Abstract: Techniques are disclosed relating to private memory management using a mapping thread, which may be persistent. In some embodiments, a graphics processor is configured to generate a pool of private memory pages for a set of graphics work that includes multiple threads. The processor may maintain a translation table configured to map private memory addresses to virtual addresses based on identifiers of the threads. The processor may execute a mapping thread to receive a request to allocate a private memory page for a requesting thread, select a private memory page from the pool in response to the request, and map the selected page in the translation table for the requesting. The processor may then execute one or more instructions of the requesting thread to access a private memory space, wherein the execution includes translation of a private memory address to a virtual address based on the mapped page in the translation table.

Abstract: Techniques are disclosed relating to private memory management using a mapping thread, which may be persistent. In some embodiments, a graphics processor is configured to generate a pool of private memory pages for a set of graphics work that includes multiple threads. The processor may maintain a translation table configured to map private memory addresses to virtual addresses based on identifiers of the threads. The processor may execute a mapping thread to receive a request to allocate a private memory page for a requesting thread, select a private memory page from the pool in response to the request, and map the selected page in the translation table for the requesting. The processor may then execute one or more instructions of the requesting thread to access a private memory space, wherein the execution includes translation of a private memory address to a virtual address based on the mapped page in the translation table.

Abstract: Techniques are disclosed relating to private memory management using a mapping thread, which may be persistent. In some embodiments, a graphics processor is configured to generate a pool of private memory pages for a set of graphics work that includes multiple threads. The processor may maintain a translation table configured to map private memory addresses to virtual addresses based on identifiers of the threads. The processor may execute a mapping thread to receive a request to allocate a private memory page for a requesting thread, select a private memory page from the pool in response to the request, and map the selected page in the translation table for the requesting. The processor may then execute one or more instructions of the requesting thread to access a private memory space, wherein the execution includes translation of a private memory address to a virtual address based on the mapped page in the translation table.

Abstract: Techniques are disclosed relating to arbitration of requests to access a register file. In one embodiment, an apparatus includes a write queue and a register file that includes multiple entries. In one embodiment, the apparatus is configured to select a request from a plurality of requests based on a plurality of request characteristics, and write data from the accepted request into a write queue. In one embodiment, the request characteristics include: whether a request is a last request from an agent for a given register file entry and whether the request finishes a previous request. In one embodiment, a final arbiter is configured to select among requests from the write queue, a read queue, and multiple execution pipelines to access banks of the register file in a given cycle.

Abstract: Techniques are disclosed relating to arbitration of requests to access a register file. In one embodiment, an apparatus includes a write queue and a register file that includes multiple entries. In one embodiment, the apparatus is configured to select a request from a plurality of requests based on a plurality of request characteristics, and write data from the accepted request into a write queue. In one embodiment, the request characteristics include: whether a request is a last request from an agent for a given register file entry and whether the request finishes a previous request. In one embodiment, a final arbiter is configured to select among requests from the write queue, a read queue, and multiple execution pipelines to access banks of the register file in a given cycle.

Abstract: A processor includes a translation lookaside buffer (TLB) including a data array and a compare unit. The data array includes a number of entries each configured to store a respective translated physical address. In response to a read access to a given entry of the TLB, the data array is configured to output within a particular clock cycle, the respective translated physical address stored in the given entry. In addition the compare unit may be configured to compare the respective translated physical address output by the data array with a number of additional addresses. The compare unit may also be configured to provide a hit indication for each of the additional addresses within the particular clock cycle.

Abstract: An Ethernet extension device is provided for metro or last mile Ethernet service via twisted pairs as opposed to fiber optics. The Ethernet extension device is implemented as a plug-in extension for existing infrastructure (e.g., in a standard electrical wall box or Type-200™ Mechanics card) that employs lighting and power cross protection required by the telephone companies for Ethernet connectivity to the telephone network (e.g., for connection between a user's building and a telephone company building over existing outdoor telephone cables).

Abstract: Techniques are disclosed relating to request arbitration between a plurality of master circuits and a plurality of target circuits. In one embodiment, an apparatus includes an arbitration unit coupled to a plurality of request queues for a target circuit. Each request queue is configured to store requests generated by a respective one of a plurality of master circuits. The arbitration unit is configured to arbitrate between requests in the plurality of request queues based on information indicative of an ordering in which requests were submitted to the plurality of request queues by master circuits. In some embodiments, each of the plurality of master circuits are configured to submit, with each request to the target circuit, an indication specifying that a request has been submitted, and the arbitration unit is configured to determine the ordering in which requested were submitted based on the submitted indications.

