Patent Application
20220121612
Embodiments of the disclosure relate generally to chiplet-based electronic systems and to communications in such systems.
Chiplets are an emerging technique for integrating various processing functionality. Generally, a chiplet system is made up of discrete chips (e.g., integrated circuits (ICs) on different substrate or die) that are integrated on an interposer and packaged together. This arrangement is distinct from single chips (e.g., ICs) that contain distinct device blocks (e.g., intellectual property (IP) blocks) on one substrate (e.g., single die), such as a system-on-a-chip (SoC), or discretely packaged devices integrated on a board. In general, chiplets provide better performance (e.g., lower power consumption, reduced latency, etc.) than discretely packaged devices, and chiplets provide greater production benefits than single die chips. These production benefits can include higher yields or reduced development costs and time.
Chiplet systems are generally made up of one or more application chiplets and support chiplets. Here, the distinction between application and support chiplets is simply a reference to the likely design scenarios for the chiplet system. Thus, for example, a synthetic vision chiplet system can include an application chiplet to produce the synthetic vision output along with support chiplets, such as a memory controller chiplet, sensor interface chiplet, or communication chiplet. In a typical use case, the synthetic vision designer can design the application chiplet and source the support chiplets from other parties. Thus, the design expenditure (e.g., in terms of time or complexity) is reduced by avoiding the design and production of functionality embodied in the support chiplets. Chiplets also support the tight integration of IP blocks that can otherwise be difficult, such as those using different feature sizes. Thus, for example, devices designed during a previous fabrication generation with larger feature sizes, or those devices in which the feature size is optimized for power, speed, or heat generation—as can happen with sensors—can be more easily integrated with devices having different feature sizes. Additionally, by reducing the overall size of the die, the yield for chiplets tends to be higher than that of more complex, single die devices.
To easily identify the discussion of any particular element or act, the most significant digit or digits in a reference number refer to the figure number in which that element is first introduced.
A variety of communications protocols can be used to communicate between a host and a memory device in a system, such as a chiplet system. Examples of such protocols can include the Open NAND Flash Interface (ONFi), eMMC, UFS, or Serial Peripheral Interface (SPI), among others. These protocols generally enable the host, primary device, or controller, to communicate commands—such as write (e.g., program), read, request the status of a command, request the status of the memory device, start or perform housekeeping operations such as intra-memory transfers, garbage collection, etc. —with the memory device. Generally, these protocols restrict initiation of communication to the host. That is, the host makes a request and the memory device responds. In some examples, the memory device can issue an exception (e.g., interrupt) that is designed to prompt the host to make a request for the status of the operation subject to the exception. In some case, the memory device can issue a signal that indicates that the memory device is busy or that the host must wait to issue further commands for some defined time period or until the memory device otherwise indicates that it has completed pending operations and is available to receive and respond to further commands.
To facilitate communication among chiplets in a system, chiplets can include multiple input-output (I/O) channels (e.g., AIB channels), such as can be arranged in columns of channels at a periphery of the chiplets. The I/O channels can be configured in a manner dependent on the particular design or system objective. For example, the I/O channels of chiplets can be configured as receive (RX) channels, transmit (TX) channels, or a mix of RX/TX channels.
In an example, one or more chiplets of a system can include initialization logic circuitry to advance a chiplet I/O interface through various stages or phases of initialization. In some examples, chiplets can include a communication interface (e.g., a serial peripheral interface or SPI) and configuration data can be communicated among the chiplets using the communication interface. In some examples, the communication interface can be implemented using auxiliary channels (AUX) of the I/O channels and the configuration data can be communicated using out-of-band signaling.
In an example, initialization logic circuitry can be configured to advance initialization of a chiplet interface sequentially through the interface layers starting with a lowest interface layer (e.g., the physical layer). The initialization can advance through multiple initialization phases with one interface layer initialized during each phase by writing initialization data to the chiplet I/O channels during each initialization phase.
Not all chiplet designs may have or use the same type of initialization. For example, individual chiplets may support only a hardware-based initialization option for the I/O channels or may support only a software-based initialization option for the I/O channels. Because a chiplet-based system can include different chiplet designs mixed in the same system, and without a standard method to support both hardware-based and software-based approaches, each system would need an individual, ad hoc hardware-based or software-based initialization method. This could result in some I/O channels of the chiplets not being interoperable with other I/O channels of the chiplets.
