High-level synthesis with distributed controller for fast timing closure

The timing closure problem (e.g., meeting timing/clock period constraint) is one of the most important problems in the design automation. However, the rapid increase of the impact of the process variation on circuit timing makes the problem much more complicated and unpredictable to tackle in synthesis. This paper addresses a new problem of high-level synthesis (HLS) that effectively takes into account the timing variation. Conventional HLS may simply avoid the timing variation problem by considering the worst case or constant delay model of hardware resources, which are certainly not viable solutions. To overcome the impact of timing variation in HLS, this paper addresses the following two problems: 1) how can the statistical static timing analysis (SSTA) used in logic synthesis be modified and effectively applied to the delay and yield computation in HLS? and 2) how can scheduling and resource binding tasks effectively solve the HLS problem with the objective of minimizing latency under yield constraint? Through an extensive experimentation, we confirm that the proposed variation-aware HLS algorithm is able to reduce the latency by 19.7% under 90% performance yield constraint, compared with the result by conventional timing variation-unaware HLS .

With the increasing cost of on-chip global communication, high-performance designs for data-intensive applications require architectures that distribute hardware resources (computing logic, memories, interconnect, etc.) throughout the chip, while restricting computations and communications to geographic proximities. In this paper, we present a methodology for high-level synthesis (HLS) of distributed logic-memory architectures, i.e., architectures that have logic and memory distributed across several partitions in a chip. Conventional HLS tools are capable of extracting parallelism from a behavior for architectures that assume a monolithic controller/datapath communicating with a memory or memory hierarchy. This paper provides techniques to extend the synthesis frontier to more general architectures that can extract both coarse and fine-grained parallelism from data accesses and computations in a synergistic manner. Our methodology selects many possible ways of organizing data and computations, carefully examines the tradeoffs (i.e., communication overheads, synchronization costs, area overheads) in choosing one solution over another, and utilizes conventional HLS techniques for intermediate steps. We have evaluated the proposed framework on several benchmarks by generating register-transfer level (RTL) implementations using an existing commercial HLS tool with and without our enhancements, and by subjecting the resulting RTL circuits to logic synthesis and layout. The results show that circuits designed as distributed logic-memory architectures using our framework achieve significant (up to 5.3/spl times/, average of 3.5/spl times/) performance improvements over well-optimized conventional designs with small area overheads (up to 19.3%, 15.1% on average). At the same time, the reduction in the energy-delay product is by an average of 5.9/spl times/ (up to 11.0/spl times/).

Group signature is one of the main theme in recent digital signature studies. Typical signature algorithm is a combination of more than 70 elliptic curve (ECC), modular (RSA), long-bit integer and hash arithmetic functions. A full H/W IP core is strongly desired for the use of group signature in SoCs in slow-clock and low-power mobile devices and embedded systems. Flexible adjustment of H/W speed and size, depending on different systems and LSI process technologies, is also required. However, for designing and verifying H/W, the group signature algorithm is too complicated to use a standard RTL (Register Transfer Level) design methodology nor any recent HLS (High Level Synthesis). Therefore, we incorporated a two-level behavioral synthesis approach, where an optimized macro-architecture is explored by a custom-made scheduler, after a database of multiple number of microarchitectures are effectively constructed by conventional HLS. We implemented the signature algorithm on a low-cost 0.25um gate-array. The H/W size is approximately 1M gates and our chip can compute a group signature at the equivalent speed (0.135 seconds@100MHz clock) with 3GHz PC S/W, while the power consumption is two orders of magnitude lower (425mW@100MHz).

Conventional High-Level Synthesis (HLS) tools exploit parallelism mostly at the Instruction Level (ILP). They statically schedule the input specifications and build centralized Finite State Machine (FSM) controllers. However, aggressive exploitation of ILP in many applications has diminishing returns and, usually, centralized approaches do not efficiently exploit coarser parallelism, because FSMs are inherently serial. In this paper we present an HLS framework able to synthesize applications that, beside ILP, also expose Task Level Parallelism (TLP). An application can expose TLP through annotations that identify the parallel functions (i.e., tasks). To generate accelerators that efficiently execute concurrent tasks, we need to solve several issues: devise a mechanism to support concurrent execution flows, exploit memory parallelism, and manage synchronization. To support concurrent execution flows, we introduce a novel adaptive controller. The adaptive controller is composed of a set of interacting control elements that independently manage the execution of a task. These control elements check dependencies and resource constraints at runtime, enabling as soon as possible execution. To support parallel access to shared memories and synchronization, we integrate with a novel Hierarchical Memory Interface (HMI). With respect to previous solutions, the proposed interface supports multi-ported memories and atomic memory operations, which commonly occur in parallel programming. Our framework can generate the hardware implementation of C functions by employing two different approaches, depending on its characteristics. If a function exposes TLP, then the framework generates hardware implementations based on the adaptive controller. Otherwise, the framework implements the function through the FSM approach, which is optimized for ILP exploitation. We evaluate our framework on a set of parallel applications and show substantial performance improvements (average speedup of 4.7) with limited area overheads (average area increase of 5.48 times)

