Abstract: The present disclosure provides systems and methods that provide improved access to content items in situations involving suboptimal network conditions. In particular, when a user attempts but fails to access a content item, a computing system can enable the user to access an alternative version of the content item that has been modified to account for the network conditions, thereby improving the accessibility and latency of content items for devices that are suffering from slow and/or unreliable network connections.

Abstract: The present disclosure provides systems and methods that provide improved access to content items in situations involving suboptimal network conditions. In particular, when a user attempts but fails to access a content item, a computing system can enable the user to access an alternative version of the content item that has been modified to account for the network conditions, thereby improving the accessibility and latency of content items for devices that are suffering from slow and/or unreliable network connections.

Abstract: Methods, systems and computer-readable storage mediums encoded with computer programs executed by one or more processors for determining cacheability of a webpage are disclosed. In an embodiment, a request for a webpage is received. A change rate associated with the webpage is determined. A cacheability determination is made as to whether the cached version of the webpage is to be provided responsive to the request based on a cached timestamp of the cached version being more recent than the change rate being subtracted from one of a current time or time at which the request was received. The cached version of the webpage is provided responsive to the request based on the cacheability determination.

Abstract: An embodiment of a computing system is configured to process data using a multithreaded SIMD architecture that includes heterogeneous processing engines to execute a program. The program is constructed of various program instructions. A first type of the program instructions can only be executed by a first type of processing engine and a second type of program instructions can only be executed by a second type of processing engine. A third type of program instructions can be executed by the first and the second type of processing engines. An instruction dispatcher is configured to identify and remove program instruction execution conflicts for the heterogeneous processing engines to improve instruction execution throughput.

Abstract: One embodiment of the present invention sets forth a technique for ensuring cache access instructions are scheduled for execution in a multi-threaded system to improve cache locality and system performance. A credit-based technique may be used to control instruction by instruction scheduling for each warp in a group so that the group of warps is processed uniformly. A credit is computed for each warp and the credit contributes to a weight for each warp. The weight is used to select instructions for the warps that are issued for execution.

Abstract: Methods, systems and computer-readable storage mediums encoded with computer programs executed by one or more processors for transitioning between remotely cached and live versions of a webpage are disclosed. A primary webpage including a link to a secondary webpage is received. A cached version of the secondary webpage is requested prior to receiving a selection of the link. The cached version of the secondary webpage is provided responsive to the selection of the link. A live version of the secondary webpage is requested in parallel with the providing of the cached version. The cached version of the secondary webpage is transitioned to the live version. Upon a completion of the transitioning, the live version of the secondary webpage is provided.

Abstract: One embodiment of the present invention sets forth a technique for ensuring cache access instructions are scheduled for execution in a multi-threaded system to improve cache locality and system performance. A credit-based technique may be used to control instruction by instruction scheduling for each warp in a group so that the group of warps is processed uniformly. A credit is computed for each warp and the credit contributes to a weight for each warp. The weight is used to select instructions for the warps that are issued for execution.