the whole harmonious society of co-operating sequential processes

advertisement
Parallelism and Distributed
Applications
Daniel S. Katz
Director, Cyberinfrastructure and User Services,
Center for Computation & Technology
Associate Research Professor, Electrical and
Computer Engineering Department
AT LOUISIANA STATE UNIVERSITY
Context
• Scientific/Engineering applications
– Complex, multi-physics, multiple time scales, multiple
spatial scales
• Physics components
– Elements such as I/O, solvers, etc.
• Computer Science components
– Parallelism across components
– Parallelism within components, particularly physics
components
• Goal: efficient application execution on both
parallel and distributed platforms
• Goal: simple, reusable programming
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
2
Types of Systems
• A lot of levels/layers to be aware of:
– Individual computers
• Many layers of memory hierarchy
• Multi-core -> many-core CPUs
– Clusters
• Used to be reasonably-tightly coupled computers (1 CPU per node)
or SMPs (multiple CPUs per node)
– Grids elements
•
•
•
•
•
•
•
Individual computers
Clusters
Networks
Instruments
Data stores
Visualization systems
Etc…
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
3
Types of Applications
• Applications can be broken up into pieces (components)
– Size (granularity) and relationship of pieces is key
• Fairly large pieces, no dependencies
– Parameter sweeps, Monte Carlo analysis, etc.
• Fairly large pieces, some dependencies
– Multi-stage applications - PHOEBUS
– Workflow applications - Montage
– Data grid apps?
• Large pieces, tight dependencies (coupling, components?)
– Distributed viz, coupled apps - Climate
• Small pieces, no dependencies
• Small pieces, some dependencies
– Dataflow?
• Small pieces, tight dependencies
– MPI apps
• Hybrids?
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
4
Parallelism within programs
• Initial parallelism: bitwise/vector (SIMD)
– “Highly computational tasks often contain substantial amounts of
concurrency. At LLL the majority of these programs use very large, twodimensional arrays in a cyclic set of instructions. In many cases, al
new array values could be computed simultaneously, rather than
stepping through one position at a time. To date, vectorization has
been the most effective scheme for exploiting this concurrency.
However, pipelining and independent multiprocessing forms of
concurrency are also available in these programs, but neither the
hardware not the software exist to make it workable.” (James R.
McGraw, Data Flow Computing: The VAL Language, MIT
Computational Structures Group Memo 188, 1980)
– Westinghouse’s Solomon introduced vector processing, early 1960s
– Continued in ILLIAC IV, ~1970s
– Goodyear MPP, 128x128 array of 1 bit processors , ~1980
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
5
Unhappy with
your
programming
model?
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
6
Parallelism across programs
• Co-operating Sequential Processes (CSP) - E. W. Dijkstra, The Structure of
the “THE”-Multiprogramming System, 1968
– “We have given full recognition of the fact that in a single sequential process …
only the time succession of the various states has a logical meaning, but not
the actual speed with which the sequential process is performed. Therefore we
have arranged the whole system as a society of sequential processes,
progressing with undefined speed ratios. To each user program … corresponds
a sequential process …”
– “This enabled us to design the whole system in terms of these abstract
"sequential processes". Their harmonious co-operation is regulated by means
of explicit mutual synchronization statements. … The fundamental
consequence of this approach … is that the harmonious co-operation of a set of
such sequential processes can be established by discrete reasoning; as a
further consequence the whole harmonious society of co-operating
sequential processes is independent of the actual number of processors
available to carry out these processes, provided the processors available can
switch from process to process.”
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
7
Parallelism within programs (2)
• MIMD
– Taxonomy from Flynn, 1972
– Dataflow parallelism
• “The data flow concept incorporates these forms of concurrency in one
basic graph-oriented system. Every computation is represented by a
data flow graph. The nodes … represent operations, the directed arcs
represent data paths.” (McGraw, ibid)
• “The ultimate goal of data flow software must be to help identify
concurrency in algorithms and map as much as possible into the
graphs.” (McGraw ibid)
– Transputer - 1984
• programmed in occam
– Uses CSP formalism, communication through named channels
– MPPs - mid 1980s
• Explicit message passing (CSP)
– Other models: actors, Petri nets, …
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