An inherent performance gap between custom designs and ASICs is one of the reasons why many designers still start their designs from register transfer level (RTL) description rather than from behavioral description, which can be synthesized to RTL via high-level synthesis (HLS). Sequencing overhead is one of the factors for this performance gap; the choice between latch and flip-flop is not typically taken into account during HLS, even though it affects all the steps of HLS. HLS-l is a new design framework that employs high-performance latches during scheduling, allocation, and controller synthesis. Its main feature is a new scheduler that is based on a concept of phase step (as opposed to conventional control step), which allows us scheduling in finer granularity, register allocation that resolves the conflict of latch being read and written at the same time, and controller synthesis that exploits dual-edge triggered storage elements to support phase step based scheduling. In experiments on benchmark designs implemented in 1.2 V, 65-nm CMOS technology, HLS-l reduced latency by 16.6% on average, with 9.5% less circuit area, compared to the designs produced by conventional HLS.

As the interconnect delay has become the most dominant factor of circuit performances, logic synthesis results without considering interconnect delay tend to lead to unsatisfactory circuit performance. In this paper, a post-layout re-synthesis scheme is proposed to tackle such a problem. In this algorithm, a local logic substitution scheme based on Alternative Logic Equivalent Block (ALEB) is used. Using the interconnect delay information provided by placement, this algorithm improves the circuit delay by substitutions of local logic blocks located on critical paths, followed by an incremental placement, is applied upon those altered blocks. The experimental results show that the algorithm reduces nearly 30% of the critical path delay without causing much area overhead.

This paper proposes an automatic design flow of datapath-dominated applications which is able to deal with optimization and equivalence checking of multi-output polynomials over Z <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</sup> . This paper also gives four main contributions: 1) proposing a complete design flow for modular equivalence checking, high level synthesis, and optimization; 2) considering hidden monomials to factorize those polynomials which do not have any common monomials; 3) combining and improving our previous optimization heuristics to eliminate multi-operand common sub-expressions as much as possible; and 4) implementing all algorithms on top of the modular Horner expansion diagram package. Experimental results have shown an average saving of 9.9% and 4.5% in the number of gates and critical path delay, respectively, after applying modular reduction over Z <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</sup> , while other optimization techniques are not used. Besides, after applying our optimization techniques, experimental comparisons with the state-of-the-art techniques show an average improvement in the area by 19.5% with an average delay decrease of 16.7%. Regarding the comparison with our previous papers, the area and delay are improved by 13.3% and 15.5%, respectively.

Interconnects (wires, buffers, clock distribution networks, multiplexers and buses) consume a significant fraction of total circuit power. In this work, we demonstrate the importance of optimizing on-chip interconnects for power during high-level synthesis. We present a methodology to integrate interconnect power optimization into high-level synthesis. Our binding algorithm not only reduces power consumption in functional units and registers in the resultant register-transfer level (RTL) architecture, but also optimizes interconnects for power. We take physical design information into account for this purpose. To estimate interconnect power consumption accurately for deep sub-micron (DSM) technologies, wire coupling capacitance is taken into. consideration. We observed that there is significant spurious (i.e., unnecessary) switching activity in the interconnects and propose techniques to reduce it. Compared to interconnect-unaware power-optimized circuits, our experimental results show that interconnect power can be reduced by 53.1% on an average, while reducing overall power by an average of 26.8% with 0.5% area overhead. Compared to area-optimized circuits, the interconnect power reduction is 72.9% and overall power reduction is 56.0% with 44.4% area overhead.

Interconnects (wires, buffers, clock distribution networks, multiplexers and buses) consume a significant fraction of total circuit power. In this work, we demonstrate the importance of optimizing on-chip interconnects for power during high-level synthesis. We present a methodology to integrate interconnect power optimization into high-level synthesis. Our binding algorithm not only reduces power consumption in functional units and registers in the resultant register-transfer level (RTL) architecture, but also optimizes interconnects for power. We take physical design information into account for this purpose. To estimate interconnect power consumption accurately for deep sub-micron (DSM) technologies, wire coupling capacitance is taken into. consideration. We observed that there is significant spurious (i.e., unnecessary) switching activity in the interconnects and propose techniques to reduce it. Compared to interconnect-unaware power-optimized circuits, our experimental results show that interconnect power can be reduced by 53.1% on an average, while reducing overall power by an average of 26.8% with 0.5% area overhead. Compared to area-optimized circuits, the interconnect power reduction is 72.9% and overall power reduction is 56.0% with 44.4% area overhead.

