The high energy consumption of current processors causes several problems, including a limited clock frequency, short battery lifetime, and reduced device reliability. It is therefore important to reduce the energy consumption of the processor. Among resources in a processor, the issue queue (IQ) is a large consumer of energy, much of which is consumed by the wakeup logic. Within the wakeup logic, the tag comparison that checks source operand readiness consumes a significant amount of energy. This paper proposes an energy reduction scheme for tag comparison, called double-stage tag comparison. This scheme first compares the lower bits of the tag and then, only if these match, compares the higher bits. Because the energy consumption of tag comparison is roughly proportional to the total number of bits compared, energy is saved by reducing this number. However, this sequential comparison increases the delay of the IQ, thereby increasing the clock cycle time. Although this can be avoided by allocating an extra cycle to the issue operation, this in turn degrades the IPC. To avoid IPC degradation, we reconfigure a small number of entries in the IQ, where several oldest instructions that are likely to have an adverse effect on performance reside, to a single stage for tag comparison. Our evaluation results for SPEC2017 benchmark programs show that the double-stage tag comparison achieves on average a 21% reduction in the energy consumed by the wakeup logic (15% when including the overhead) with only 3.0% performance degradation.

The region that includes the register file is a hot spot in high-performance cores that limits the clock frequency. Although multibanking drastically reduces the area and energy consumption of the register files of superscalar processor cores, it suffers from low IPC due to bank conflicts. Our skewed multistaging drastically reduces not the bank conflict probability but the pipeline disturbance probability by the second stage. The evaluation results show that, compared with NORCS, which is the latest research on a register file for area and energy efficiency, a proposed register file with 18 banks achieves a 39.9% and 66.4% reduction in circuit area and in energy consumption, while maintaining a relative IPC of 97.5%.

Out-of-order superscalar processors have high performance but consume a large amount of energy for dynamic instruction scheduling. We propose a front-end execution architecture (FXA) for improving the energy efficiency of out-of-order superscalar processors. FXA has two execution units: an out-of-order execution unit (OXU) and an in-order execution unit (IXU). The OXU is the execution core of a common out-of-order superscalar processor. In contrast, the IXU consists only of functional units and a bypass network only. The IXU is placed at the processor front end and executes instructions in order. The IXU functions as a filter for the OXU. Fetched instructions are first fed to the IXU, and the instructions are executed in order if they are ready to execute. The instructions executed in the IXU are removed from the instruction pipeline and are not executed in the OXU. The IXU does not include dynamic scheduling logic, and thus its energy consumption is low. Evaluation results show that FXA can execute more than 50% of the instructions by using IXU, thereby making it possible to shrink the energy-consuming OXU without incurring performance degradation. As a result, FXA achieves both high performance and low energy consumption. We evaluated FXA and compared it with conventional out-of-order/in-order superscalar processors after ARM big.LITTLE architecture. The results show that FXA achieves performance improvements of 7.4% on geometric mean in SPECCPU INT 2006 benchmark suite relative to a conventional superscalar processor (big), while reducing the energy consumption by 17% in the entire processor. The performance/energy ratio (the inverse of the energy-delay product) of FXA is 25% higher than that of a conventional superscalar processor (big) and 27% higher than that of a conventional in-order superscalar processor (LITTLE).

Out-of-order superscalar processors rename register numbers to remove false dependencies between instructions. A renaming logic for register renaming is a high-cost module in a superscalar processor, and it consumes considerable energy. A renamed trace cache (RTC) was proposed for reducing the energy consumption of a renaming logic. An RTC caches and reuses renamed operands, and thus, register renaming can be omitted on RTC hits. However, conventional RTCs suffer from several performance, energy consumption, and hardware overhead problems. We propose a semi-global renamed trace cache (SGRTC) that caches only renamed operands that are short distance from producers outside traces, and solves the problems of conventional RTCs. Evaluation results show that SGRTC achieves 64% lower energy consumption for renaming with a 0.2% performance overhead as compared to a conventional processor.