8
PHOEBUS
MPP MACHINE
MESH
Radar Cross Section
170
220
270
320
20
70
120
+
170
220
270
0
50
100
150
200 250
320
300
COLU MN I NDEX
COLU MN I NDEX
BB B
BB
B
B
BBB
B
BB
B
BB BBBBBB
BBBB B B
BBBB
BB
BB
BB
BBBB
BBB
BB
BB
BB
BBB
BB
BBB
BBBB
B BB
BB
B
B
BBBB
BB
BB
BB
BBB
BBB
B
B
BB
B
B B
BB
B BBBB BB
B BB
BBBB B B B
BBB
B
BBB BB
BBB
BBB BB
BBB
BB
B
BBB
BBBB BB
B B BB
BB B
BB
B
BBBB
BB
BB
B
BBB
BBBBB
BB
BBB
BBB
BB B BBB
BB
BBB B BBB
BBBB
B B BB
BBB
BB B BBBB
BBBB
B BB B
B B BBBB
BB
BB BBB
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
BBBBBB
BB
B BB B
B BB
B BB
BB
B
BB
B
BB
BBB
B
BBB
BBBBB
BB
BBB
BBB
BBB
BBB
BB
B
BBB BB
B
B
BBB
BBBBB
BB
BBB
BBB
B BB BBBB BBB
BB
BBB
BBB
BB
BB B
B
BBB BBB
B B BB
BB
B
B
BBB
B
BBB
BBBBB
BB
BBB
BBB
BB
BB B
BB
BBB
BB
BBB
BBBBB
BB
BBB
BBB
B B BBB
B BB
BB
B
B
B
B BB
B B B
BB BBBB
B B
BB
BB
B
B
BB
B
BBB
BBBBB
BB
BBB
BBB
B BB
BB
BB
B
BBB
B
B
B BBB B
BB
BB
BB
BB
BB
B BB BB B
B
BB B
B
BBBB
B
BB
BB
BBBBB
BB
B
BB
BB
B
B
B
B BB
BB BB
B BB
B B B
B BB
B BBBB
BB
B B BBB
B B
BB
BBBB B
BBBB
BB
BBB
BBBBB
BBBB BB
BBB
BBBBB
BB
BBB
BBB
B
B BBBB
BB
BB BB BB
BB
BB BBB
BBBBB
BB
BBB
BBB
B BBBB
BB
BB
B B
BB
B
B BB BBBBBBBBB
BB
BBB
BBBBB
BB
BBB
BBB
B B BBBB
BB
B BB
BB
BBBB B
BBB
BBBBB
BB
BBB
BBB
B B B B
BBB
BBB
B
BB BB B BB BBBB
BB
BBBB
B
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
B
BB BBBB
BBB
BB
BB
B B
BB B B
BB
BBB
BBBBB
BB
BBB
BBB
B BBB
B
BBB B B
BB
B
B
B
BBB
BBBBB
BBBBB
BB
BBB
BBB
B
BB
BB B BB B BB
B
BB
BBB
B
BB
B
BBBB
BB
BBB
BBB
B B B
BB
BB
BB
BBB
BBB
BB
BB
BBBB
B
B
B
B
BB B
BB
BBBB
B BB B
BBB
BB
B
B BBB
BB
B
BB
BB B B
B BB B B B
B
B BB BB B
B BBBBBB
BB
B B BBB B B
B
BBB
BBBBB
BB
BBB
BBB
B
B BB BB
B B BB
BB
BBB
B B BB
B
BBB
BBBBB
BB
BBB
BBB
B B BBB BB B
B
B B BBB
BBB
BB
B
BB
BB
BBB
B
BB
BBB
BBBBB
BB
BBB
BBB
B
BB
B
B
BBBBB
BB
BB B
B B BB
B
B
BBB
BBBBB
BB
BBB
BBB
B BB
BB
BBB
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
B BBBBB BB
BB
BB
BB
BB
B
BB
BBBB
B
BBB B BBBB
BB
B B B BB
BBB
B
B
B BB
B B BB BBB BB
B
BBB
BBBBB
BB
BBB
BBB
B
B
B B BB BB
B B
B BB
BB
B BB
BB BB
BBB
B
BBB
BB BB
BB B B
B
B
BBB
BBBBB
BB
BBB
BBB
B
BBB
B BB
BB
BB
BBB
BB
BB BB BB
BB
BB
B
B BB
BB
BB
BBB
B
B
B
BB BB
BBB
BB
B
BB
BBBBB
B
B B
BB
BBBB B
BB BB
B
BB B
BB
BBBB
BB BB
B
BBB
BBBBB
BB
BBB
BBB
B
B
B BBBBB
BBB
BBB
BB
BB
BB
BB
BBB
B
BB
BBB
BBBBB
BB
BBB
BBB
B
B
B B
BBBBB
BBB
BB
BB
BBBB B
BBB
BBBBB
BB
BBB
BBB
B
B BB BB BB BBB BBBB
B
BBB
BBBBB
BB
BBB
BBB
B
BB B BBBB
B B
BB BBB
B
B
B BB
B BB
BB BBBBBB
BBB
BB
B
BB
B
BBB
BBBBB
BB
BBB
BBB
BB
BB
B
BBBB
BB
BBBB
BBBB
B BB
BBB
B
BB
B
BB
BBBB
BBB
BB
BB
BBB
B
BB
B
B
BB
BB
BBB B B BB
B
BB
B
B
BBBB
BB
B
BB B B
B BB B B B
B
BBB B B
B BBBBBB
BBB
BBB B BB B
B
B BBBBBB BBB
BB
BBB
B
BB
B BBBBBB BBB
B
BBB B
B B BB
BBB
BB
B
BB B BB
B
B B BBBB
BB
BB BB
BBB
B BBBBBB BBB
B
BB
B
B
BBBB
BB
BB
BBB
BB
B BBBB
BBB
B
B BBBBBB BBB
BBB
BBB
BB
B BBBBBB BBB
BB B B
BBB
BB
B
BBBB
B
BB
BB B B
B BBBBBB BBB
B
B BBBB B BBB
BBB
BBB B
B
BBB
B
BB BBB BB
BB B B BB
B BBBBBB BBB
B
BB
B
B B
B B B B BBB B
BB
B
B BB
B BBBBBB BBB
BBB
B B
BBB BBBB
BB BB
BB B B B
B
B BBBBBB BBB
B
B
B
BBB
B B
BB
BB
BBB
BB
BB B BB
BB
B
B B
BB
BB
BBB
B
B B
B B
B
B BB
BBB
BB
B
BB
BBB B
B B B B B
BB B
B
BB B
BB
BBB
BB BBB
B
B BBBBBB BBB
B
B B
BBBBB
BBB
B BB
B
B BB BBB
BB
BBBB
BB
BB
BB
B B BBBBBB BBB
B
B B
B B
BBBBB
BBBB
BB
BB B
B
BB B BB BBBBB
BB
BBB
BB
B
B BBBBBB BBB
B
B B BBBBBB
BB B B
BB BB
B
B
BB
B BB
BB BBBBB
BB
BBB
BB
BB BB BBBBBB BBB
BB
BB
BB
B
BBB
BBB
BBB
B BB
BBB
B
BB
B
BBBB
B
BB
BB
B B
B
B
B
BB
BBB
BBBB
B
B B BB
B BBB
B
B
B
BBBB B B
B BBBBB
BB
B BB BBB
B
B BBBBBB BBB
B
BBB B
BB BB
BB
BBBB
B B
B BBBBBB BBB
B
B
BB
BB BBBBBBBB
B
BBB
BB
B BB
BB
B
BB BBBBBB BBB
BB
BB
B
BB B
B
BBBBBBBB
B B BB
BB BBBBBB BBB
BB
BB
B
BB BBB BB
BBB
B
B
B B
B
B
B B BBBB
BBB
BB
BB
BB
BBB BBBBBB BBB
B
B
BBB B
B
B BBBB
BB BBB B BB
BBB
B BBBBBB BBB
BBB
BBB
B
BBBB
B
BBBB BB
BBB
B
BB
BBBBB
BBBB BB
BBB BB
BB BB
BBB
BB
BBB
BB
B BB BBBBBB BBB
B
B B
B BB BB
B BBBB
BB
B BB
BB
BB
B
B BBBBB
BB
BBB
BBB BBBBBB BBB
B
BB
B
BB BBBBBBBB
BB
BB
BB BBBBBB BBB
B
B
B
B B
B BB B BB BBBBBBBBB
BBB BBBBBB BBB
B
B
B
B
B B
B B
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB B B B B
BB BBB B
B B B BB
B B BBB
B B B BBB B B B B
B BBB B B BBB
B
BB
B
B BB
BB
BB
B B B
BB
B BBB B
B BB
BB BB
B B BBB
B BB
BB BBB B BB
BB B
B BBB B BB
BBBB
B
BB
B
B
B
B
B B
B B
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
BBB
BBB
B
BBB
BBB
BBBBB
BBBBB
BB
BBB
BBBBB
BBBBB
BBB
BB
BBB
BB
BBBBB
BBBBB
BBB
BBB
BBBBB
BBBBB
BB
BBB
BB
BBB
=
0
50
100
150
200 250
300
20
70
120
+
ROW INDEX
ROW INDEX
120
BB B
BB
B
B
BBB
B
BB
B
BB BBBBBB
BBBB B B
BBBB
BB
BB
BB
BBBB
BBB
BB
BB
BB
BBB
BB
BBB
BBBB
B BB
BB
B
B
BBBB
BB
BB
BB
BBB
BBB
B
B
BB
B
B B
BB
B BBBB BB
B BB
BBBB B B B
BBB
B
BBB BB
BBB
BBB BB
BBB
BB
B
BBB
BBBB BB
B B BB
BB B
BB
B
BBBB
BB
BB
B
BBB
BBBBB
BB
BBB
BBB
BB B BBB
BB
BBB B BBB
BBBB
B B BB
BBB
BB B BBBB
BBBB
B BB B
B B BBBB
BB
BB BBB
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
BBBBBB
BB
B BB B
