SIMD Vectorization of Histogram Functions

Multicore processors usually communicate via shared memory, which is backed up by a shared level 2 cache and a cache coherence protocol. However, this solution is not a good fit for real-time systems, where we need to provide tight guarantees on execution and memory access times. In this paper, we propose a shared scratchpad memory as a time-predictable communication and synchronization structure, instead of the level 2 cache. The shared on-chip memory is accessed via a time division multiplexing arbiter, isolating the execution time of load and store instructions between processing cores. Furthermore, the arbiter supports an extended time slot where an atomic load and store instruction can be executed to implement synchronization primitives. In the evaluation we show that a shared scratchpad memory is an efficient communication structure for a small number of processors; in our setup, 9 cores. Furthermore, we evaluate the efficiency of the synchronization support for implementation of classic locks.

Data dependencies have become one of the main bottlenecks of current superscalar processors. Data speculation is gaining popularity as a mechanism to avoid the ordering imposed by data dependencies. Loads and stores are very good candidates for data speculation since their effective address has a regular behavior and then, they are highly predictable. In this paper we propose a mechanism called Address Prediction and Data Prefetching that allows load instructions to obtain their data at the decode stage. Besides, the effective address of load and store instructions is also predicted. These instructions and those dependent on them are speculatively executed. The technique has been evaluated for an out-of-order processor with a realistic configuration. The performance gain is about 19% in average and it is much higher for some benchmarks (up to 35%).

Chapter 6 - Instruction Usage Examples

The widespread deployment of exploit mitigations such as CFI and shadow stacks are making code-reuse attacks increasingly difficult. This has forced adversaries to consider data-only attacks against which the venerable ASLR remains the primary deployed defense.Data-Space Randomization (DSR) techniques raise the bar against data-only attacks by making it harder for adversaries to inject malicious data flows into vulnerable applications. DSR works by masking memory load and store instructions. Masks are chosen (i) to not interfere with intended data flows and (ii) such that masking likely interferes with unintended flows introduced by malicious program inputs. In this paper, we show two new attacks that bypass all existing static DSR approaches; one that directly discloses memory and another using speculative execution. We then present CoDaRR, the first dynamic DSR scheme resilient to disclosure attacks. CoDaRR continuously rerandomizes the masks used in loads and stores, and re-masks all memory objects to remain transparent w.r.t. program execution. Our evaluation confirms that CoDaRR successfully thwarts these attacks with limited run-time overhead in standard benchmarks as well as real-world applications.

In this work, an approach to generate test programs for functional verification of memory management units of microprocessors is proposed. The approach is based on formal specification of memory access instructions, namely load and store instructions, and memory devices such as cache units and address translation buffers. The use of formal specifications helps automate development of test program generators and makes verification systematic due to clear definition of testing goals. In the suggested approach, test programs are constructed by using combinatorial techniques, which means that stimuli - sequences of loads and stores - are created by enumerating all feasible combinations of instructions, situations (instruction execution paths) and dependencies (sets of conflicts between instructions). It is of importance that test situations and dependencies are automatically extracted from specifications. The approach has been used in a number of industrial projects and allowed to discover critical bugs in memory management mechanisms.

Emerging non-volatile memory (NVM) offers non-volatility, byte-addressability and fast access at the same time. To make the best use of these properties, it has been shown by empirical evidence that programs should access NVM directly through CPU load and store instructions, so that the overhead of a traditional file system or database can be avoided. Thus, durable transactions become a common choice of applications for accessing persistent memory data in a crash consistent manner. However, existing durable transaction systems employ either undo logging, which requires a fence for every memory write, or redo logging, which requires intercepting all memory reads within transactions.