Dynamic instruction window resizing (DIWR) is a scheme that effectively exploits both memory-level parallelism and instruction-level parallelism by configuring the instruction window size appropriately for exploiting each parallelism. Although a previous study has shown that the DIWR processor achieves a significant speedup, power consumption has not been explored. The power consumption is increased in DIWR because the instruction window resources are enlarged in memory-intensive phases. If the power consumption exceeds the power budget determined by certain requirements, the DIWR processor must save power and thus, the performance previously presented cannot be achieved. In this paper, we explore to what extent the DIWR processor can achieve improved performance for a given power budget, assuming that dynamic voltage and frequency scaling (DVFS) is introduced as a power saving technique. Evaluation results using the SPEC2006 benchmark programs show that the DIWR processor, even with a constrained power budget, achieves a speedup over the conventional processor over a wide range of given power budgets. At the most important power budget point, i.e., when the power a conventional processor consumes without any power constraint is supplied, DIWR achieves a 16% speedup.

Single-thread performance has not improved much over the past few years, despite an ever increasing transistor budget. One of the reasons for this is that there is a speed gap between the processor and main memory, known as the memory wall. A promising method to overcome this memory wall is aggressive out-of-order execution by extensively enlarging the instruction window resources to exploit memory-level parallelism (MLP). However, simply enlarging the window resources lengthens the clock cycle time. Although pipelining the resources solves this problem, it in turn prevents instruction-level parallelism (ILP) from being exploited because issuing instructions requires multiple clock cycles. This paper proposed a dynamic scheme that adaptively resizes the instruction window based on the predicted available parallelism, either ILP or MLP. Specifically, if the scheme predicts that MLP is available during execution, the instruction window is enlarged and the window resources are pipelined, thereby exploiting MLP. Conversely, if the scheme predicts that less MLP is available, that is, ILP is exploitable for improved performance, the instruction window is shrunk and the window resources are de-pipelined, thereby exploiting ILP. Our evaluation results using the SPEC2006 benchmark programs show that the proposed scheme achieves nearly the best performance possible with fixed-size resources. On average, our scheme realizes a performance improvement of 21% over that of a conventional processor, with additional cost of only 6% of the area of the conventional processor core or 3% of that of the entire processor chip. The evaluation results also show 8% better energy efficiency in terms of 1/EDP (energy-delay product).

This paper evaluates the delay of the issue queue in a superscalar processor to aid microarchitectural design, where quick quantification of the complexity of the issue queue is needed to consider the tradeoff between clock cycle time and instructions per cycle. Our study covers two aspects. First, we introduce banking tag RAM, which comprises the issue queue, to reduce the delay. Unlike normal RAM, this is not straightforward, because of the uniqueness of the issue queue organization. Second, we explore and identify the correct critical path in the issue queue. In a previous study, the critical path of each component in the issue queue was summed to obtain the issue queue delay, but this does not give the correct delay of the issue queue, because the critical paths of the components are not connected logically. In the evaluation assuming 32-nm LSI technology, we obtained the delays of issue queues with eight to 128 entries. The process of banking tag RAM and identifying the correct critical path reduces the delay by up to 20% and 23% for 4- and 8-issue widths, respectively, compared with not banking tag RAM and simply summing the critical path delay of each component.

Modern microprocessors achieve high application performance at the acceptable level of power dissipation. In terms of power to performance trade-off, the instruction window is particularly important. This is because enlarging the window size achieves high performance but naive scaling of the conventional instruction window can severely increase the complexity and power consumption. In this paper, we propose low-power instruction window techniques for contemporary microprocessors. First, the small reorder buffer (SROB) reduces power dissipation by deferred allocation and early release. The deferred allocation delays the SROB allocation of instructions until their all data dependencies are resolved. Then, the instructions are executed in program order and they are released faster from the SROB. This results in higher resource utilization and low power consumption. Second, we replace a conventional issue queue by a direct lookup table (DLT) with an efficient tag translation technique. The translation scheme resolves the instruction dependency, especially for the case of one producer to multiple consumers. The efficiency of the translation scheme stems from the fact that the vast majority of instruction dependency exists within a basic block. Experimental results show that our proposed design reduces the power consumption significantly for SPEC2000 benchmarks.

The understanding of instruction set usage in typical DOS/Windows applications plays a very important role in designing high performance x86 compatible microprocessors. This paper presents the tools to such analysis, the analysis results, and their implications on the design of a RISC-based superscalar processor for efficient x86 instruction execution. The analyzed results are used to optimize the execution of frequently executed instructions and micro operations.