B BB
B BB
BB
B
BB
B
BB
BBB
B
BBB
BBBBB
BB
BBB
BBB
BBB
BBB
BB
B
BBB BB
B
B
BBB
BBBBB
BB
BBB
BBB
B BB BBBB BBB
BB
BBB
BBB
BB
BB B
B
BBB BBB
B B BB
BB
B
B
BBB
B
BBB
BBBBB
BB
BBB
BBB
BB
BB B
BB
BBB
BB
BBB
BBBBB
BB
BBB
BBB
B B BBB
B BB
BB
B
B
B
B BB
B B B
BB BBBB
B B
BB
BB
B
B
BB
B
BBB
BBBBB
BB
BBB
BBB
B BB
BB
BB
B
BBB
B
B
B BBB B
BB
BB
BB
BB
BB
B BB BB B
B
BB B
B
BBBB
B
BB
BB
BBBBB
BB
B
BB
BB
B
B
B
B BB
BB BB
B BB
B B B
B BB
B BBBB
BB
B B BBB
B B
BB
BBBB B
BBBB
BB
BBB
BBBBB
BBBB BB
BBB
BBBBB
BB
BBB
BBB
B
B BBBB
BB
BB BB BB
BB
BB BBB
BBBBB
BB
BBB
BBB
B BBBB
BB
BB
B B
BB
B
B BB BBBBBBBBB
BB
BBB
BBBBB
BB
BBB
BBB
B B BBBB
BB
B BB
BB
BBBB B
BBB
BBBBB
BB
BBB
BBB
B B B B
BBB
BBB
B
BB BB B BB BBBB
BB
BBBB
B
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
B
BB BBBB
BBB
BB
BB
B B
BB B B
BB
BBB
BBBBB
BB
BBB
BBB
B BBB
B
BBB B B
BB
B
B
B
BBB
BBBBB
BBBBB
BB
BBB
BBB
B
BB
BB B BB B BB
B
BB
BBB
B
BB
B
BBBB
BB
BBB
BBB
B B B
BB
BB
BB
BBB
BBB
BB
BB
BBBB
B
B
B
B
BB B
BB
BBBB
B BB B
BBB
BB
B
B BBB
BB
B
BB
BB B B
B BB B B B
B
B BB BB B
B BBBBBB
BB
B B BBB B B
B
BBB
BBBBB
BB
BBB
BBB
B
B BB BB
B B BB
BB
BBB
B B BB
B
BBB
BBBBB
BB
BBB
BBB
B B BBB BB B
B
B B BBB
BBB
BB
B
BB
BB
BBB
B
BB
BBB
BBBBB
BB
BBB
BBB
B
BB
B
B
BBBBB
BB
BB B
B B BB
B
B
BBB
BBBBB
BB
BBB
BBB
B BB
BB
BBB
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
B BBBBB BB
BB
BB
BB
BB
B
BB
BBBB
B
BBB B BBBB
BB
B B B BB
BBB
B
B
B BB
B B BB BBB BB
B
BBB
BBBBB
BB
BBB
BBB
B
B
B B BB BB
B B
B BB
BB
B BB
BB BB
BBB
B
BBB
BB BB
BB B B
B
B
BBB
BBBBB
BB
BBB
BBB
B
BBB
B BB
BB
BB
BBB
BB
BB BB BB
BB
BB
B
B BB
BB
BB
BBB
B
B
B
BB BB
BBB
BB
B
BB
BBBBB
B
B B
BB
BBBB B
BB BB
B
BB B
BB
BBBB
BB BB
B
BBB
BBBBB
BB
BBB
BBB
B
B
B BBBBB
BBB
BBB
BB
BB
BB
BB
BBB
B
BB
BBB
BBBBB
BB
BBB
BBB
B
B
B B
BBBBB
BBB
BB
BB
BBBB B
BBB
BBBBB
BB
BBB
BBB
B
B BB BB BB BBB BBBB
B
BBB
BBBBB
BB
BBB
BBB
B
BB B BBBB
B B
BB BBB
B
B
B BB
B BB
BB BBBBBB
BBB
BB
B
BB
B
BBB
BBBBB
BB
BBB
BBB
BB
BB
B
BBBB
BB
BBBB
BBBB
B BB
BBB
B
BB
B
BB
BBBB
BBB
BB
BB
BBB
B
BB
B
B
BB
BB
BBB B B BB
B
BB
B
B
BBBB
BB
B
BB B B
B BB B B B
B
BBB B B
B BBBBBB
BBB
BBB B BB B
B
B BBBBBB BBB
BB
BBB
B
BB
B BBBBBB BBB
B
BBB B
B B BB
BBB
BB
B
BB B BB
B
B B BBBB
BB
BB BB
BBB
B BBBBBB BBB
B
BB
B
B
BBBB
BB
BB
BBB
BB
B BBBB
BBB
B
B BBBBBB BBB
BBB
BBB
BB
B BBBBBB BBB
BB B B
BBB
BB
B
BBBB
B
BB
BB B B
B BBBBBB BBB
B
B BBBB B BBB
BBB
BBB B
B
BBB
B
BB BBB B
BB
B B B BB
B BBBBBB BBB
B
BB
B
B B
B B B B BBB B
BB
B
B BB
B BBBBBB BBB
BBB
B B
BBB BBBB
BB BB
BB B B B
B
B BBBBBB BBB
B
B
B
BBB
B B
BB
BB
BBB
BB
BB B BB
BB
B
B B
BB
BB
BBB
B
B B
B B
B
B BB
BBB
BB
B
BB
BBB B
B B B B B
BB B
B
BB B
BB
BBB
BB BBB
B
B BBBBBB BBB
B
B B
BBBBB
BBB
B BB
B
B BB BBB
BB
BBBB
BB
BB
BB
B B BBBBBB BBB
B
B B
B B
BBBBB
BBBB
BB
BB B
B
BB B BB BBBBB
BB
BBB
BB
B
B BBBBBB BBB
B
B B BBBBBB
BB B B
BB BB
B
B
BB
B BB
BB BBBBB
BB
BBB
BB
BB BB BBBBBB BBB
BB
BB
BB
B
BBB
BBB
BBB
B BB
BBB
B
BB
B
BBBB
B
BB
BB
B B
B
B
B
BB
BBB
BBBB
B
B B BB
B BBB
B
B
B
BBBB B B
B BBBBB
BB
B BB BBB
B
B BBBBBB BBB
B
BBB B
BB BB
BB
BBBB
B B
B BBBBBB BBB
B
B
BB
BB BBBBBBBB
B
BBB
BB
B BB
BB
B
BB BBBBBB BBB
BB
BB
B
BB B
B
BBBBBBBB
B B BB
BB BBBBBB BBB
BB
BB
B
BB BBB BB
BBB
B
B
B B
B
B
B B BBBB
BBB
BB
BB
BB
BBB BBBBBB BBB
B
B
BBB B
B
B BBBB
BB BBB B BB
BBB
B BBBBBB BBB
BBB
BBB
B
BBBB
B
BBBB BB
BBB
B
BB
BBBBB
BBBB BB
BBB BB
BB BB
BB
B
BB
B
BB
BB
B BB BBBBBB BBB
B
B B
B BB BB
B BBBB
BB
B BB
BB
BB
B
B BBBB
BB
BBB
B
BBB BBBBBB BBB
B
BB
B
BB BBBBBBB
BB
BB
B
BB BBBBBB BBB
B
B
B
B B
B BB B BB BBBBBBBB
BB
BB BBBBBB BBB
B
B
B
B
B B
B B
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB B B B B
BB BBB B
B B B BB
B B BBB
B B B BBB B B B B
B BBB B B BBB
B
BB
B
B BB
BB
BB
B B B
BB
B BBB B
B BB
BB BB
B B BBB
B BB
BB BBB B BB
BB B
B BBB B BB
BBBB
B
BB
B
B
B
B
B B
B B
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
BBB
BBB
B
BBB
BBB
BBBBB
BBBBB
BB
BBB
BBBBB
BBBBB
BBB
BB
BBB
BB
BBBBB
BBBBB
BBB
BBB
BBBBB
BBBBB
BB
BBB
BB
BBB
ROW INDEX
70
=170
220
270
320
BB B
BB
B
B
BBB
B
BB
B
BB BBBBBB
BBBB B B
BBBB
BB
BB
BB
BBBB
BBB
BB
BB
BB
BBB
BB
BBB
BBBB
B BB
BB
B
B
BBBB
BB
BB
BB
BBB
BBB
B
B
BB
B
B B
BB
B BBBB BB
B BB
BBBB B B B
BBB
B
BBB BB
BBB
BBB BB
BBB
BB
B
BBB
BBBB BB
B B BB
BB B
BB
B
BBBB
BB
BB
B
BBB
BBBBB
BB
BBB
BBB
BB B BBB
BB
BBB B BBB
BBBB
B B BB
BBB
BB B BBBB
BBBB
B BB B
B B BBBB
BB
BB BBB
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
BBBBBB
BB
B BB B
B BB
B BB
BB
B
BB
B
BB
BBB
B
BBB
BBBBB
BB
BBB
BBB
BBB
BBB
BB
B
BBB BB
B
B
BBB
BBBBB
BB
BBB
BBB
B BB BBBB BBB
BB
BBB
BBB
BB
BB B
B
BBB BBB
B B BB
BB
B
B
BBB
B
BBB
BBBBB
BB
BBB
BBB
BB
BB B
BB
BBB
BB
BBB
BBBBB
BB
BBB
BBB
B B BBB
B BB
BB
B
B
B
B BB
B B B
BB BBBB
B B
BB
BB
B
B
BB
B
BBB
BBBBB
BB
BBB
BBB
B BB
BB
BB
B
BBB
B
B
B BBB B
BB
BB
BB
BB
BB
B BB BB B
B
BB B
B
BBBB
B
BB
BB
BBBBB
BB
B
BB
BB
B
B
B
B BB
BB BB
B BB
B B B
B BB
B BBBB
BB
B B BBB
B B
BB
BBBB B
BBBB
BB
BBB
BBBBB
BBBB BB
BBB
BBBBB
BB
BBB
BBB
B
B BBBB
BB
BB BB BB
BB
BB BBB
BBBBB
BB
BBB
BBB
B BBBB
BB
BB
B B
BB
B
B BB BBBBBBBBB
BB
BBB
BBBBB