The crosstalk noise induced on a net depends on its wire size, its driver size and the sizes of the aggressor gates driving its coupled nets. The problem of crosstalk noise optimization at post-route level is difficult due to the cyclic dependency on aggressor and victim driver sizes, resulting in a conflicting situation. Game theory provides a natural framework for handling conflicting objectives and allows simultaneous optimization of multiple parameters. The formulation of a convex objective function is a requirement in order to obtain better optimization in a game theoretic framework. This property is exploited to solve the cyclic dependency of crosstalk noise on its gate and wire sizes, while modeling the problem of simultaneous optimization of interconnect delay and crosstalk noise. A game is modeled with the crosstalk noise and interconnect delay of the chosen net as the players, the possible gate and wire sizes as the strategy set and the analytical expressions for crosstalk noise and interconnect delay as their respective expected payoffs. The time and space complexity of the proposed integrated sizing algorithm is linear in terms of the number of gates and wires in the design. Experimental results on several medium and large open core designs indicate average improvements of 27.52% and 34.03% for interconnect delay and crosstalk noise respectively, achieved by game theoretic gate sizing based on noise criticality over and above the optimization from the Cadence place and route tools without any area overhead or the need for rerouting

In deep-submicron designs, the interconnects are equally as or more important than the logic gates. In particular, to achieve timing closure, it is necessary and critical to consider the interconnect delay at an early stage of the synthesis process. It has been known that resource sharing in high-level synthesis is one of the major synthesis tasks which greatly affect the final synthesis/layout results. In this paper, we propose a new layout-aware resource sharing approach to overcome some of the limitations of the previous works in which the effects of layout on the synthesis have never been taken into account or considered in local and limited ways, or whose computation time is excessively large. The proposed approach consists of two steps: (Step 1) We relax the integrated resource sharing and placement into an efficient linear programming (LP) formulation based on the concept of $discretisizing$ placement space; (Step 2) We derive a feasible solution from the solution obtained in Step 1. Then, we employ an iterative mechanism based on the two steps to tightly integrate resource sharing and placement tasks so that the slack time violation due to interconnect delay (determined by placement) as well as logic delay (determined by resource sharing) should be minimized. From experiments using a set of benchmark designs, it is shown that the approach is effective, and efficient, completely removing the slack time violation produced by conventional methods.

Interconnect delays represent an increasingly dominant portion of overall circuit delays. During timing-driven physical synthesis process, timing analysis is repeatedly performed over several hundred thousand components. Thus, fast and accurate estimation of interconnect delays is crucial. Traditionally, lumped and elmore delay models have been widely used for computing interconnect delays in physical synthesis due to their computational efficiency. However, these delay models are known to be inaccurate since they ignore slew and resistive shielding effects. In this paper, we propose a new iterative refinement based delay estimation approach that considers resistive shielding along with driver slew. Experimental results show that the proposed approach gives not only highly accurate results for far end RC-line delays but also compares very favorably to more difficult to match source end delays and source slews. In addition, use of the proposed delay model in physical synthesis yields significant performance improvement on several large industrial ASICs.

Summary form only given. Incorporating physical information into earlier architectural and logic synthesis stages is highly desirable since it allows more realistic exploration of the design space and the generation of solutions with predictable metrics. We present a forward-looking synthesis methodology in which we weigh all nets in the control data flow graph (CDFG) according to their criticality. We cluster operations in the CDFG into macros while satisfying logical and physical constraints. We perform relational placement on these macros. We have evaluated the proposed approach using a set of benchmark designs by comparing it with the results of a traditional synthesis flow. The results show that our methodology achieves up to 26% improvement in clock frequency without any area overhead, and average 12.7% improvement in critical path delay with no or little place-and-route time overhead.

High-level synthesis for FPGA designs (FPGA-HLS) is recently required in various applications. Since wire delays are becoming a design bottleneck in FPGA, we need to handle interconnection delays and clock skews in FPGA-HLS flow. In this paper, we propose an FPGA-HLS algorithm optimizing critical path with interconnection-delay and clock-skew consideration. By utilizing HDR architecture, we floorplan circuit modules in HLS flow and, based on the result, estimate interconnection delays and clock skews. To reduce the critical-path delay(s) of a circuit, we propose two novel methods for FPGA-HLS. Experimental results demonstrate that our algorithm can improve circuit performance by up to 24% compared with conventional approaches.

The importance of effective lower bound estimation (LBE) techniques is well established in high-level synthesis (HLS) since it allows more efficient exploration of the design space while providing other HLS tools with the capability of predicting the effect of specific tools on the design space. Much of the previous work has focused on LBE techniques that use very simple cost models which primarily focus on the functional unit resources. With the push toward submicron technologies, simple models that use functional unit resources alone are not accurate enough to allow effective design space exploration since the effects of storage and interconnect can indeed dominate the cost function. In this paper, we present an integrated approach aimed at predicting lower bounds on hardware resources needed to implement a behavioral description within a given amount of time. Our area cost model accounts for storage (register) and interconnect resources (buses) in addition to functional resources. Our timing model uses a finer granularity that permits the modeling of functional unit, register, and interconnect delays. Our approach is integrated because we consider the dependencies between the different types of resources as well as the ordering in which the resources are allocated. We tested our technique for functional unit, storage, and interconnect requirements on several high-level synthesis benchmarks, and observed near-optimal results. We believe that our comprehensive LBE approach can lead to better quality HLS solutions in less time, and we demonstrate this approach in our paper.