Abstract: Disclosed are an instruction execution method and apparatus for graph computation. The method includes the following steps: S1: sending operators of each node in a computational graph used for neural network computation to an operator interpreter; S2: building, by the operator interpreter, instructions in operation; S3: defining an instruction dependency relationship; S4: building an instruction dependency relationship graph; S5: building a topological order of parallel instructions; S6: scheduling the parallel instructions to hardware resources; S7: building shortest schedules for the parallel instructions: the shortest time required to execute the parallel instructions under the condition of limited hardware resources; and S8: releasing the completed instructions.

Abstract: The present disclosure belongs to the field of intelligent computing technologies, and relates to a multi-policy intelligent scheduling methods and apparatuses oriented to heterogeneous computing power. The method includes: step 1, setting an execution policy of a task based on heterogeneity of computing clusters, differences of computing tasks and a user requirement, and constructing a Markov decision process model by adopting a reinforcement learning method combined with the execution policy; step 2, adopting a proximal policy optimization to solve an optimal task scheduling policy of the task input by the user based on the constructed Markov decision process model; step 3, scheduling the task to a corresponding computing cluster for execution based on the optimal task scheduling policy.

Abstract: The disclosure discloses an intermediate representation method for compiling computation graphs, including: step 1: compiling a neural network into a computation graph for neural network computation; step 2: constructing a node for each tensor variable in the computation graph; step 3: associating the node representing the tensor variable in the computation graph to a set of pointers to the tensor variable; step 4: analyzing constraint relationships between the tensor variables in the computation graph; step 5: iteratively constructing a topological graph of the intermediate representation based on the constraint relationships between the tensor variables in the computation graph; and step 6: analyzing the tensor variables with different aliases pointing to a same memory location based on the intermediate representation, and allocating a register for the tensor variables with different aliases.

Abstract: Disclosed are an intermediate representation method and apparatus for parallel execution of graph computation. The method includes the following steps: S1: compiling a neural network into a computational graph on a computer; S2: defining branch states of tensor variables in the computational graph; S3: defining a data dependency relationship of the tensor variables in the computational graph; S4: defining a control dependency relationship of the tensor variables in the computational graph; S5: building a data dependency relationship graph of the tensor variables in the computational graph; S6: building a control dependency relationship graph of the tensor variables in the computational graph; and S7: transforming control dependencies into data dependencies. The present application derives, based on the dependency relationship, a parallel computing method that can execute the branch threads in parallel in the global computational graph, and optimizes the compilation efficiency of the computational graph.

Abstract: A component-oriented web mashup system and method for communicating between component-oriented Web gadgets to facilitate secure Web mashups. Embodiments of the system and method redefine the traditional definition of gadget to mean a Web component having a verifiable controlled communication channel (a CompoWeb gadget). A CompoWeb gadget is created and defined using new HTML tags and global script objects and functions that extend the functions of the browser. CompoWeb gadget content is treated as a component that is isolated from other gadgets and frames by a browser, and only those allowed access can view data and code therein. Called functions of a CompoWeb gadget are run in the callee's environment instead of the caller's environment. This adds security, because all the requesting CompoWeb gadget receives is the run result. Embodiments of the system and method also include delayed binding of CompoWeb gadgets, such that binding is performed at run time.

Abstract: A method for disassembling an executable binary (binary). In one implementation, a plurality of potential address references may be identified based on the binary and a plurality of storage addresses containing the binary. A plurality of assembler source code instructions (instructions) may be generated by disassembling the binary. The binary may be disassembled at one or more sequential addresses starting at each of the plurality of potential address references.

Abstract: A cooperative rendering cache browser (CRC-Browser) for a mobile device may cooperatively manage cached content with a proxy server to reduce redundant transmissions of processed Web data. Additionally, the CRC-Browser may provide stateful, thin-client Web browsing to maintain synchronized information about rendering objects of a Web page. Further, the CRC-Browser may cache only stable and/or visible document object model (DOM) elements and communicate a list of IDs of the cached elements to the proxy. Such stateful thin-client browsers may be useful for reducing battery drain, poor responsiveness, and high wireless network costs of mobile Web browsing.

Abstract: A cooperative rendering cache browser (CRC-Browser) for a mobile device may cooperatively manage cached content with a proxy server to reduce redundant transmissions of processed Web data. Additionally, the CRC-Browser may provide stateful, thin-client Web browsing to maintain synchronized information about rendering objects of a Web page. Further, the CRC-Browser may cache only stable and/or visible document object model (DOM) elements and communicate a list of IDs of the cached elements to the proxy. Such stateful thin-client browsers may be useful for reducing battery drain, poor responsiveness, and high wireless network costs of mobile Web browsing.

Abstract: Methods and systems for parallel Web page processing are usable to parallelize Web page document parsing, Web page layout calculations, Web page style formatting, and Web page script engine processing. Such parallelized parsers may be used to enhance Web page processing and exploit multi-core and multi-processor computing device resources. The parallelized script engine may be used to enhance Web page processing when independent scripting events exist in the Web page document. Additionally, the parallelized layout calculations and style formatting may be used to further enhance Web page processing by allowing multi-core and multi-processor computing devices to take advantage of their parallel processing abilities.

Abstract: Methods and systems for style and/or layout caching of Web content are usable to build reusable style caching trees and cacheable layout calculations. Such style caching trees may be used to avoid recalculating style content of Web pages for document object model (DOM) elements that have not changed. Additionally, the cacheable layout calculations may be used to avoid recalculating the layout content of Web pages that are subsequently accessed.