BB
BBB
BBB
B B BBBB
BB
B BB
BB
BBBB B
BBB
BBBBB
BB
BBB
BBB
B B B B
BBB
BBB
B
BB BB B BB BBBB
BB
BBBB
B
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
B
BB BBBB
BBB
BB
BB
B B
BB B B
BB
BBB
BBBBB
BB
BBB
BBB
B BBB
B
BBB B B
BB
B
B
B
BBB
BBBBB
BBBBB
BB
BBB
BBB
B
BB
BB B BB B BB
B
BB
BBB
B
BB
B
BBBB
BB
BBB
BBB
B B B
BB
BB
BB
BBB
BBB
BB
BB
BBBB
B
B
B
B
BB B
BB
BBBB
B BB B
BBB
BB
B
B BBB
BB
B
BB
BB B B
B BB B B B
B
B BB BB B
B BBBBBB
BB
B B BBB B B
B
BBB
BBBBB
BB
BBB
BBB
B
B BB BB
B B BB
BB
BBB
B B BB
B
BBB
BBBBB
BB
BBB
BBB
B B BBB BB B
B
B B BBB
BBB
BB
B
BB
BB
BBB
B
BB
BBB
BBBBB
BB
BBB
BBB
B
BB
B
B
BBBBB
BB
BB B
B B BB
B
B
BBB
BBBBB
BB
BBB
BBB
B BB
BB
BBB
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
B BBBBB BB
BB
BB
BB
BB
B
BB
BBBB
B
BBB B BBBB
BB
B B B BB
BBB
B
B
B BB
B B BB BBB BB
B
BBB
BBBBB
BB
BBB
BBB
B
B
B B BB BB
B B
B BB
BB
B BB
BB BB
BBB
B
BBB
BB BB
BB B B
B
B
BBB
BBBBB
BB
BBB
BBB
B
BBB
B BB
BB
BB
BBB
BB
BB BB BB
BB
BB
B
B BB
BB
BB
BBB
B
B
B
BB BB
BBB
BB
B
BB
BBBBB
B
B B
BB
BBBB B
BB BB
B
BB B
BB
BBBB
BB BB
B
BBB
BBBBB
BB
BBB
BBB
B
B
B BBBBB
BBB
BBB
BB
BB
BB
BB
BBB
B
BB
BBB
BBBBB
BB
BBB
BBB
B
B
B B
BBBBB
BBB
BB
BB
BBBB B
BBB
BBBBB
BB
BBB
BBB
B
B BB BB BB BBB BBBB
B
BBB
BBBBB
BB
BBB
BBB
B
BB B BBBB
B B
BB BBB
B
B
B BB
B BB
BB BBBBBB
BBB
BB
B
BB
B
BBB
BBBBB
BB
BBB
BBB
BB
BB
B
BBBB
BB
BBBB
BBBB
B BB
BBB
B
BB
B
BB
BBBB
BBB
BB
BB
BBB
B
BB
B
B
BB
BB
BBB B B BB
B
BB
B
B
BBBB
BB
B
BB B B
B BB B B B
B
BBB B B
B BBBBBB
BBB
BBB B BB B
B
B BBBBBB BBB
BB
BBB
B
BB
B BBBBBB BBB
B
BBB B
B B BB
BBB
BB
B
BB B BB
B
B B BBBB
BB
BB BB
BBB
B BBBBBB BBB
B
BB
B
B
BBBB
BB
BB
BBB
BB
B BBBB
BBB
B
B BBBBBB BBB
BBB
BBB
BB
B BBBBBB BBB
BB B B
BBB
BB
B
BBBB
B
BB
BB B B
B BBBBBB BBB
B
B BBBB B BBB
BBB
BBB B
B
BBB
B
BB BBB B
BB
B B B BB
B BBBBBB BBB
B
BB
B
B B
B B B B BBB B
BB
B
B BB
B BBBBBB BBB
BBB
B B
BBB BBBB
BB BB
BB B B B
B
B BBBBBB BBB
B
B
B
BBB
B B
BB
BB
BBB
BB
BB B BB
BB
B
B B
BB
BB
BBB
B
B B
B B
B
B BB
BBB
BB
B
BB
BBB B
B B B B B
BB B
B
BB B
BB
BBB
BB BBB
B
B BBBBBB BBB
B
B B
BBBBB
BBB
B BB
B
B BB BBB
BB
BBBB
BB
BB
BB
B B BBBBBB BBB
B
B B
B B
BBBBB
BBBB
BB
BB B
B
BB B BB BBBBB
BB
BBB
BB
B
B BBBBBB BBB
B
B B BBBBBB
BB B B
BB BB
B
B
BB
B BB
BB BBBBB
BB
BBB
BB
BB BB BBBBBB BBB
BB
BB
BB
B
BBB
BBB
BBB
B BB
BBB
B
BB
B
BBBB
B
BB
BB
B B
B
B
B
BB
BBB
BBBB
B
B B BB
B BBB
B
B
B
BBBB B B
B BBBBB
BB
B BB BBB
B
B BBBBBB BBB
B
BBB B
BB BB
BB
BBBB
B B
B BBBBBB BBB
B
B
BB
BB BBBBBBBB
B
BBB
BB
B BB
BB
B
BB BBBBBB BBB
BB
BB
B
BB B
B
BBBBBBBB
B B BB
BB BBBBBB BBB
BB
BB
B
BB BBB BB
BBB
B
B
B B
B
B
B B BBBBB
BBB
B
BB
BB
BBB BBBBBB BBB
B
B
BBB B
B
B BBBB
BB BBB B BB
BBB
B BBBBBB BBB
BBB
BBB
B
BBBB
B
BBBB BB
BBB
B
BB
BBBBB
BBBB BB
BBB BB
BB BB
BB
B
BB
B
BB
BB
B BB BBBBBB BBB
B
B B
B BB BB
B BBBB
BB
B BB
BB
BB
B
B BBBB
BB
BBB
B
BBB BBBBBB BBB
B
BB
B
BB BBBBBBB
BB
BB
B
BB BBBBBB BBB
B
B
B
B B
B BB B BB BBBBBBBB
BB
BB BBBBBB BBB
B
B
B
B
B B
B B
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB B B B B
BB BBB B
B B B BB
B B BBB
B B B BBB B B B B
B BBB B B BBB
B
BB
B
B BB
BB
BB
B B B
BB
B BBB B
B BB
BB BB
B B BBB
B BB
BB BBB B BB
BB B
B BBB B BB
BBBB
B
BB
B
B
B
B
B B
B B
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
BBB
BBB
B
BBB
BBB
BBBBB
BBBBB
BB
BBB
BBBBB
BBBBB
BBB
BB
BBB
BB
BBBBB
BBBBB
BBB
BBB
BBBBB
BBBBB
BB
BBB
BB
BBB
0
50
100
150
200 250
300
COLU MN I NDEX
20
70
120
ROW INDEX
COLU MN I NDEX
20
Sigma/lambda**2 (dB)
25
170
220
270
320
BB B
BB
B
B
BBB
B
BB
B
BB BBBBBB
BBBB B B
BBBB
BB
BB
BB
BBBB
BBB
BB
BB
BB
BBB
BB
BBB
BBBB
B BB
BB
B
B
BBBB
BB
BB
BB
BBB
BBB
B
B
BB
B
B B
BB
B BBBB BB
B BB
BBBB B B B
BBB
B
BBB BB
BBB
BBB BB
BBB
BB
B
BBB
BBBB BB
B B BB
BB B
BB
B
BBBB
BB
BB
B
BBB
BBBBB
BB
BBB
BBB
BB B BBB
BB
BBB B BBB
BBBB
B B BB
BBB
BB B BBBB
BBBB
B BB B
B B BBBB
BB
BB BBB
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
BBBBBB
BB
B BB B
B BB
B BB
BB
B
BB
B
BB
BBB
B
BBB
BBBBB
BB
BBB
BBB
BBB
BBB
BB
B
BBB BB
B
B
BBB
BBBBB
BB
BBB
BBB
B BB BBBB BBB
BB
BBB
BBB
BB
BB B
B
BBB BBB
B B BB
BB
B
B
BBB
B
BBB
BBBBB
BB
BBB
BBB
BB
BB B
BB
BBB
BB
BBB
BBBBB
BB
BBB
BBB
B B BBB
B BB
BB
B
B
B
B BB
B B B
BB BBBB
B B
BB
BB
B
B
BB
B
BBB
BBBBB
BB
BBB
BBB
B BB
BB
BB
B
BBB
B
B
B BBB B
BB
BB
BB
BB
BB
B BB BB B
B
BB B
B
BBBB
B
BB
BB
BBBBB
BB
B
BB
BB
B
B
B
B BB
BB BB
B BB
B B B
B BB
B BBBB
BB
B B BBB
B B
BB
BBBB B
BBBB
BB
BBB
BBBBB
BBBB BB
BBB
BBBBB
BB
BBB
BBB
B
B BBBB
BB
BB BB BB
BB
BB BBB
BBBBB
BB
BBB
BBB
B BBBB
BB
BB
B B
BB
B
B BB BBBBBBBBB
BB
BBB
BBBBB
BB
BBB
BBB
B B BBBB
BB
B BB
BB
BBBB B
BBB
BBBBB
BB
BBB
BBB
B B B B
BBB
BBB
B
BB BB B BB BBBB
BB
BBBB
B
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
B
BB BBBB
BBB
BB
BB
B B
BB B B
BB
BBB
BBBBB
BB
BBB
BBB
B BBB
B
BBB B B
BB
B
B
B
BBB
BBBBB
BBBBB
BB
BBB
BBB
B
BB
BB B BB B BB
B
BB
BBB
B
BB
B
BBBB
BB
BBB
BBB
B B B
BB
BB
BB
BBB
BBB
BB
BB
BBBB
B
B
B
B
BB B
BB
BBBB
B BB B
BBB
BB
B
B BBB
BB
B
BB
BB B B
B BB B B B
B
B BB BB B
B BBBBBB
BB
B B BBB B B
B
BBB
BBBBB
BB
BBB
BBB
B
B BB BB
B B BB
BB
BBB
B B BB
