Clustering is one solution to the demand for wideissue machines and fast clock cycles because it allows for smaller, less ported register files and simpler bypass logic while remaining scaleable. Much of the previous work on scheduling for clustered architectures has focused on acyclic code. While minimizing schedule length of acyclic code is paramount, the primary objective when scheduling cyclic code is to maximize the throughput or steady state performance. This paper investigates a pre-modulo scheduling pass that performs cluster assignment in a way that minimizes performance degradation do to explicit communication required as the loops are split over clusters. The proposed cluster assignment algorithm annotates and adjusts the graph for use by the scheduler so that any traditional modulo scheduling algorithm, having no knowledge of clustering, can produce a valid and efficient schedule for a clustered machine.

Clustered ILP processors are characterized by a large number of non-centralized on-chip resources grouped into clusters. Traditional code generation schemes for these processors consist of multiple phases for cluster assignment, register allocation and instruction scheduling. Most of these approaches need additional re-scheduling phases because they often do not impose finite resource constraints in all phases of code generation. These phase-ordered solutions have several drawbacks, resulting in the generation of poor performance code. Moreover, the iterative/back-tracking algorithms used in some of these schemes have large running times. In this paper we present CARS, a code generation framework for Clustered ILP processors, which combines the cluster assignment, register allocation, and instruction scheduling phases into a single code generation phase, thereby eliminating the problems associated with phase-ordered solutions. The CARS algorithm explicitly takes into account all the resource constraints at each cluster scheduling step to reduce spilling and to avoid iterative re-scheduling steps. We also present a new on-the-fly register allocation scheme developed for CARS. We describe an implementation of the proposed code generation framework and the results of a performance evaluation study using the SPEC95/2000 and MediaBench benchmarks.

Clustered architectures are a solution to the bottleneck of centralized register files in superscalar and VLIW processors. The main challenge associated with clustered architectures is compiler support to effectively partition operations across the available resources on each cluster. In this work, we present a novel technique for clustering operations based on graph partitioning methods. Our approach incorporates new methods of assigning weights to nodes and edges within the dataflow graph to guide the partitioner. Nodes are assigned weights to reflect their resource usage within a cluster, while a slack distribution method intelligently assigns weights to edges to reflect the cost of inserting moves across clusters. A multilevel graph partitioning algorithm, which globally divides a dataflow graph into multiple parts in a hierarchical manner, uses these weights to efficiently generate estimates for the quality of partitions. We found that our algorithm was able to achieve an average of 20 % improvement in DSP kernels and 5 % improvement in SPECint2000 for a four-cluster architecture. Categories and Subject Descriptors D.3.4 [Programming Languages]: Processors—code generation, retargetable compilers; C.1.1 [Processor Architectures]:

applicationspecific design, architecture synthesis, bitwidth, clustering, embedded system, hardware accelerator, operation scheduling, resource allocation PICO is a system for automatically synthesizing embedded hardware accelerators from loop nests specified in the C programming language. A key issue confronted when designing such accelerators is the optimization of hardware by exploiting information that is known about the varying number of bits required to represent and process operands. In this paper, we describe the handling and exploitation of integer bitwidth in PICO. A bitwidth analysis procedure is used to determine bitwidth requirements for all integer variables and operations in a C application. Given known bitwidths for all variables, complex problems arise when determining a program schedule that specifies on which function unit and at what time each operation executes. If operations are assigned to function units with no knowledge of bitwidth, bitwidth-related cost benefit is lost when each unit is built to accommodate the widest operation assigned. By carefully placing operations of similar width on the same unit, hardware costs are decreased. This problem is addressed using a preliminary clustering of operations that is based jointly on width and implementation cost. These clusters are then honored during resource allocation and operation scheduling to create an efficient widthconscious design. Experimental results show that exploiting integer bitwidth substantially reduces the gate count of PICO-synthesized hardware accelerators across a range of applications.

