACK yourself, how could the installation process be improved?
Did not get it to run, stopped bothering, handpicked some other routines from netlib.

15. How frequently do you refer to the ScaLAPACK Users Guide?
Never

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
Linux, Mac OS X


Additional Information

5.Description of related activities
PDE's in control, boundary control, conservation laws
6.Additional Comments/Suggestions
I like Lapack, thank you.
7. Use DOE-lab resourcesYes



Response #29: 2005-08-29

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

3. What dense matrix sizes are most important or time-consuming for your application?
100s X 10,000s

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: Less than 10
b. Distributed shared-memory: More than 100
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, Symmetric-multi-procs, 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: Somewhat useful
b. Using optional arguments in the language interface: Not useful
c. Automatic memory allocation of the work space: Not useful
d. More complicated matrix data structures: Very useful

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, 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, Other

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

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

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

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

14. If you have installed ScaLAPACK yourself, how could the installation process be improved?
Our users do report to us difficultes to have compatible and efficient versions of SCalappack/BLACS/MPI running correctly on their computer.

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

Targeted Environment Specifics

1. Under which operating system environments do your applications run?
AIX, Linux, Solaris, Unicos, Windows (cygwin)

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
Sparse linear algebra
6.Additional Comments/Suggestions
Functionalities for symmetric indefinite matrices are missing in Scalappack.
7. Use DOE-lab resourcesYes



Response #30: 2005-09-14

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: Very Frequently
c. More than double precision: Rarely
d. Complex single precision: Rarely
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?
10,000s X 10,000s

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 1,000
c. Distributed memory: More than 1,000

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

7. Do you use any other sequential or parallel dense linear algebra packages other than LAPACK or ScaLAPACK?
PLAPACK

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

a. User defined matrix types: Somewhat 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: Somewhat useful

9. Do your applications solve linear algebra problems of the type?
Linear positive definite systems, General linear systems, Symmetric eigenvalue, SVDs

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?
Netlib

6. Do your applications make direct LAPACK calls?
Yes

7. Do your applications use libraries which depend on LAPACK?
Yes

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, Fortran 2003

14. If you have installed LAPACK yourself, how could the installation process be improved?
see Scalapack comments below.

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?
See Scalapack comments below.

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?
Netlib

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 77

12. Please describe any tools or helper functions that you frequently implement to assist your applications in using ScaLAPACK?
PBLAS redistribution routines: pdelset, pdelget, indxl2g etc.

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
Get rid of the dependance on BLACS. BLACS contexts in particular are unwieldy and difficult to use and understand.

14. If you have installed ScaLAPACK yourself, how could the installation process be improved?
Clearer options for compiling Single/Double/Real/Complex libraries only. Better and clearer validation routines.

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

16. What information in the ScaLAPACK guide is hard to find or is missing, if any?
More information on the underlying algorithms used in the various routines would be useful.

Targeted Environment Specifics

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

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
Atomic, Molecular and Optical Physics, Computational Chemistry.
6.Additional Comments/Suggestions
I would be very interested in using the MRRR (Multiple Relatively Robust Representations) eigensolver algorithm. I would be very grateful if you could feedback to me any plans you have to incorporate this into future releases of Scalapack, as this could influence the direction of my future research.
7. Use DOE-lab resourcesYes








Thu May 23 02:19:51 2013
1 seconds