B
BBB
BBBBB
BB
BBB
BBB
B B BBB BB B
B
B B BBB
BBB
BB
B
BB
BB
BBB
B
BB
BBB
BBBBB
BB
BBB
BBB
B
BB
B
B
BBBBB
BB
BB B
B B BB
B
B
BBB
BBBBB
BB
BBB
BBB
B BB
BB
BBB
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
B BBBBB BB
BB
BB
BB
BB
B
BB
BBBB
B
BBB B BBBB
BB
B B B BB
BBB
B
B
B BB
B B BB BBB BB
B
BBB
BBBBB
BB
BBB
BBB
B
B
B B BB BB
B B
B BB
BB
B BB
BB BB
BBB
B
BBB
BB BB
BB B B
B
B
BBB
BBBBB
BB
BBB
BBB
B
BBB
B BB
BB
BB
BBB
BB
BB BB BB
BB
BB
B
B BB
BB
BB
BBB
B
B
B
BB BB
BBB
BB
B
BB
BBBBB
B
B B
BB
BBBB B
BB BB
B
BB B
BB
BBBB
BB BB
B
BBB
BBBBB
BB
BBB
BBB
B
B
B BBBBB
BBB
BBB
BB
BB
BB
BB
BBB
B
BB
BBB
BBBBB
BB
BBB
BBB
B
B
B B
BBBBB
BBB
BB
BB
BBBB B
BBB
BBBBB
BB
BBB
BBB
B
B BB BB BB BBB BBBB
B
BBB
BBBBB
BB
BBB
BBB
B
BB B BBBB
B B
BB BBB
B
B
B BB
B BB
BB BBBBBB
BBB
BB
B
BB
B
BBB
BBBBB
BB
BBB
BBB
BB
BB
B
BBBB
BB
BBBB
BBBB
B BB
BBB
B
BB
B
BB
BBBB
BBB
BB
BB
BBB
B
BB
B
B
BB
BB
BBB B B BB
B
BB
B
B
BBBB
BB
B
BB B B
B BB B B B
B
BBB B B
B BBBBBB
BBB
BBB B BB B
B
B BBBBBB BBB
BB
BBB
B
BB
B BBBBBB BBB
B
BBB B
B B BB
BBB
BB
B
BB B BB
B
B B BBBB
BB
BB BB
BBB
B BBBBBB BBB
B
BB
B
B
BBBB
BB
BB
BBB
BB
B BBBB
BBB
B
B BBBBBB BBB
BBB
BBB
BB
B BBBBBB BBB
BB B B
BBB
BB
B
BBBB
B
BB
BB B B
B BBBBBB BBB
B
B BBBB B BBB
BBB
BBB B
B
BBB
B
BB BBB B
BB
B B B BB
B BBBBBB BBB
B
BB
B
B B
B B B B BBB B
BB
B
B BB
B BBBBBB BBB
BBB
B B
BBB BBBB
BB BB
BB B B B
B
B BBBBBB BBB
B
B
B
BBB
B B
BB
BB
BBB
BB
BB B BB
BB
B
B B
BB
BB
BBB
B
B B
B B
B
B BB
BBB
BB
B
BB
BBB B
B B B B B
BB B
B
BB B
BB
BBB
BB BBB
B
B BBBBBB BBB
B
B B
BBBBB
BBB
B BB
B
B BB BBB
BB
BBBB
BB
BB
BB
B B BBBBBB BBB
B
B B
B B
BBBBB
BBBB
BB
BB B
B
BB B BB BBBBB
BB
BBB
BB
B
B BBBBBB BBB
B
B B BBBBBB
BB B B
BB BB
B
B
BB
B BB
BB BBBBB
BB
BBB
BB
BB BB BBBBBB BBB
BB
BB
BB
B
BBB
BBB
BBB
B BB
BBB
B
BB
B
BBBB
B
BB
BB
B B
B
B
B
BB
BBB
BBBB
B
B B BB
B BBB
B
B
B
BBBB B B
B BBBBB
BB
B BB BBB
B
B BBBBBB BBB
B
BBB B
BB BB
BB
BBBB
B B
B BBBBBB BBB
B
B
BB
BB BBBBBBBB
B
BBB
BB
B BB
BB
B
BB BBBBBB BBB
BB
BB
B
BB B
B
BBBBBBBB
B B BB
BB BBBBBB BBB
BB
BB
B
BB BBB BB
BBB
B
B
B B
B
B
B B BBBB
BBB
BB
BB
BB
BBB BBBBBB BBB
B
B
BBB B
B
B BBBB
BB BBB B BB
BBB
B BBBBBB BBB
BBB
BBB
B
BBBB
B
BBBB BB
BBB
B
BB
BBBBB
BBBB BB
BBB BB
BB BB
BBB
BB
BBB
BB
B BB BBBBBB BBB
B
B B
B BB BB
B BBBB
BB
B BB
BB
BB
B
B BBBBB
BB
BBB
BBB BBBBBB BBB
B
BB
B
BB BBBBBBB
BB
BB
B
BB BBBBBB BBB
B
B
B
B B
B BB B BB BBBBBBBB
BB
BB BBBBBB BBB
B
B
B
B
B B
B B
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB B B B B
BB BBB B
B B B BB
B B BBB
B B B BBB B B B B
B BBB B B BBB
B
BB
B
B BB
BB
BB
B B B
BB
B BBB B
B BB
BB BB
B B BBB
B BB
BB BBB B BB
BB B
B BBB B BB
BBBB
B
BB
B
B
B
B
B B
B B
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
BBB
BBB
B
BBB
BBB
BBBBB
BBBBB
BB
BBB
BBBBB
BBBBB
BBB
BB
BBB
BB
BBBBB
BBBBB
BBB
BBB
BBBBB
BBBBB
BB
BBB
BB
BBB
=
0
+
50
100
150
200 250
=
300
20
15
10
5
0
-5
-10
0
70
ROW INDEX
120
170
220
270
320
BB B
BB
B
B
BBB
B
BB
B
BB BBBBBB
BBBB B B
BBBB
BB
BB
BB
BBBB
BBB
BB
BB
BB
BBB
BB
BBB
BBBB
B BB
BB
B
B
BBBB
BB
BB
BB
BBB
BBB
B
B
BB
B
B B
BB
B BBBB BB
B BB
BBBB B B B
BBB
B
BBB BB
BBB
BBB BB
BBB
BB
B
BBB
BBBB BB
B B BB
BB B
BB
B
BBBB
BB
BB
B
BBB
BBBBB
BB
BBB
BBB
BB B BBB
BB
BBB B BBB
BBBB
B B BB
BBB
BB B BBBB
BBBB
B BB B
B B BBBB
BB
BB BBB
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
BBBBBB
BB
B BB B
B BB
B BB
BB
B
BB
B
BB
BBB
B
BBB
BBBBB
BB
BBB
BBB
BBB
BBB
BB
B
BBB BB
B
B
BBB
BBBBB
BB
BBB
BBB
B BB BBBB BBB
BB
BBB
BBB
BB
BB B
B
BBB BBB
B B BB
BB
B
B
BBB
B
BBB
BBBBB
BB
BBB
BBB
BB
BB B
BB
BBB
BB
BBB
BBBBB
BB
BBB
BBB
B B BBB
B BB
BB
B
B
B
B BB
B B B
BB BBBB
B B
BB
BB
B
B
BB
B
BBB
BBBBB
BB
BBB
BBB
B BB
BB
BB
B
BBB
B
B
B BBB B
BB
BB
BB
BB
BB
B BB BB B
B
BB B
B
BBBB
B
BB
BB
BBBBB
BB
B
BB
BB
B
B
B
B BB
BB BB
B BB
B B B
B BB
B BBBB
BB
B B BBB
B B
BB
BBBB B
BBBB
BBB
BBBBB
BBBB BB
BBB
BBBBB
BB
BBB
BBB
B
B BBBB
BB
BB BB BB
BB
BB BBB
BBBBB
BB
BBB
BBB
B BBBB
BB
BB
B B
BB
B
B BB BBBBBBBBB
BB
BBB
BBBBB
BB
BBB
BBB
B B BBBB
BB
B BB
BB
BBBB B
BBB
BBBBB
BB
BBB
BBB
B B B B
BBB
BBB
B
BB BB B BB BBBB
BB
BBBB
B
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
B
BB BBBBB
BBB
B
BB
B B
BB B B
BB
BBB
BBBBB
BB
BBB
BBB
B BBB
B
BBB B B
BB
B
B
B
BBB
BBBBB
BBBBB
BB
BBB
BBB
B
BB
BB B BB B BB
B
BB
BBB
B
BB
B
BBBB
BB
BBB
BBB
B B B
BB
BB
BB
BBB
BBB
BB
BB
BBBB
B
B
B
B
BB B
BB
BBBB
B BB B
BBB
BB
B
B BBB
BB
B
BB
BB B B
B BB B B B
B
B BB BB B
B BBBBBB
BB
B B BBB B B
B
BBB
BBBBB
BB
BBB
BBB
B
B BB BB
B B BB
BB
BBB
B B BB
B
BBB
BBBBB
BB
BBB
BBB
B B BBB BB B
B
B B BBB
BBB
BB
B
BB
BB
BBB
B
BB
BBB
BBBBB
BB
BBB
BBB
B
BB
B
B
BBBBB
BB
BB B
B B BB
B
B
BBB
BBBBB
BB
BBB
BBB
B BB
BB
BBB
BB
BB
BBB
BBBBB
BB
BBB
BBB
B
B BBBBB BB
BB
BB
BB
BB
B
BB
BBBB
B
BBB B BBBB
BB
B B B BB
BBB
B
B
B BB
B B BB BBB BB
B
BBB
BBBBB
BB
BBB
BBB
B
B
B B BB BB
B B
B BB
BB
B BB
BB BB
BBB
BBB
BB BB
BB B B
B
B
BBB
BBBBB
BB
BBB
BBB
B
BBB
B BB
BB
BB
BBB
BB
BB BB BB
BB
BB
B
B BB
BBB
BBB
B
B
B
BB BB
BBB
BB
B
BB
BBBBB
B
B B
BB
BBBB B
BB BB
B
BB B
BB
BBBB
BB BB
B
BBB
BBBBB
BB
BBB
BBB
B
B
B BBBBB
BBB
BBB
