1. Deadlock Detection for Distributed Process Networks
Alex G. Olson & Brian L. Evans
The University of Texas at Austin
ICASSP 2005

2. Motivation for Formal Models
Applications may require higher input/output and computational rates than one CPU can handle
Exploit parallelism for high performance
Parallel (one machine) or distributed (many machines)
Pitfalls of parallel/distributed programming
Synchronization, shared memory, and deadlock
Debugging concurrent code on many processors
Formal models have provable properties
Determinacy: programs are correct by construction
Validation: only debug each component separately
Scalability: faster execution with more CPUs
ICASSP 2005

3. Applications Application Input Data Rate Computation Rate
Output Data Rate
Sonar Beamforming
[Allen & Evans 00]
160 MB/s
4-20 GFLOPS
72 MB/s
Bzip2 (block-zip) Compression
1-4 MB/s
~1-4 GIPS (approx)
MPEG4 Encoding (4CIF)
18 MB/s
~2 GIPS
~1 MB/s
H.264 Video Server (QCIF) [Banerjee 02]
1 MB/s
~1 GIPS
~40 KB/s x
Design Space Exploration [Vissers & Wolf, 1999]
Image Processing [Webb et al., 1999]
ICASSP 2005

4. Process Networks [Kahn, 1974]
Concurrently executing processes
Communicate only over one-way unbounded channels (FIFO queues)
Read one input port at a time
Node execution suspended until enough data available
Data that has been read is dequeued from channel
Samples (tokens) flow along arcs
Samples have value but not time information
Flow of (untimed) data drives computation
Determinate execution
Any scheduling algorithm that obeys above rules will produce same history of tokens on arcs
ICASSP 2005

5. Bounding Size of PN Queues
Bounded Scheduling [Parks & Lee, 1995]
Write to a full queue suspends node execution
On global deadlock, resize smallest queue
Favors incomplete bounded execution (non-determinate)
Computational PN [Allen & Evans, 2000]
Processes may consume fewer tokens than read
All memory allocation can be handled by queues
Bounded Scheduling [Geilen & Basten, 2003]
Show local deadlock may not lead to global deadlock
Artificial deadlock
Deadlock detection required for bounded communication, but no framework detects local deadlock
ICASSP 2005

6. Deadlock Detection Algorithm
Mitchell & Merritt’s algorithm [1984]
Detects local and global deadlocks
Exactly one process detects deadlock
Simplifies deadlock resolution
Pair of labels (numbers) used for deadlock detection
Deadlock detected when a label makes a “round-trip” among set of blocked processes
ICASSP 2005

7. Mitchell-Merritt Example
BUSY
BUSY
Write to B
Read from C
1,1
1,2
2,1
3,3
4,2
1,4 A B A B
Blocking Step
Initial State
1,3
1,4 C D C D
Public (count, pid)
Private (count, pid)
BUSY
BUSY
Read from A
Arrows indicate waiting. Artificial deadlock without feedback.
ICASSP 2005

8. Mitchell-Merritt Example
Transmit Step
4,2
2,1
3,3
1,4
Deadlock Detected
4,2
2,1
3,3
1,4 A B A B C D C D
Public Label
Private Label
ICASSP 2005

9. Implementation
Distributed framework for Computational Process Networks
TCP sockets for communication
Transmit and receive queues (zero-copy)
C++, POSIX threads
ICASSP 2005

10. Execution Performance
Overhead <1μs per read/write
ICASSP 2005

11. Execution Performance
Overhead <1μs per read/write
ICASSP 2005

12. Conclusion
Formal models simplify parallel design, implementation, and debugging
Communication in PN model follows “Single-Resource” semantics
Mitchell-Merritt algorithm applicable to non-distributed, parallel, and distributed PN’s
Can be used to implement bounded-memory scheduling algorithms
ICASSP 2005