Abstract: There is described a method of shading a group of pixels in a fragment shader in a raster graphics pipeline. At least one first pilot pixel of the group of pixels is shaded under a first precision. At least one second pilot pixel of the group of pixels is shaded under a second precision. An error value representing a difference between the first and second pilot pixels is calculated. At least one other pixel of the group of pixels is shaded under the first precision if the error value is greater than an error threshold. The at least one other pixel is shaded under the second precision if the error value is smaller than the error threshold.

Abstract: The present disclosure provides methods, systems and apparatus for handling control flow structures in data-parallel architectures. According to a first aspect, a method is provided. The method includes receiving, by a processing unit (PU), a program for execution. The method further includes applying, by the PU, a branching solution to the program to obtain data on control flow structures of the program. The method further includes determining, by the PU and based at least in part on the obtained data, one or more control flow structures of the program to predicate. The method further includes applying, by the PU, predication to the one or more control flow structures of the program.

Abstract: There is described a method of shading a group of pixels in a fragment shader in a raster graphics pipeline. At least one first pilot pixel of the group of pixels is shaded under a first precision. At least one second pilot pixel of the group of pixels is shaded under a second precision. An error value representing a difference between the first and second pilot pixels is calculated. At least one other pixel of the group of pixels is shaded under the first precision if the error value is greater than an error threshold. The at least one other pixel is shaded under the second precision if the error value is smaller than the error threshold.

Abstract: Systems and methods are described for extending a live range for a virtual scalar register during compiling of a program, comprising: receiving an intermediate representation (IR) of a source code configured for implementing single-instruction-multiple-thread (SIMT) execution, the IR representing the source code as control flow graph including a plurality of basic blocks (BB); and when a virtual scalar register defined in a first BB of the IR is last used in a second BB of the IR that is a divergent BB, modifying the IR to extend the live range of the virtual scalar register.

Abstract: The disclosed systems, structures, and methods are directed to optimizing address calculations in a computer. This is achieved in a compiler that identifies an address calculation in code that is being compiled and transforms the code by splitting the address calculation into a first portion in which an offset is determined and a second portion, in which the offset is combined with a base pointer to generate an address. The address and the base pointer have a first bit-length, and the offset has a second bit-length shorter than the first bit-length. The offset is determined using an operation performed at the second bit-length. In some implementations the first bit-length is 64 bits and the second bit-length is 32 bits.

Abstract: Systems and methods for network traffic metering credit distribution and packet processing in a network device having multiple processing units are provided. According to an embodiment, management of multiple meters is distributed among multiple processing units of a network device. Each meter is implemented in a form of a master entry and a slave entry. Responsive to receipt by one of the processing units of a packet subject to rate-limiting by a meter, an action to be taken on the packet is made with reference to a slave entry managed by the processing unit based on available credit of the slave entry. When the action indicates the packet is to be passed: (i) credits associated with passing the packet are deducted from the available credit; and (ii) the packet is passed to a subsequent stage of packet processing; otherwise, the packet is dropped.

Abstract: The disclosed systems, structures, and methods are directed to optimizing memory access to constants in heterogeneous parallel computers, including systems that support OpenCL. This is achieved in an optimizing compiler that transforms program scope constants and constants at the outermost scope of kernels into implicit constant pointer arguments. The optimizing compiler also attempts to determine access patterns for constants at compile-time and places the constants in a variety of memory types available in a compute device architecture based on these access patterns.

Abstract: Systems and methods for network traffic metering credit distribution and packet processing in a network device having multiple processing units are provided. According to an embodiment, management of multiple meters is distributed among multiple processing units of a network device. Each meter is implemented in a form of a master entry and a slave entry. Responsive to receipt by one of the processing units of a packet subject to rate-limiting by a meter, an action to be taken on the packet is made with reference to a slave entry managed by the processing unit based on available credit of the slave entry. When the action indicates the packet is to be passed: (i) credits associated with passing the packet are deducted from the available credit; and (ii) the packet is passed to a subsequent stage of packet processing; otherwise, the packet is dropped.

Abstract: Systems and methods are described for extending a live range for a virtual scalar register during compiling of a program, comprising: receiving an intermediate representation (IR) of a source code configured for implementing single-instruction-multiple-thread (SIMT) execution, the IR representing the source code as control flow graph including a plurality of basic blocks (BB); and when a virtual scalar register defined in a first BB of the IR is last used in a second BB of the IR that is a divergent BB, modifying the IR to extend the live range of the virtual scalar register.

Abstract: The disclosed systems, structures, and methods are directed to optimizing memory access to constants in heterogeneous parallel computers, including systems that support OpenCL. This is achieved in an optimizing compiler that transforms program scope constants and constants at the outermost scope of kernels into implicit constant pointer arguments. The optimizing compiler also attempts to determine access patterns for constants at compile-time and places the constants in a variety of memory types available in a compute device architecture based on these access patterns.