BB
BB
BB
BB
BBB
B
BB
BBB
BBBBB
BB
BBB
BBB
B
B
B B
BBBBB
BBB
BB
BB
BBBB B
BBB
BBBBB
BB
BBB
BBB
B
B BB BB BB BBB BBBB
B
BBB
BBBBB
BB
BBB
BBB
B
BB B BBBB
B B
BB BBB
B
B
B BB
B BB
BB BBBBBB
BBB
BB
BB
B
BBB
BBBBB
BB
BBB
BBB
BB
BB
B
BBBB
BB
BBBB
BBBB
B BB
BBB
B
BB
B
BB
BBBB
BBB
BB
BB
BBB
B
BB
B
B
BB
BB
BBB B B BB
B
BB
B
B
BBBB
BB
B
BB B B
B BB B B B
B
BBB B B
B BBBBBB
BBB
BBB B BB B
B
B BBBBBB BBB
BB
BBB
B
BB
B BBBBBB BBB
B
BBB B
B B BB
BBB
BB
B
BB B BB
B
B B BBBB
BB
BB BB
BBB
B BBBBBB BBB
B
BB
B
B
BBBBB
BB
BBB
BB
B BBBB
BBB
B
B BBBBBB BBB
BBB
BBB
BB
B BBBBBB BBB
BB B B
BBB
BB
B
BBBB
B
BB
BB B B
B BBBBBB BBB
B
B BBBB B BBB
BBB
BBB B
B
BBB
B
BB BBB B
BB
B B B BB
B BBBBBB BBB
B
BB
B
B B
B B B B BBB B
BB
B
B BB
B BBBBBB BBB
BBB
B B
BBB BBBB
BB BB
BB B B B
B
B BBBBBB BBB
B
B
B
BBB
B B
BB
BB
BBB
BB
BB B BB
BB
B
B B
BBB
BBB
B
B B
B B
B
B BB
BBB
BB
B
BB
BBB B
B B B B B
BB B
B
BB B
BB
BBB
BB BBB
B
B BBBBBB BBB
B
B B
BBBBB
BBB
B BB
B
B BB BBB
BB
BBBB
BB
BB
BB
B B BBBBBB BBB
B
B B
B B
BBBBB
BBBB
BB
BB B
B
BB B BB BBBBB
BB
BBB
BB
B
B BBBBBB BBB
B
B B BBBBBB
BB B B
BB BB
B
B
BB
B BB
BB BBBBBBBB
BB BB BBBBBB BBB
BB
BB
BB
B
BBB
BBB
BBB
B BB
BBB
B
BB
B
BBBB
B
BB
BB
B B
B
B
B
BB
BBB
BBBB
B
B B BB
B BBB
B
B
B
BBBB B B
B BBBBB
BB
B BB BBB
B
B BBBBBB BBB
B
BBB B
BB BB
BB
BBBB
B B
B BBBBBB BBB
B
B
BB
BB BBBBBBBB
B
BBB
BB
B BB
BB
B
BB BBBBBB BBB
BB
BB
B
BB B
B
BBBBBBBB
B B BB
BB BBBBBB BBB
BB
BB
B
BB BBB BB
BBB
B
B
B B
B
B
B B BBBBB
BBB
B
BB
BB
BBB BBBBBB BBB
B
B
BBB B
B
B BBBB
BB BBB B B
BB
BB
B BBBBBB BBB
BBB
BBB
B
BBBB
B
BBBB BB
BBB
B
BB
BB
BBBB
B B BB
BBB BB
BB BB
BBBBB BB BBBBBB BBB
B
B B
B BB BB
B BBBB
BB
B BB
BB
BB
B
B BBBBB
BB
BBB
BBB BBBBBB BBB
B
BB
B
BB BBBBBBB
BB
BB
B
BB BBBBBB BBB
B
B
B
B B
B BB B BB B BBBBBB
BB
BB BBBBBB BBB
B
B
B
B
B B
B B
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
B
B
B
B
B
B B
B B
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB B
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BBB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
B BBB BBB B BBB B
B BBB B BBBBB
BB
BB BBBB B B
BB BBB B
B BBB BB
B B BBB
BBB
BBB
B
BBB
BBB
BBBBB
BBBBB
BB
BBB
BBBBB
BBBBB
BBB
BB
BBB
BB
BBBBB
BBBBB
BBB
BBB
BBBBB
BBBBB
BB
BBB
BB
BBB
0
50
100
150
200
250
300
=
SYSTEM OF EQUATIONS
 K
C
0 H   0 

  

C†
0
Z0 M   0 

  

 0

V 
Z
Z
J




 inc 
M
J 
H  K 1CM
Z0 M  0 
   
Z J J  V 
This matrix problem is filled and solved by PHOEBUS
–
–
–
60
P_SLICE
P_SOLVE
70
80 100 120 140 160 180
THETA (deg)
P_FIELD
256 PEs
60
50
256 PEs
40
30
 † 1
C K C

 Z

M
40
Dielectric Cylinder
r = 1.0 cm, h =10.0 cm, eps = 4.0, 2.5 GHz
Time (min)
COLU MN I NDEX
20
20
128 PEs
256 PEs
100694
271158
417359
Size of problem (edges)
20
10
0
579993
The K submatrix is a sparse finite element matrix
The Z submatrices are integral equation matrices
The C submatrices are coupling matrices between the FE and IE equations
1996! - 3 Executable, 2+ programming models, executables run sequentially
Credit: Katz, Cwik, Zuffada, Jamnejad
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
9
Cholesky Factorization
• SuperMatrix work - Chan and van de Geijn, Univ. of
Texas, in progress
• Based on FLAME library
• Aimed at NUMA systems, OpenMP programming model
• Initial realization: poor performance of LAPACK
(w/ multithreaded BLAS)
could be fixed by choosing
a different variant
Credit: Ernie Chan
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
10
Cholesky Factorization
Iteration 1
Iteration 2
Iteration 3
Chol
Trsm
Syrk
Trsm
Gemm
Chol
Syrk
Trsm
Chol
Syrk
Chol
• Can represent as DAG
Trsm
Syrk
…
Credit: Ernie Chan
Gemm
…
Syrk
Trsm
Chol
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
11
Cholesky SuperMatrix
• Execute DAG tasks in parallel, possibly “out-oforder”
– Similar in concept
to Tomasulo’s
algorithm and
instruction-level
parallelism on
blocks of
computation
• Superscalar ->
SuperMatrix
Credit: Ernie Chan
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
12
Uintah Framework
• de St. German, McCorquedale, Parker, Johnson
at SCI Institute, Univ. of Utah
• Based on task graph model
– Each algorithm define a description of computation
• Required inputs and outputs
• Callbacks to perform a task on a single region of space
– Communication performed at graph edges
– Graph created by Uintah
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
13
Uintah Tensor Product Task Graph
•
•
•
•
Each task is replicated over regions in space
Expresses data parallelism and task parallelism
Resulting detailed graph is tensor product of
master graph and spatial regions
Efficient:
– Detailed tasks not replicated on all processors
•
Scalable:
– Control structure known globally
– Communication structure known locally
•
Dependencies specified implicitly w/ simple algebra
Master Graph
(explicitly defined)
– Spatial dependencies
• Computes:
– Variable (name, type)
– Patch subset
• Requires:
QuickTime™ and a
TIFF (LZW) decompressor
are needed to see this picture.