Emerging non-volatile memory (NVM) offers non-volatility, byte-addressability and fast access at the same time. To make the best use of these properties, it has been shown by empirical evidence that programs should access NVM directly through CPU load and store instructions, so that the overhead of a traditional file system or database can be avoided. Thus, durable transactions become a common choice of applications for accessing persistent memory data in a crash consistent manner. However, existing durable transaction systems employ either undo logging , which requires a fence for every memory write, or redo logging , which requires intercepting all memory reads within transactions. This paper presents DUDETM, a crash-consistent durable transaction system that avoids the drawbacks of both undo logging and redo logging. DUDETM uses shadow DRAM to decouple the execution of a durable transaction into three fully asynchronous steps. The advantage is that only minimal fences and no memory read instrumentation are required. This design also enables an out-of-the-box transactional memory (TM) to be used as an independent component in our system. The evaluation results show that DUDETM adds durability to a TM system with only 7.4 ~ 24.6% throughput degradation. Compared to the existing durable transaction systems, DUDETM provides 1.7times to 4.4times higher throughput. Moreover, DUDETM can be implemented with existing hardware TMs with minor hardware modifications, leading to a further 1.7times speedup.

Recent developments in register allocation, mostly linked to static single assignment (SSA) form, have shown the benefits of decoupling the problem in two phases: a first spilling phase places load and store instructions so that the register pressure at all program points is small enough, and a second assignment and coalescing phase maps the variables to physical registers and reduces the number of move instructions among registers. This article focuses on the first phase, for which many open questions remain: in particular, we study the notion of optimal spilling (what can be expressed?) and the impact of SSA form (does it help?). To identify the important features for optimal spilling on load-store architectures, we develop a new integer linear programming formulation, more accurate and expressive than previous approaches. Among other features, we can express SSA ϕ-functions, memory-to-memory copies, and the fact that a value can be stored simultaneously in a register and in memory. Based on this formulation, we present a thorough analysis of the results obtained for the SPECINT 2000 and EEMBC 1.1 benchmarks, from which we draw, among others, the following conclusions: (1) rematerialization is extremely important; (2) SSA complicates the formulation of optimal spilling, especially because of memory coalescing when the code is not in conventional SSA (CSSA); (3) microarchitectural features are significant and thus have to be accounted for; and (4) significant savings can be obtained in terms of static spill costs, cache miss rates, and dynamic instruction counts.

The design of software acceptance tests is as important to meeting contract goals as is the design of algorithms. This statement is particularly significant on fixed price contracts with tight schedules. An extreme instance of the demand placed on acceptance testing can be found in software projects wherein the only rigorous testing that required the Computer Program Configuration Item (CPCI) to exercise its repertoire of load and store instructions was the Formal Qualification Test (FQT). This paper is about such a project, the lessons learned from it, and provides an effective test approach for fixed price contracts. A word or two about the project is appropriate to establish the context that underscores the impact of the above assertion. Initially 30K (core words), 16-bit program instructions were to be developed within one year using a Varian 73 computer with 32K words of memory for a Command and Control application under a fixed price contract. A set of a priori conditions existed that tended to convey the impression that the inherent risks of this endeavor were reasonable. They were the “facts” that: Of the 30K (core words) to be written, 30% of this code already existed and would be used. Contractor standards would be allowed for documentation with limited use of Military Specifications No formal Design Reviews or audits would accompany the deliverable CPCI. Existent executive software would suffice. A competent and enthusiastic team was committed to the effort.

Treegraph-based Instruction Scheduling for Stack-based Virtual Machines

In Speculative Multithreading, data dependence that limits the speedup of speculative parallelization needs to be resolved to achieve a high performance. This paper designs a synchronization execution model, with the support of compiler, to synchronize store and load instructions that frequently have data dependence on each other. We use hardware profiler to gather dependence violation information of memory data, and the profiler information is fed back to the compiler. The compiler analyzes the synchronization efficiency, select store/load pairs of great synchronization potential, and inserts synchronization instructions using insertion algorithm. Loop threads and non-loop threads can both be synchronized. The hardware support is also given in the paper. The experimental results show that the synchronization under the compiler support can effectively resolve some memory data dependence and improve the performance of the speculative execution.

