Response #1: 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:

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:

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-memory, Hybrid-shared

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

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

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

LAPACK Usage

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

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

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?
Sun Performance Library (Solaris 8, 9) had too many bugs. We were on the phone for half and hour or more each time we called in bug reports to Sun and finally decided that their implementation was worthless; we stopped using it and reverted to netlib sources.

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?
FLAME interfaces

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

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

13. How could the LAPACK interface be improved to feel more natural to your application and implementation language?
Hide work arrays, leading dimensions, matrix dimensions, and possibly other matrix information (symmetric, triangular) beneath an abstraction layer (a la FLAME objects). Do not require user to provide workspace.

14. If you have installed LAPACK yourself, how could the installation process be improved?
LAPACK currently has no configuration support. Manually editing makefiles is fine for most wonks like us, but it's not very "professional" for an NSF project funded at LAPACK's level. Autoconf configure scripts would be very slick and help provide built-in support for building LAPACK on various frequently-used architectures.

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

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

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

13. How could the ScaLAPACK interface be improved to feel more natural to your application and implementation language?
Same as LAPACK: desperately needed abstraction from details, memory allocation, etc.

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, SUPER-UX

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?
PLAPACK