Media processing has become one of the dominant computing workloads. In this context, SIMD instructions have been introduced in current processors to raise performance, often the main goal of microprocessor designers. Today, however, designers have become concerned with the power consumption, and in some cases low power is the main design goal (laptops). In this paper, we show that SIMD ISA extensions on a superscalar processor can be one solution to reduce power consumption and keeping a high performance level. We reduce the average power consumption by decreasing the number of instructions, the number of cache references, and using dynamic power management to transform the speedup in performance in power consumption reduction.

Cache memories are small fast memories used to temporarily hold the contents of main memory that are likely to be referenced by processors so as to reduce instruction and data access time. In study of cache performance, most of previous works have employed simulation-based methods. However, that kind of researches cannot precisely explain the obtained results. Moreover, when a new processor is designed, huge simulations must be performed again with several different parameters. This research classifies cache structures for superscalar processors into four types, and then represents analytical model of instruction fetch process for each cache type considering various kinds of architectural parameters such as the frequency of branch instructions in program, cache miss rate, cache miss penalty, branch misprediction frequency, and branch misprediction penalty, and etc. To prove the correctness of the proposed models, we performed extensive simulations and compared the results with the analytical models. Simulation results showed that the proposed model can estimate the expected instruction fetch rate accurately within 10% error in most cases. This paper shows that the increase of cache misses reduces the instruction fetch rate more severely than that of branch misprediction does. The model is also able to provide exact relationship between cache miss and branch misprediction for the instruction fetch analysis. The proposed model can explain the causes of performance degradation that cannot be uncovered by the simulation method only.

Wire delays, instead of gate delays, are moving into dominance in modern VLSI design. Current synchronous processors have the critical path not in the ALU function but in the cache access. Since the cache performance enhancement is limited by the memory access delay which mainly consists of wire delays, a reduction in gate delays may no longer imply any enhancement in processor performance. To solve this problem, this paper presents a novel architecture, called the Cascade ALU. The Cascade ALU allows super-scalar processors with future technologies to move the critical path into the ALU part. Therefore the Cascade ALU can enjoy the expected progress in future device speed. Since the delay of the Cascade ALU varies depending on the executed instructions, an asynchronous system is shown to be suitable for implementing the Cascade ALU. However an asynchronous system may have a large handshake overhead, this paper also presents an asynchronous Fine Grain Pipeline technique that hides the handshake overhead. Finally, this paper presents results of performance and area evaluation for an asynchronous implementation of the cascade ALU. The results show that the cascade ALU architecture has a good performance scalability on the reduction of the ALU latency and imposes little area penalty compared with current synchronous processors.

The reorder buffer is usually employed to maintain the instruction execution in the correct order for a superscalar pipeline with out-of-order issue. In this paper, we propose a reorder buffer structure with shelter buffer for out-of-order issue superscalar processors not only to control stagnation efficiently, but also to reduce the buffer size. We can get remarkable performance improvement with only one or two buffers. Simulation results show that if the size of reorder buffer is between 8 and 32, performance gain obtained from the shelter is noticeable. For the shelter buffer of size 4, there is no performance improvement compared to that of size 2, which means that the shelter buffer of size 2 is large enough to handle most of the stagnation. If the shelter buffer of size 2 is employed, we can reduce the reorder buffer by 44% in Whetstone, 50% in FFT, 60% in FM, and 75% in Linpack benchmark program without loss of any throughput. Execution time is also improved by 19.78% in Whetstone, 19.67% in FFT, 23.93% in FM, and 8.65% in Linpack benchmark when the shelter buffer is used.

In this paper, an instruction-cache scheme called Multi-Path Tracing is proposed to enhance the trace cache. Paths are classified to improve the trace cache hit ratio by reducing the path conflict and basic blocks are joined to reduce the hardware cost needed to implement the trace cache. Simulation results for various SPEC integer benchmarks show that the proposed scheme increases the hit ratio by more than 25% and the effective fetch size by 10%.