Abstract: A security-enhanced login technique that provides a convenient and easy-to-use two factor technique to enhance the security of passwords without requiring any changes on the server side of a client-server network. The technique employs a convenient and easy-to-use two-factor technique to generate strong passwords for Web and other applications. In this technique, a convenient or personal device such as a mouse is used as the other factor besides a user password. A secret stored in the mouse or other personal device is hashed together with the password entered by a user and the server ID, to generate a strong, server-specific password which is used to authenticate the user to the server. This password enhancement operation is carried out inside the personal device.

Abstract: A cross-module detection system and method for detecting and monitoring control flow transfers between software modules in a computer system. The system and method detect and monitor control flows entering and exiting the software modules. For a particular module, a checking model is extracted from the binary file of that module. In addition, a relaxed shadow stack is generated. If the module is an original module, meaning that the control flow originated from that module, then the checking model is used to check the validity of the control flow transfer. Otherwise, the relaxed shadow stack is used. An interception module is used to intercept and terminate invalid control flow transfers. If an invalid control flow transfer is detected, then the transfer is terminated. Otherwise, the control flow transfer is allowed to continue.

Abstract: A “Performance Evaluator” provides various techniques for tracking system events to diagnose root causes of application performance anomalies. In general, traces of system events involved in inter-thread interactions are collected at application runtime. These traces are then used to construct inter-thread dependency patterns termed “control patterns.” Control patterns are then evaluated to determine root causes of performance anomalies. Where an application terminates abnormally or full traces cannot be collected for some reason, partial control patterns are constructed for that application. In various embodiments, “fingerprints” are then generated from full or partial control patterns and are matched to fingerprints corresponding to operations in other control patterns extracted from reference traces collected on the same or similar systems. Matched fingerprints or control patterns are then used to deduce the root cause of application performance anomalies associated with full or partial traces.

Abstract: A cross-module detection system and method for detecting and monitoring control flow transfers between software modules in a computer system. The system and method detect and monitor control flows entering and exiting the software modules. For a particular module, a checking model is extracted from the binary file of that module. In addition, a relaxed shadow stack is generated. If the module is an original module, meaning that the control flow originated from that module, then the checking model is used to check the validity of the control flow transfer. Otherwise, the relaxed shadow stack is used. An interception module is used to intercept and terminate invalid control flow transfers. If an invalid control flow transfer is detected, then the transfer is terminated. Otherwise, the control flow transfer is allowed to continue.

Abstract: A system for facilitating online shopping that includes one or more content providers of a first domain, one or more online merchants of a second domain, and a framework. Web pages may be aggregated from the content providers. Descriptions of products for sale may be aggregated from the online merchants. The products for sale may be matched with the web pages based on relevancy. A list of descriptions of products that are relevant to each web page may be provided. A shopping cart list may be maintained by the framework. The shopping cart list may facilitate payment for the products for sale selected from different web pages.

Abstract: A component-oriented web mashup system and method for communicating between component-oriented Web gadgets to facilitate secure Web mashups. Embodiments of the system and method redefine the traditional definition of gadget to mean a Web component having a verifiable controlled communication channel (a CompoWeb gadget). A CompoWeb gadget is created and defined using new HTML tags and global script objects and functions that extend the functions of the browser. CompoWeb gadget content is treated as a component that is isolated from other gadgets and frames by a browser, and only those allowed access can view data and code therein. Called functions of a CompoWeb gadget are run in the callee's environment instead of the caller's environment. This adds security, because all the requesting CompoWeb gadget receives is the run result. Embodiments of the system and method also include delayed binding of CompoWeb gadgets, such that binding is performed at run time.

Abstract: A security-enhanced login technique that provides a convenient and easy-to-use two factor technique to enhance the security of passwords without requiring any changes on the server side of a client-server network. The technique employs a convenient and easy-to-use two-factor technique to generate strong passwords for Web and other applications. In this technique, a convenient or personal device such as a mouse is used as the other factor besides a user password. A secret stored in the mouse or other personal device is hashed together with the password entered by a user and the server ID, to generate a strong, server-specific password which is used to authenticate the user to the server. This password enhancement operation is carried out inside the personal device.

Abstract: A method for disassembling an executable binary (binary). In one implementation, a plurality of potential address references may be identified based on the binary and a plurality of storage addresses containing the binary. A plurality of assembler source code instructions (instructions) may be generated by disassembling the binary. The binary may be disassembled at one or more sequential addresses starting at each of the plurality of potential address references.

Abstract: Distributed system simulation is enhanced by extending the simulation window. In a described implementation, the simulation window extension is facilitated with a slow message relaxation scheme. For example, especially when the simulation window is extended, slow unscheduled events can arrive at a logical process with a timestamp that is prior to (e.g., less than) the local time of a receiving logical process that is participating in a simulation. To ameliorate issues created by a slow unscheduled message and its corresponding slow unscheduled event, a current logical time of the receiving logical process is substituted for the original timestamp of the slow unscheduled event to transform it into a punctual unscheduled event.

Abstract: Distributed system simulation is enhanced by extending the simulation window. In a described implementation, the simulation window extension is facilitated with a slow message relaxation scheme. For example, especially when the simulation window is extended, slow unscheduled events can arrive at a logical process with a timestamp that is prior to (e.g., less than) the local time of a receiving logical process that is participating in a simulation. To ameliorate issues created by a slow unscheduled message and its corresponding slow unscheduled event, a current logical time of the receiving logical process is substituted for the original timestamp of the slow unscheduled event to transform it into a punctual unscheduled event.