In an example, a standardized method of initialization of the I/O channels of the chiplets can allow all chiplets of a multi-chiplet system and I/O interface to be interoperable within a single system. The standardized method can be used for both hardware-based and software-based initialization mechanisms to guarantee interoperability of the interconnected chiplet I/O channels. In an example, the initialization methods can include or use communication among chiplets using an SPI bus. The SPI bus can be used to provide communication between at least a primary device and a secondary device. As used in this description, a primary device, such as a host, may manage or control communications with or operations of one or more secondary devices. The relationship between primary and secondary devices may be an asymmetric one that has sometimes been referred to in the art using the antiquated terms master and slave. Such relationships may also be referred to as parent/child, supervisor/worker, controller/peripheral, or the like.
An issue with traditional SPI communications arises from a bus configuration that can include respective chip select (CS) channels or ports for each device in the network. That is, a traditional SPI bus can include a CS channel for each secondary device such that respective different CS signals can be provided to each secondary device. A primary device configured to use a traditional SPI bus can include a number of CS channel output ports that corresponds to the number of secondary devices on the bus so that the primary device can issue respective chip select signals to each secondary device.
To address the issue, an SPI communication system (or an analogous communication system) can be configured to include or use respective static identifiers to address or identify each secondary device that is in communication with a particular primary device. In an example, each secondary device can include or use a unique identifier, such as can be set using a static strapping port at each device. The port can be statically strapped to a unique value for each different secondary device on the SPI bus or interface to distinguish each device from others on the same bus.
In an example of a system that includes or uses static peripheral identifiers, a chip select port can be eliminated (or unused) at each secondary device. In this example, a particular secondary device can determine whether to attend to or disregard a particular message or instruction from a primary device based on the identifier included in the message from the primary device.
In another example, a group of chip select lines or channels in an SPI bus can be consolidated to a single multi-drop chip select channel. That is, a single chip select channel can be provided to concurrently enable (or disable) one or multiple secondary devices, and a particular secondary device can determine whether to attend to or disregard a particular message from a primary device based on the identifier included in the message. In this example, SPI-based communication can optionally be aborted by the primary device by changing a state of the signal on the single chip select channel.
The application chiplet 106 is illustrated as including a network-on-chip (NOC 108) to support an inter-chiplet communications network, or chiplet network 122. In example embodiments, NOC 108 may be included on the application chiplet 106. In some examples, NOC 108 may be defined in response to selected support chiplets (e.g., the host interface chiplet 112, memory controller chiplet 114, or memory device chiplet 118) thus enabling a designer to select an appropriate number or chiplet network connections or switches for the NOC 108. In an example, the NOC 108 can be located on a separate chiplet, or even within the interposer 120. In examples as discussed herein, the NOC 108 implements an inter-chiplet communications network as a chiplet protocol interface (CPI) network.
The CPI is a packet-based network that supports virtual channels to enable a flexible and high-speed interaction between chiplets. CPI enables bridging from intra-chiplet networks to the chiplet network 122. For example, the Advanced eXtensible Interface (AXI) is a widely used specification to design intra-chip communications. AXI specifications, however, cover a great variety of physical design options, such as the number of physical channels, signal timing, power, etc. Within a single chip, these options are generally selected to meet design goals, such as power consumption, speed, etc. However, to achieve the flexibility of the chiplet system, an adapter, such as CPI, is used to interface between the various AXI design options that can be implemented in the various chiplets. By enabling a physical channel to virtual channel mapping and encapsulating time-based signaling with a packetized protocol, CPI successfully bridges intra-chiplet networks across the chiplet network 122.