Abstract: Controllable arrays in a memory may be activated and deactivated independently. In one embodiment, a processor may include a memory. The memory may be a de-interleaved memory with independently selectable arrays. Based on an address bit of an address used to access data from the memory, a wordline and downstream components may be activated while another wordline and downstream components may be deactivated.

Abstract: Sense amplifiers in a memory may be activated and deactivated. In one embodiment, a processor may include a memory. The memory may include a number of sense amplifiers. Based on a late arriving address bit of an address used to access data from the memory, a sense amplifier may be activated while another sense amplifier may be deactivated.

Abstract: Techniques are disclosed relating to buffer circuits. In one embodiment, a buffer circuit is disclosed that includes memory unit and an output register. The memory unit is configured to store a plurality of buffer entries and a first pointer to a current one of the plurality of buffer entries. The output register is coupled to an output of the memory unit. The buffer circuit is configured to perform a read operation by outputting a current value of the output register and storing a value of the current buffer entry in the output register. The buffer circuit is configured to update the first pointer in response to the read operation.

Abstract: Techniques are disclosed relating to request arbitration between a plurality of master circuits and a plurality of target circuits. In one embodiment, an apparatus includes an arbitration unit coupled to a plurality of request queues for a target circuit. Each request queue is configured to store requests generated by a respective one of a plurality of master circuits. The arbitration unit is configured to arbitrate between requests in the plurality of request queues based on information indicative of an ordering in which requests were submitted to the plurality of request queues by master circuits. In some embodiments, each of the plurality of master circuits are configured to submit, with each request to the target circuit, an indication specifying that a request has been submitted, and the arbitration unit is configured to determine the ordering in which requested were submitted based on the submitted indications.

Abstract: A dual port memory includes a mechanism for preventing collisions. The memory includes dual port bit cells arranged in rows and columns and each bit cell stores a data bit. The memory also includes a wordline unit that may provide a respective write wordline signal and a respective read wordline signal to each row of bit cells. The wordline unit may also selectively inhibit the read wordline signal for a given row based upon address information that is indicative of whether a write operation will be performed to the given row.

Abstract: A dual port memory includes a mechanism for preventing collisions. The memory includes dual port bit cells arranged in rows and columns and each bit cell stores a data bit. The memory also includes a wordline unit that may provide a respective write wordline signal and a respective read wordline signal to each row of bit cells. The wordline unit may also selectively inhibit the read wordline signal for a given row based upon address information that is indicative of whether a write operation will be performed to the given row.

Abstract: Controllable arrays in a memory may be activated and deactivated independently. In one embodiment, a processor may include a memory. The memory may be a de-interleaved memory with independently selectable arrays. Based on an address bit of an address used to access data from the memory, a wordline and downstream components may be activated while another wordline and downstream components may be deactivated.

Abstract: Sense amplifiers in a memory may be activated and deactivated. In one embodiment, a processor may include a memory. The memory may include a number of sense amplifiers. Based on a late arriving address bit of an address used to access data from the memory, a sense amplifier may be activated while another sense amplifier may be deactivated.

Abstract: A processor includes a translation lookaside buffer (TLB) including a data array and a compare unit. The data array includes a number of entries each configured to store a respective translated physical address. In response to a read access to a given entry of the TLB, the data array is configured to output within a particular clock cycle, the respective translated physical address stored in the given entry. In addition the compare unit may be configured to compare the respective translated physical address output by the data array with a number of additional addresses. The compare unit may also be configured to provide a hit indication for each of the additional addresses within the particular clock cycle.

Abstract: An Ethernet extension device is provided for metro or last mile Ethernet service via twisted pairs as opposed to fiber optics. The Ethernet extension device is implemented as a plug-in extension for existing infrastructure (e.g., in a standard electrical wall box or Type-200™ Mechanics card) that employs lighting and power cross protection required by the telephone companies for Ethernet connectivity to the telephone network (e.g., for connection between a user's building and a telephone company building over existing outdoor telephone cables).

Abstract: An Ethernet extension device is provided for metro or last mile Ethernet service via twisted pairs as opposed to fiber optics. The Ethernet extension device is implemented as a plug-in extension for existing infrastructure (e.g., in a standard electrical wall box or Type-200™ Mechanics card) that employs lighting and power cross protection required by the telephone companies for Ethernet connectivity to the telephone network (e.g., for connection between a user's building and a telephone company building over existing outdoor telephone cables).

Abstract: Systems and methods are disclosed for locking code in cache. A processor comprises a cache and a cache controller. The cache is configured to store a temporary copy of code residing in main memory. Also, the cache is divided into a number of cache ways, where each cache way is further divided into a number of cache way portions. The cache controller is configured to utilize a first signal and a second signal. The first signal designates one of the cache ways as a partial cache way and the second signal defines which ones of the cache way portions of the partial cache way are to be locked.