Abstract: Methods and systems for a more efficient transmission of network traffic are provided. According to one embodiment, presence of outbound payload data, distributed across a first and second payload buffer, within a user memory space of a network device that has been generated by a user process is determined by a bus/memory interface or a network interface unit. The payload data is fetched by performing direct virtual memory addressing of the user memory space including mapping virtual addresses of the payload buffers to corresponding physical addresses, including: (i) when the payload buffers are noncontiguous, then retrieving the outbound payload data with reference to multiple buffer descriptors having starting virtual addresses of the payload buffers and (ii) when they are contiguous, then retrieving the outbound payload data with reference to a single buffer descriptor. The outbound payload data is then segmented across one or more TCP packets.

Abstract: Methods and systems for a more efficient transmission of network traffic are provided. According to one embodiment, presence of outbound payload data, distributed across a first and second payload buffer, within a user memory space of a network device that has been generated by a user process is determined by a bus/memory interface or a network interface unit. The payload data is fetched by performing direct virtual memory addressing of the user memory space including mapping virtual addresses of the payload buffers to corresponding physical addresses, including: (i) when the payload buffers are noncontiguous, then retrieving the outbound payload data with reference to multiple buffer descriptors having starting virtual addresses of the payload buffers and (ii) when they are contiguous, then retrieving the outbound payload data with reference to a single buffer descriptor. The outbound payload data is then segmented across one or more TCP packets.

Abstract: Methods and systems for a more efficient transmission of network traffic are provided. According to one embodiment, presence of outbound payload data, distributed across a first and second payload buffer, within a user memory space of a network device that has been generated by a user process is determined by a bus/memory interface or a network interface unit. The payload data is fetched by performing direct virtual memory addressing of the user memory space including mapping virtual addresses of the payload buffers to corresponding physical addresses, including: (i) when the payload buffers are noncontiguous, then retrieving the outbound payload data with reference to multiple buffer descriptors having starting virtual addresses of the payload buffers and (ii) when they are contiguous, then retrieving the outbound payload data with reference to a single buffer descriptor. The outbound payload data is then segmented across one or more TCP packets.

Abstract: Methods and systems for a more efficient transmission of network traffic are provided. According to one embodiment, presence of outbound payload data, distributed across a first and second payload buffer, within a user memory space of a network device that has been generated by a user process is determined by a bus/memory interface or a network interface unit. The payload data is fetched by performing direct virtual memory addressing of the user memory space including mapping virtual addresses of the payload buffers to corresponding physical addresses, including: (i) when the payload buffers are noncontiguous, then retrieving the outbound payload data with reference to multiple buffer descriptors having starting virtual addresses of the payload buffers and (ii) when they are contiguous, then retrieving the outbound payload data with reference to a single buffer descriptor. The outbound payload data is then segmented across one or more TCP packets.

Abstract: Methods and systems for a more efficient transmission of network traffic are provided. According to one embodiment, payload data originated by a user process running on a host processor of a network device is fetched by an interface of the network device by performing direct virtual memory addressing of a user memory space of a system memory of the network device on behalf of a network interface unit of the network device. The direct virtual memory addressing maps physical addresses of various portions of the payload data to corresponding virtual addresses. The payload data is segmented by the network interface unit across one or more packets.

Abstract: Methods and systems for a more efficient transmission of network traffic are provided. According to one embodiment, payload data originated by a user process running on a host processor of a network device is fetched by an interface of the network device by performing direct virtual memory addressing of a user memory space of a system memory of the network device on behalf of a network interface unit of the network device. The direct virtual memory addressing maps physical addresses of various portions of the payload data to corresponding virtual addresses. The payload data is segmented by the network interface unit across one or more packets.

Abstract: A compiler method for exploiting data value locality for computation reuse. When a code region having single entry and exit points and in which a potential computation reuse opportunity exists is identified during runtime, a helper thread is created separate from the master thread. One of the helper thread and master thread performs a computation specified in the code region, and the other of the helper thread and master thread looks up a value of the computation previously executed and stored in a lookup table. If the value of the computation previously executed is located in the lookup table, the other thread retrieves the value from the table, and ignores the computation performed by the thread. If the value of the computation is not located, the other thread obtains a result of the computation performed by the thread and stores the result in the lookup table for future computation reuse.

Abstract: Methods and systems for a more efficient transmission of network traffic are provided. According to one embodiment, payload data originated by a user process running on a host processor of a network device is fetched by an interface of the network device by performing direct virtual memory addressing of a user memory space of a system memory of the network device on behalf of a network interface unit of the network device. The direct virtual memory addressing maps physical addresses of various portions of the payload data to corresponding virtual addresses. The payload data is segmented by the network interface unit across one or more packets.

Abstract: Methods and systems for a more efficient transmission of network traffic are provided. According to one embodiment, payload data originated by a user process running on a host processor of the computer system is fetched by an interface of the computer system by performing direct virtual memory addressing of a user memory space of a system memory of the computer system on behalf of a network processor of the computer system. The direct virtual memory addressing maps a physical address of the payload data to a virtual address. The payload data is segmented by the network processor across one or more packets.

Abstract: Methods and systems for a more efficient transmission of network traffic are provided. According to one embodiment, payload data originated by a user process running on a host processor of a network device is fetched by an interface of the network device by performing direct virtual memory addressing of a user memory space of a system memory of the network device on behalf of a network interface unit of the network device. The direct virtual memory addressing maps physical addresses of various portions of the payload data to corresponding virtual addresses. The payload data is segmented by the network interface unit across one or more packets.