– Variable (name, type)
– Patch subset
– Halo specification
– Other dependencies: AMR, others
Credit: Steve Parker
Detailed Graph (implicitly defined)
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
14
Uintah - How It Works
Credit: Steve Parker
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
15
Uintah - More Details
• Task graphs can be complex
– Can include loops, nesting, recursion
• Optimal scheduling is NP-hard
– “Optimal enough” scheduling isn’t
too hard
• Creating schedule can be
expensive
– But may not be done too often
• Overall, good
scaling and
performance
has been
obtained with
this approach
Credit: Steve Parker
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
16
Applications and Grids
• How to map applications to grids?
• Some applications are Grid-unaware - they just want to
run fast
– May run on Grid-aware (Grid-enabled?) programming
environments, e.g. MPICH-G2, MPIg
• Other apps are Grid-aware themselves
– This is where SAGA fits in, as an API to permit the apps to
interact with the middleware
Grid-unaware applications
Grid-aware applications
Grid-enabled tools/environments
Simple API (SAGA)
Middleware
Grid resources, services, platforms
Credit: Thilo Kielmann
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
17
Common Grid Applications
• Data processing
– Data exists on the grid, possibly replicated
– Data is staged to a single set of resources
– Application starts on that set of resources
• Parameter sweeps
– Lots of copies of a sequential/parallel job launched on
independent resources, with different inputs
– Controlling process start jobs and gathers outputs
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
18
More Common Grid Applications
• Workflow applications
– Multiple units of work, either sequential or parallel,
either small or large
– Data often transferred between tasks by files
– Task sequence described as a graph, possibly a DAG
– Abstract graph doesn’t include resource information
– Concrete graph does
– Some process/service converts graph from abstract to
concrete
• Often all at once, ahead of job start - static mapping
• Perhaps more gradually (JIT?) - dynamic mapping
• Pegasus from ISI is an example of this, currently static
– (Note: Parameter sweeps are very simple workflows)
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
19
Montage - a Workflow App
•
•
•
An astronomical image mosaic service
for the National Virtual Observatory
http://montage.ipac.caltech.edu/
Delivers custom, science grade image mosaics
–
–
–
•
Image mosaic: combine many images so
that they appear to be a single image from a
single telescope or spacecraft
User specifies projection, coordinates,
spatial sampling, mosaic size, image rotation
Preserve astrometry (to 0.1 pixels) & flux (to 0.1%)
100 µm sky; aggregation of COBE and IRAS maps (Schlegel, Finkbeiner and
Davis, 1998). Covers 360 x 180 degrees in CAR projection.
Modular, portable “toolbox” design
–
–
Loosely-coupled engines
Each engine is an executable compiled from ANSI C
Supernova remnant S147, from IPHAS: The INT/WFC Photometric
H-alpha Survey of the Northern Galactic Plane
David Hockney Pearblossom Highway 1986
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
20
Montage Workflow
mProject 1
mDiff 1 2
D12
3
2
1
mProject 2
Final Mosaic
(Overlapping Tiles)
mAdd 2
mAdd 1
mProject 3
mBackground 1
mDiff 2 3
D23
mBackground 2
mBackground 3
a1x + b1y + c1 = 0
a2x + b2y + c2 = 0
a3x + b3y + c3 = 0
mFitplane D12
mFitplane D23
mBgModel
ax + by + c = 0
dx + ey + f = 0
mConcatFit
ax + by + c = 0
dx + ey + f = 0
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
21
Montage on the Grid Using Pegasus
(Planning for Execution on Grids)
Example DAG for 10 input files
Maps an abstract workflow
to an executable form
mProject
Pegasus
mDiff
http://pegasus.isi.edu/
mFitPlane
mConcatFit
Grid Information
Systems
mBgModel
mBackground
Information about
available resources,
data location
mAdd
Condor DAGMan
Data Stage-in nodes
Executes the workflow
Montage compute nodes
Data stage-out nodes
Registration nodes
MyProxy
Grid
User’s grid credentials
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
22
Montage Performance
• MPI version on a single cluster is baseline
• Grid version on a single cluster has similar
performance for large problems
• Grid version on multiple clusters has
performance dominated by data transfer
between stages
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
23
Workflow Application Issues
• Apps need to map processing to clusters
• Depending on mapping, various data movement is
needed, so the mapping either leads to networking
requirements or is dependent on the available
networking
• Prediction (and mapping) needs some intelligence
• One way to do this is through Pegasus, which currently
does static mapping of an abstract workflow to a
concrete workflow, but will do more dynamic mapping at
some future point
• Networking resources and availability could be inputs to
Pegasus, or Pegasus could be used to request network
resources at various times during a run.
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
24
Making Use of Grids
• In general, groups of users (communities) want to run
applications
• Code/User/Infrastructure is aware of environment and
does:
–
–
–
–
Discover resources available now (or perhaps later)
Start my application
Have access to data and storage
Monitor and possibly steer the application
• Other things that could be done:
– Migrate app to faster resources that are now available
– Recover from hardware failure by continuing with fewer
processors or by restarting from checkpoint on different
resources
– Use networks as needed (reserve them for these times)
Credit: Thilo Kielmann and Gabrielle Allen
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
25
Less Common Grid Applications
• True distributed MPI application over multiple
resources/clusters
• Other applications that use multiple coupled
clusters
• Uncommon because these jobs run poorly
without sufficient network bandwidth, and there
has been no good way for users to reserve
bandwidth when needed
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
26
SPICE
• Used for analyzing RNA translocation through protein pores
• Using “standard” molecular dynamics would need millions of CPU hours
• Instead, use Steered Molecular Dynamics and Jarzynski’s Equation
(SMD-JE)
– Uses static visualization to understand structural features
– Uses interactive simulations to determine
“near-optimal” parameters
• Uses Haptic interaction - requires low-latency
bi-directional communication between user
and simulation
– Uses “near-optimal” parameters and many
large parallel simulations to determine
“optimal” parameters
• ~75 simulations on 128/256 processors
– Uses “optimal” parameters to calculate full
free energy profile along axis of pore
• ~100 simulations on 2500 processors
Credit: Shantenu Jha, et. al.
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
27
NEKTAR
• Simulates arterial blood flow
• Uses hybrid approach
– 3D detailed CFD computed at bifurcations
– Waveform coupling between bifurcations modeled
w/ reduced set of 1D equations
– 55 largest arteries in human body w/ 27 bifurcations
would require about 7 TB memory
– Parallelized across and within clusters
NCSAPSC
SDSC TACC
Credit: Shantenu Jha, et. al.