Convergent scheduling is a general framework for cluster assignment and instruction scheduling on spatial architectures. A convergent scheduler is composed of independent passes, each implementing a heuristic that addresses a particular problem or constraint. The passes share a simple, common interface that provides spatial and temporal preference for each instruction. Preferences are not absolute; instead, the interface allows a pass to express the confidence of its preferences, as well as preferences for multiple space and time slots. A pass operates by modifying these preferences. By applying a series of passes that address all the relevant constraints, the convergent scheduler can produce a schedule that satisfies all the important constraints. Because all passes are independent and need to understand only one interface to interact with each other, convergent scheduling simplifies the problem of handling multiple constraints and codeveloping different heuristics. We have applied convergent scheduling to two spatial architectures: the Raw processor and a clustered VLIW machine. It is able to successfully handle traditional constraints such as parallelism, load balancing, and communication minimization, as well as constraints due to preplaced instructions, which are instructions with predetermined cluster assignment. Convergent scheduling is able to obtain an average performance improvement of 21% over the existing space-time scheduler of the Raw processor, and an improvement of 14% over state-of-the-art assignment and scheduling techniques on a clustered VLIW architecture.

retargetable compilers, table-driven compilers, machine description, processor description, instruction-level parallelism, EPIC processors, VLIW processors, EPIC compilers, VLIW compilers, code generation, scheduling, register allocation In the past, due to the restricted gate count available on an inexpensive chip, embedded DSPs have had limited parallelism, few registers and irregular, incomplete interconnectivity. More recently, with increasing levels of integration, embedded VLIW processors have started to appear. Such processors typically have higher levels of instruction-level parallelism, more registers, and a relatively regular interconnect between the registers and the functional units. The central challenges faced by a code generator for an EPIC (Explicitly Parallel Instruction Computing) or VLIW processor are quite different from those for the earlier DSPs and, consequently, so is the structure of a code generator that is designed to be easily retargetable. In this report, we explain the nature of the challenges faced by an EPIC or VLIW compiler and present a strategy for performing code generation in an incremental fashion that is best suited to generating high-quality code efficiently. We also describe the Operation Binding Lattice, a formal model for incrementally binding the opcodes and register assignments in an EPIC code generator. As we show, this reflects the phase structure of the EPIC code generator. It also defines the structure of the machine-description database, which is queried by the code generator for the information that it needs about the target processor. Lastly, we discuss general features of our implementation of these ideas and techniques in Elcor, our EPIC compiler research infrastructure.

In conventional superscalar microarchitectures with partitioned integer and floating-point resources, all floating-point resources are idle during execution of integer programs. Palacharla and Smith [26] addressed this drawback and proposed that the floating-point subsystem be augmented to support integer operations. The hardware changes required are expected to be fairly minimal. To exploit these

Modern computers have taken advantage of the instruction-level parallelism (ILP) available in programs with advances in both architecture and compiler design. Unfortunately, large amounts of ILP hardware and aggressive instruction scheduling techniques put great demands on a machine's register resources. With increasing ILP, it becomes difficult to maintain a single monolithic register bank and a high clock rate. To provide support for large amounts of ILP while retaining a high clock rate, registers can be partitioned among several different register banks. Each bank is directly accessible by only a subset of the functional units with explicit inter-bank copies required to move data between banks. Therefore, a compiler must deal not only with achieving maximal parallelism via aggressive scheduling, but also with data placement to limit inter-bank copies. Our approach to code generation for ILP architectures with partitioned register resources provides flexibility by representing machine dependent features as node and edge weights and by remaining independent of scheduling and register allocation methods. Experimentation with our framework has shown a degradation in execution performance of 10% on average when compared to an unrealizable monolithic-register-bank architecture with the same level of ILP.

This paper explores hierarchical instruction scheduling for a tiled processor. Our results show that at the top level of the hierarchy, a simple profile-driven algorithm effectively minimizes operand latency. After this schedule has been partitioned into large sections, the bottom-level algorithm must more carefully analyze program structure when producing the final schedule. Our analysis reveals that at this bottom level, good scheduling depends upon carefully balancing instruction contention for processing elements and operand latency between producer and consumer instructions. We develop a parameterizable instruction scheduler that more effectively optimizes this trade-off. We use this scheduler to determine the contention-latency sweet spot that generates the best instruction schedule for each application. To avoid this application-specific tuning, we also determine the parameters that produce the best performance across all applications. The result is a contention-latency setting that generates instruction schedules for all applications in our workload that come within 17 % of the best schedule for each.

This paper presents a new modulo scheduling algorithm for clustered microarchitectures. The main feature of the proposed scheme is that the assignment of instructions to clusters is done by means of graph partitioning algorithms that are guided by a pseudo-scheduler. This pseudo-scheduler is a simplified version of the full instruction scheduler and estimates key constraints that would be encountered in the final schedule.