Systems and methods for an interface with a widened interface-to-fabric shoreline between semiconductor circuits and a narrower interface-to-memory controller shoreline. The interface providing transitions from a first clock of a first circuit (e.g., field-programmable gate array (FPGA)), a second clock of a second circuit (e.g., high-bandwidth memory generation 2 (HBM2) stack, and a third clock of a physical layer of the second circuit. A first transfer between the first clock and the second clock may use a first set of first-in first-outs (FIFO) buffers, such as rate-matching FIFO buffers. A second transfer between the second clock and the third clock may use a second set of FIFO buffers, such as phase compensation FIFOs.

An apparatus is described. The apparatus includes a rank of memory chips to couple to a memory channel. The memory channel is characterized as having eight transfers of eight bits of raw data per burst access. The rank of memory chips has first, second and third X4 memory chips. The X4 memory chips conform to a JEDEC dual data rate (DDR) memory interface specification. The first and second X4 memory chips are to couple to an eight bit raw data portion of the memory channel's data bus. The third X4 memory chip to couple to an error correction coding (ECC) information portion of the memory channel's data bus.

Representative apparatus, method, and system embodiments are disclosed for configurable computing. A representative memory interface circuit comprises: a plurality of registers storing a plurality of tables, a state machine circuit, and a plurality of queues. The plurality of tables include a memory request table, a memory request identifier table, a memory response table, a memory data message table, and a memory response buffer. The state machine circuit is adapted to receive a load request, and in response, to obtain a first memory request identifier from the load request, to store the first memory request identifier in the memory request identifier table, to generate one or more memory load request data packets having the memory request identifier for transmission to the memory circuit, and to store load request information in the memory request table. The plurality of queues store one or more data packets for transmission.

Implementations described herein provide apparatus and methods for storing data in, and retrieving data from, storage buffer having an odd number of storage locations using minimal additional logic. A binary address symbol with a maximum value of one less than twice the number of storage locations is used to allow use of Gray code in transferring the storage location pointers between clock domains. An offset value is added to the binary address symbol to further facilitate use of Gray code. The Gray code is converted back to a binary symbol at the read side, the offset value is subtracted therefrom, and a pointer to a particular storage location is resolved.

A method for providing an asynchronous execution queue for accelerator hardware includes replacing a malloc operation in an execution queue to be sent to an accelerator with an asynchronous malloc operation that returns a unique reference pointer. Execution of the asynchronous malloc operation in the execution queue by the accelerator allocates a requested memory size and adds an entry to a look-up table accessible by the accelerator that maps the reference pointer to a corresponding memory address.

Representative apparatus, method, and system embodiments are disclosed for configurable computing. A representative memory interface circuit comprises: a plurality of registers storing a plurality of tables, a state machine circuit, and a plurality of queues. The plurality of tables include a memory request table, a memory request identifier table, a memory response table, a memory data message table, and a memory response buffer. The state machine circuit is adapted to receive a load request, and in response, to obtain a first memory request identifier from the load request, to store the first memory request identifier in the memory request identifier table, to generate one or more memory load request data packets having the memory request identifier for transmission to the memory circuit, and to store load request information in the memory request table. The plurality of queues store one or more data packets for transmission.

A method receives an inbound request to be processed based on multiple outbound service invocations of multiple outbound services. The method accesses expected response times for the inbound request for each of the multiple outbound services. The method determines which one or more of the multiple outbound services to invoke asynchronously and which one or more of the multiple outbound services to invoke synchronously based on the expected response times for the inbound request for each of the multiple outbound services. The method invokes asynchronously the one or more of the multiple outbound services determined to be invoked asynchronously, invokes synchronously the one or more of the multiple outbound services determined to be invoked synchronously.

Some embodiments include apparatuses and methods of using the apparatuses. One of the apparatuses includes first-in first-out (FIFO) cells included in an asynchronous FIFO unit and first and second circuits included in the asynchronous FIFO unit. The first circuit provides first information based on a value of a first bit from each of the FIFO cells in order to select one of the FIFO cells to be a selected FIFO cell for storing data information in the selected FIFO cell. The second circuit provides information based on a value of a second bit from each of the FIFO cells in order to select one of the FIFO cells to be a selected FIFO cell for reading data information from the selected FIFO cell.

An apparatus is described. The apparatus includes a rank of memory chips to couple to a memory channel. The memory channel is characterized as having eight transfers of eight bits of raw data per burst access. The rank of memory chips has first, second and third X4 memory chips. The X4 memory chips conform to a JEDEC dual data rate (DDR) memory interface specification. The first and second X4 memory chips are to couple to an eight bit raw data portion of the memory channel's data bus. The third X4 memory chip to couple to an error correction coding (ECC) information portion of the memory channel's data bus.

A method for stack timing adjustment for serial communications is provided. The method includes receiving a USB communication, decoding the USB communication into UART frames, and adjusting the timing of the UART frames according to a serial protocol.