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
28
Cactus
• Freely available, modular, portable and manageable environment for
collaboratively developing parallel, high-performance multidimensional simulations (components-based)
• Developed for numerical relativity, but now general framework for
parallel computing (CFD, astro, climate, chem. eng., quantum
gravity, etc.)
• Finite difference, AMR, FE/FV, multipatch
• Active user and developer communities, main development now at
LSU and AEI
• Science-driven design issues
• Open source, documentation, etc.
• Just over 10 years old
Credit: Gabrielle Allen
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
29
Cactus Structure
remote steering
Plug-In “Thorns”
(modules)
extensible APIs
ANSI C
parameters
driver
scheduling
equations of state
Core “Flesh”
input/output
error handling
interpolation
SOR solver
Fortran/C/C++
black holes
make system
grid variables
wave evolvers
multigrid
boundary conditions
coordinates
Credit: Gabrielle Allen
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
30
Cactus and Grids
• HTTPD thorn, allows web browser to connect to running
simulation, examine state of running simulation, change
parameters
• Worm thorn, makes Cactus app self-migrating
• Spawner thorn, any routine can be done on another
resource
• TaskFarm, allows distributing of apps on Grid
• Run a single app using distributed MPI
Credit: Gabrielle Allen, Erik Schnetter
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
31
EnLIGHTened
•
•
•
•
Network research, driven by concrete application projects, all of which
critically require progress in network technologies and tools that utilize them
EnLIGHTened testbed: 10 Gbps optical networks running over NLR. Four
all-photonic Calient switches are interconnected via Louisiana Optical
Network Initiative (LONI), EnLIGHTened wave, and the Ultralight wave, all
using GMPLS control plane technologies.
Global alliance of partners
Will develop, test, and disseminate advanced software and underlying
technologies to:
– Provide generic applications with the ability to be aware of their network, Grid
environment and capabilities, and to make dynamic, adaptive and optimized use
(monitor & abstract, request & control) of networks connecting various high end
resources
– Provide vertical integration from the application to the optical control plane,
including extending GMPLS
•
Will examine how to distribute the network intelligence among the network
control plane, management plane, and the Grid middleware
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
32
EnLIGHTened Team
•
•
•
•
•
•
•
•
•
•
Yufeng Xin
Steve Thorpe
Bonnie Hurst
Joel Dunn
Gigi Karmous-Edwards
Mark Johnson
John Moore
Carla Hunt
Lina Battestilli
Andrew Mabe
•
•
•
•
•
•
•
•
•
•
Ed Seidel
Gabrielle Allen
Seung Jong Park
Jon MacLaren
Andrei Hutanu
Lonnie Leger
Dan Katz
Savera Tanwir
Harry Perros
Mladen Vouk
•
•
Olivier Jerphagnon
John Bowers
•Steven Hunter
•Dan Reed
•Alan Blatecky
•Chris Heermann
•
•
•
•
•
Javad Boroumand
Russ Gyurek
Wayne Clark
Kevin McGrattan
Peter Tompsu
•Rick Schlichting
•John Strand
•Matti Hiltunen
•Yang Xia
•Xun Su
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
33
EnLIGHTened Testbed
To Canada
To Asia
To Europe
SEA
POR
BOI
CAVE wave
OGD
EnLIGHTened wave
(Cisco/NLR)
DEN
KAN
SVL
PIT
CHI
CLE
WDC
Cisco/UltraLight wave
LONI wave
San Diego
TUL
VCL @NCSU
DAL
Members:
- MCNC GCNS
- LSU CCT
- NCSU
- RENCI
Official Partners:
- AT&T Research
- SURA
- NRL
- Cisco Systems
- Calient Networks
- IBM
HOU
NSF Project Partners
- OptIPuter
- UltraLight
- DRAGON
- Cheetah
International
Partners
•Phosphorus - EC
•G-lambda - Japan
-GLIF
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
34
HARC: Highly Available Robust Co-allocator
• Extensible, open-sourced co-allocation system
• Can already reserve:
– Time on supercomputers (advance reservation), and
– Dedicated paths on GMPLS-based networks with simple topologies
• Uses Paxos Commit to atomically reserve multiple resources, while
providing a highly-available service
• Used to coordinate bookings across EnLIGHTened and G-lambda
testbeds in largest demonstration of its kind to date (more later)
• Used for setting up the
network for Thomas Sterling’s
HPC Class which
goes out live in
HD (more later)
Credit: Jon MacLaren
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
35
Request Network
bandwidth
and Computers
CRM
CRM
Application
(Visualization)
Request Network
bandwidth
and Computers
Reservation
From xx:xx to yy:yy
KDDI
NRM
US
JAPAN
Application
(MPI)
Reservation
From xx:xx to yy:yy
EL
NRM
NTT
NRM
CRM
CRM
CRM
CRM
CRM
CRM
Cluster CENTER
Cluster FOR COMPUTATION
Cluster Cluster
Cluster
Cluster
36
& TECHNOLOGY
AT LOUISIANA
STATE Cluster
UNIVERSITY Cluster
Data grid applications
• Remote visualization
– Data is somewhere, needs to flow quickly and
smoothly to a visualization app
– Data could be simulation results, or measured data
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
37
Distributed Viz/Collaboration
• iGrid 2005 demo
• Visualization at
LSU
• Interaction among
San Diego, LSU,
Brno
• Data on remote
LONI machines
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
38
Video for visualization
• But also for videoconference between the three
sites
• 1080i (1920x1080, 60fps interlaced):
• 1.5 Gbps / unidirectional stream, 4.5 Gbps each site (two
incoming, one outgoing streams)
• Jumbo frames (9000 bytes), Layer 2 lossless (more or less)
dedicated network
• Hardware capture:
• DVS Centaurus (HD-SDI) + DVI -> HD-SDI converter
from Doremi
Credit: Andrei Hutanu
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
39
Hardware setup – one site
Credit: Andrei Hutanu
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
40
Video distribution
• Done in software (multicast not up to speed,
optical multicast complicated to set up). Can do
1:4 distribution with high-end Opteron
workstations.
• HD class 1-to-n
– Only one stream is distributed - the one showing the
presenter (Thomas Sterling) - others are just to LSU
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
41
Data analysis
• Future scenario motivated by increases in
network speed
• Possibilities of simulations to store results locally
are limited
– Downsampling the output, not storing all data
• Use remote (distributed, possibly virtual) storage
– Can store all data
– This will enable new types of data analysis
Credit: Andrei Hutanu
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
42
Components
• Storage
– high-speed distributed file systems or virtual RAM disks
– potential use cases: global checkpointing facility; data analysis
using the data from this storage
• distribution could be
determined by the
analysis routines
• Data access
– Various data selection
routines gather data
from the distributed storage
elements (storage supports
app-specific operations)
Credit: Andrei Hutanu
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
43
More Components
• Data transport
– Components of the storage are connected by various networks.
May need to use different transport protocols
• Analysis (visualization or
numerical analysis)
– Initially single-machine but can
also be distributed
• Data source
– computed in advance and
preloaded on the distributed
storage initially
– or live streaming from the
distributed simulation
Credit: Andrei Hutanu
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
44
Conclusions
• Applications exist where infrastructure exists that
enables them
• Very few applications (and application authors)
can afford to get ahead of the infrastructure
• We can run the same (grid-unaware)
applications on more resources
– Perhaps add features such as fault tolerance
• Use SAGA to help here?
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
45
SAGA
• Intent: SAGA is to grid apps what MPI is to parallel apps
• Questions/metrics:
– Does SAGA enable rapid development of new apps?
– Does it allow complex apps with less code?
– Is it used in libraries?
• Roots: Reality Grid (ReG Steering Library), GridLab (GAT), and
others came together at GGF
• Strawman API:
– Uses SIDL (from Babel, CCA)
– Language independent spec.
– OO base design - can adapt to procedural languages
• Status:
– Started between GGF 11 & GGF 12 (July/Aug 2004)
– Draft API submitted to OGF early Oct. 2006
– Currently, responding to comments…
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
46
More Conclusions
• Infrastructure is getting better
• Middleware developers are working on some of
the right problems
– If we want to keep doing the same things better
– And add some new things (grid-aware apps)
• Web 3.1 is coming soon…
– We’re not driving the distributed computing world
– Have to keep trying new things
CENTER FOR COMPUTATION & TECHNOLOGY AT LOUISIANA STATE UNIVERSITY
47
Download