Region-based Hierarchical Operation Partitioning for Multicluster Processors

1
Region-based Hierarchical Operation Partitioning
for Multicluster Processors

• Michael Chu, Kevin Fan, Scott Mahlke
• University of Michigan
• Presented by Cristian Petrescu-Prahova

2
Clustered Register Files

• Why?
• Register file cost and access time grows with the
  square of he number of register ports
• Bypass logic grows quadratically with the number
  of operations issued per cycle
• Distance separating FUs from register file
  increases with a large number of FUs
• gt Clustered register files
• Decentralized architecture with several small
  register files
• Each register file supplies operands to a subset
  of FUs
• Multiflow Trace, Alpha 21264, TI C6x, Analog
  Tigersharc (two clusters) reconfigurable meshes?

3
Goal

• Partition operations across the resources
  available on each cluster to maximize ILP
• Minimize inter-cluster communication
• Rule of thumb
• 2 identical clusters processor loose 20
  performance
• 4 identical clusters processor loose 30
  performance
• Nonidentical clusters lead to even more
  performance loss

4
Well Known TechniqueBottom-Up Greedy

• Recurse along DFG, critical path first
• Assign each operation a cluster based on
  estimates of when the operation and its
  predecessors can complete earliest (from
  scheduler)
• Problem 1 makes local decisions (see figure)
• Problem 2 is slow - needs to query accurate
  cluster status info for each operation considered

5
Region-Based Hierarchical Operation Partitioning

• Works on acyclic DFGs extracted from the complete
  program based on region decomposition. I assume
  region loop (?!?)
• Two phases
• Weigth calculation Node and Edge
• Partitioning Coarsening and Refining

6
Node Weight Calculation

• Reflects the quantity of resources per operation
• Ignores dependencies
• Individual weight (FUs)
• Shared weight (ports, buses)

7
Edge Weight Calculation

• Measure of criticalness
• Based on the notion of slack
• First come first serve slack distribution

8
Coarsening Partitioning

• Multilevel graph partitioning algorithm (Chaco,
  Metis)
• Works by coarsening highly related nodes into
  partitions, takes in account only edge weights
• Takes a snapshot of each step for refining step

9
Refinement Partitioning

• Traverse back the coarsening stages, making
  improvements to the initial partition
• At each stage the coarsened nodes available at
  that point are considered for movement to another
  cluster
• Highly related operations are grouped together at
  each stage because we follow the coarsening
  process backwards
• Metrics
• Cluster weight
• estimate of the load per cluster
• the cluster with highest weight is denoted the
  imbalanced cluster
• System load
• Estimates the load across all clusters
• Gain
• The gain of moving operations into other clusters

10
Cluster Weight

• Individual resource constraint per cluster, per
  cycle (op groups)
• Total node weight per cluster per cycle (shared
  constraints)
• Cycle weight per cluster
• Cluster weight

11
Sytem Load

• Inter-cluster move overhead
• Total load, based on cycle by cycle estimation

12
Gain

• Load gain
• Edge gain
• Move gain

13
Example
14
Evaluation

• Implemented using Trimaran tool set
• Compared with BUG algorithm
• 5 DSP benchmarks (high ILP), SPECint2000 (low
  ILP)
• 5 configurations, functional units integer (I),
  float (F), memory (M), branch (B)

15
Improvement in dynamic total cycles of RHOP over
BUG
16
Comparison of BUG and RHOP clustering performance
versus a 1-cluster machine
2-1111 processor
4-1111 processor
17
Histogram of RHOP versus BUG
Achieved schedule length versus critical path
length. Numbers of top are dynamic execution
percentage
18
Compiling performance number of calls to the
resource table