Modern micro-architectures employ superscalar techniques to enhance system performance. Since the superscalar microprocessors must fetch at least one instruction cache line at a time to support high issue rate and large amount speculative executions. There are cases that multiple branches are often encountered in one cycle. And in practical implementation this would cause serious problem while there are variable number of instruction addresses that look up the Branch Target Buffer simultaneously. In this paper, we propose a Range Associative Branch Target Buffer (RABTB) that can recognize and predict multiple branches in the same instruction cache line for a wide-issue micro-architecture. Several configurations of the RABTB are simulated and compared using the SPECint95 benchmarks. We show that with a reasonable size of prediction scope, branch prediction can be improved by supporting multiple / up to 8 branch predictions in one cache line in one cycle. Our simulation results show that the optimal RABTB should be 2048 entry, 8-column range-associate and 8-entry modified ring buffer architecture using PAs prediction algorithm. It has an average 5.2 IPC_f and branch penalty per branch of 0.54 cycles. This is almost two times better than a mechanism that makes prediction only on the first encountered branch.

This research presents a novel analytic model to predict the instruction execution rate of superscalar processors using the queuing model with finite-buffer size and synchronous operation mode. The proposed model is also able to analyze the performance relationship between cache and pipeline. The proposed model takes into account various kinds of architectural parameters such as instruction-level parallelism, branch probability, the accuracy of branch prediction, cache miss, and etc. To prove the correctness of the model, we performed extensive simulations and compared the results with the analytic model. Simulation results showed that the proposed model can estimate the average execution rate accurately within 10% error in most cases. The proposed model can explain the causes of performance bottleneck which cannot be uncovered by the simulation method only. The model is also able to show the effect of the cache miss on the performance of out-of-order issue superscalar processors, which can provide an valuable information in designing a balanced system.

This paper describes an algorithm and its prototype system--VeriProc/1. 1--which can prove the correctness of pipelined and superscalar processor controls automatically without a pipeline invariant, human interaction, or additional information. This algorithm is based on behavior-covering and partial unfolding. No timing relations such as an abstract function or β-relation is required. The only information required is to specify the location of the selectors in the design. Partial unfolding makes it possible to derive superscalar specifications from conventional specifications. Correctness proof of the partial unfolding is given. The prototype system can verify various superscalar control designs of simple processors.

In order to improve microprocessor performance, we propose to utilize histories of dynamic instruction sequences. A lot of special purpose memories integrated in a processor chip hold the histories. In this paper, we describe the usefulness of using two special purpose memories: Non-Consecutive basic block Buffer (NCB) and Reference Prediction Table (RPT). The NCB improves instruction fetching efficiency in order to relieve control dependences. The RPT predicts data addresses in order to speculate data dependences. From the simulation study, it has been found that the proposed mechanisms improve processor performance by up to 49. 2%.

The new technique for reducing the load latency is presented. This technique, named tunneling-load, utilizes the register specifier buffer in order to reduce the load latency without fetching the data cache speculatively, and thus eliminates the drawback of any load address prediction techniques. As a consequence of the trend toward increasing clock frequency, the internal cache is no longer able to fill the speed gap between the processor and the external memory, and the data cache latency degrades the processor performance. In order to hide this latency, several techniques predicting the load address have been proposed. These techniques carry out the speculative data cache fetching, which causes the explosion of the memory traffic and the pollution of the data cache. The tunneling-load solves these problems. We have evaluated the effects of the tunneling-load, and found that in an in-order-issue superscalar platform the instruction level parallelism is increased by approximately 10%.

This paper suggests a novel analytical model to predict average issue rate of both in-order and out-of-order issue policies. Most of previous works have employed only simulation methods to measure the instruction-level parallelism for performance. However these methods cannot disclose the cause of the performance bottle-neck. In this paper, the proposed model takes into account such factors as issue policy, instruction-level parallelism, branch probability, the accuracy of branch prediction, instruction window size, and the number of pipeline units to estimate the issue rate more accurately. To prove the correctness of the model, extensive simulations were performed with Intel 80386/80387 instruction traces. Simulation results showed that the proposed model can estimate the issue rate accurately within 3-10% differences. The analytical model and simulations show that the out-of-order issue can improve the superscalar performance by 70-206% compared to the in-order issue. The model employs parameters to characterize the behavior of programs and the structure of superscalar that cause performance bottle-neck. Thus, it can disclose the cause of the disproportion in performance and reduce the burden of excess simulations that should be performed whenever a new processor is designed.