SummaryThe primary goal of the register allocation phase in a compiler is to minimize register spills to memory. Spill decisions by the allocator are often made based on the costs of spilling a virtual register and, therefore, on an assumed placement of spill instructions. However, because most allocators make these decisions incrementally, placement opportunities can change as allocation proceeds, calling into question the basis for the original spill decision. An alternative heuristic to placement costs for spill decisions focuses on where program execution will lead. Spilling the virtual register with the Furthest Next Use is known to lead to the minimum number of loads under certain conditions in straight‐line code. While it has been implemented in register allocation in different forms, none of these implementations fully exploits profiling information. We present a register allocator that can adapt to improved profiling information, using branch probabilities to compute an Expected Distance to Next Use for making spill decisions and block frequency information to optimize post‐allocation spill instruction placement. Spill placement is optimized after allocation using a novel method for minimizing spill instruction costs on the control flow graph. Our evaluation of the allocator compared with LLVM recognizes more than 36% and 50% reductions, on average, in the number of dynamically executed store and load instructions, respectively, when using statically derived profiling information. When using dynamically gathered profiling, these improvements increase to 50% and 60% reductions, on average, for stores and loads, respectively. Copyright © 2016 John Wiley & Sons, Ltd.

Energy consumption is one of the most important issues in resource-constrained embedded systems. Many such systems run Java-based applications due to Java's architecture-independent format (bytecode). Standard techniques for executing bytecode programs, e.g. interpretation or just-in-time compilation, have performance or memory issues that make them unsuitable for resource-constrained embedded systems.A superoperator-extended, lightweight Java Virtual Machine (JVM) can be used in resource-constrained embedded systems to improve performance and reduce memory consumption. This paper shows that such a JVM also significantly reduces energy consumption. This is due primarily to a considerable reduction in the number of memory accesses and thus in energy consumption in the instruction and data TLBs and caches and, in most cases, in DRAM energy consumption. Since the fraction of processor energy dissipated in these units is approximately 60%, the energy savings achieved are significant.The paper evaluates the number of load, store, and computational instructions eliminated by the use of proposed superoperators as compared to a simple interpreter on a set of embedded benchmarks. Using cache and DRAM per access energy we estimate the total processor/DRAM energy saved by using our JVM. Our results show that with 32KB caches the reduction in energy consumption ranges from 40% to 60% of the overall processor, plus DRAM energy. Even higher savings may be achieved with smaller caches and increased access to DRAM as DRAM access energy is fairly high.

In this article, we illustrate that the boundary of a general-purpose node can be extended into the network by extracting information from network traffic generated by that general-purpose node to infer the state of its hardware components. This information is represented in a delay signature latent within the network traffic. In contrast, the traditional approach to determine the internal state of a node’s resources meant that a software application with internal processes had to be resident on the node. The aforementioned delay signature is the keystone that provides a correlation between network traffic and the internal state of the source node. We characterize this delay signature by (1) identifying the different types of assembly language instructions that source this delay and (2) describing how architectural techniques, such as instruction pipelining and caching, give rise to this delay signature. In theory, highly utilized nodes (due to multiple threads) will contain excessive context switching and contention for shared resources. One important shared resource is main memory, and excessive use of this resource by applications and internal processes eventually leads to a decrease in cache efficiency that eventually stalls the instruction pipeline. Our results support this theory; specifically, we have observed that excessive context switching in active applications increases the effective memory access time and wastes precious CPU cycles, thus adding additional delay to the execution of load, store, and other instructions. Because the operating system (OS) kernel accesses memory to send network packets, the delay signature is induced into network traffic in situations where user-level utilization is high. We demonstrate this theory in two case studies: (1) resource discovery in cluster grids and (2) network-based detection of bitcoin mining on compromised nodes.

In this paper, we introduce an Array Memory (AM) processor. The AM processor uses a shared memory network amenable to on-chip 3D stacking. Node couplings use a 1 to K adjacency of connections in each dimension of communication of an array of nodes, such as an R×C array where R ≥ K and C ≥ K and K is a positive odd integer. This design also provides data sharing between processors within sub-arrays of the R × C array to support high-performance programmable application specific processing. A new instruction set architecture is proposed that has arithmetic instructions that do not require the specification of any source or target operand addresses. The source operands and target values are provided by separate load, store, and arithmetic instructions that are appropriately scheduled with the arithmetic instruction to be executed to reduce the storage of temporary variables for lower power implementations.