CPI can use a variety of different physical layers to transmit packets. The physical layer can include simple conductive connections, or can include drivers to increase the voltage, or otherwise facilitate transmitting the signals over longer distances. An example of one such physical layer can include the Advanced Interface Bus (AIB), which in various examples, can be implemented in the interposer 120. AIB transmits and receives data using source synchronous data transfers with a forwarded clock. Packets are transferred across the AIB at single data rate (SDR) or dual data rate (DDR) with respect to the transmitted clock. Various channel widths are supported by AIB. AIB channel widths are in multiples of 20 bits when operated in SDR mode (20, 40, 60, . . . ), and multiples of 40 bits for DDR mode: (40, 80, 120, . . . ). The AIB channel width includes both transmit and receive signals. The channel can be configured to have a symmetrical number of transmit (TX) and receive (RX) input/outputs (I/Os), or have a non-symmetrical number of transmitters and receivers (e.g., either all transmitters or all receivers). The channel can act as an AIB controller or peripheral depending on which chiplet provides the controller clock. AIB I/O cells support three clocking modes: asynchronous (i.e. non-clocked), SDR, and DDR. In various examples, the non-clocked mode is used for clocks and some control signals. The SDR mode can use dedicated SDR only I/O cells, or dual use SDR/DDR I/O cells.
In an example, CPI packet protocols (e.g., point-to-point or routable) can use symmetrical receive and transmit I/O cells within an AIB channel. The CPI streaming protocol allows more flexible use of the AIB I/O cells. In an example, an AIB channel for streaming mode can configure the I/O cells as all TX, all RX, or half RX and half RX. CPI packet protocols can use an AIB channel in either SDR or DDR operation modes. In an example, the AIB channel is configured in increments of 80 I/O cells (i.e. 40 TX and 40 RX) for SDR mode and 40 I/O cells for DDR mode. The CPI streaming protocol can use an AIB channel in either SDR or DDR operation modes. Here, in an example, the AIB channel is in increments of 40 I/O cells for both SDR and DDR modes. In an example, each AIB channel is assigned a unique interface identifier. The identifier is used during CPI reset and initialization to determine paired AIB channels across adjacent chiplets. In an example, the interface identifier is a 20-bit value comprising a seven-bit chiplet identifier, a seven-bit column identifier, and a six-bit link identifier. The AIB physical layer transmits the interface identifier using an AIB out-of-band shift register. The 20-bit interface identifier is transferred in both directions across an AIB interface using bits 32-51 of the shift registers.
AIB defines a stacked set of AIB channels as an AIB channel column. An AIB channel column has some number of AIB channels, plus an auxiliary channel. The auxiliary channel contains signals used for AIB initialization. All AIB channels (other than the auxiliary channel) within a column are of the same configuration (e.g., all TX, all RX, or half TX and half RX, as well as having the same number of data I/O signals). In an example, AIB channels are numbered in continuous increasing order starting with the AIB channel adjacent to the AUX channel. The AIB channel adjacent to the AUX is defined to be AIB channel zero.
Generally, CPI interfaces on individual chiplets can include serialization-deserialization (SERDES) hardware. SERDES interconnects work well for scenarios in which high-speed signaling with low signal count are desirable. SERDES, however, can result in additional power consumption and longer latencies for multiplexing and demultiplexing, error detection or correction (e.g., using block level cyclic redundancy checking (CRC)), link-level retry, or forward error correction. However, when low latency or energy consumption is a primary concern for ultra-short reach, chiplet-to-chiplet interconnects, a parallel interface with clock rates that allow data transfer with minimal latency may be utilized. CPI includes elements to minimize both latency and energy consumption in these ultra-short reach chiplet interconnects.
For flow control, CPI employs a credit-based technique. A recipient, such as the application chiplet 106, provides a sender, such as the memory controller chiplet 114, with credits that represent available buffers. In an example, a CPI recipient includes a buffer for each virtual channel for a given time-unit of transmission. Thus, if the CPI recipient supports five messages in time and a single virtual channel, the recipient has five buffers arranged in five rows (e.g., one row for each unit time). If four virtual channels are supported, then the recipient has twenty buffers arranged in five rows. Each buffer holds the payload of one CPI packet.
When the sender transmits to the recipient, the sender decrements the available credits based on the transmission. Once all credits for the recipient are consumed, the sender stops sending packets to the recipient. This ensures that the recipient always has an available buffer to store the transmission.
As the recipient processes received packets and frees buffers, the recipient communicates the available buffer space back to the sender. This credit return can then be used by the sender to allow transmitting of additional information.
Also illustrated is a chiplet mesh network 124 that uses a direct, chiplet-to-chiplet technique without the need for the NOC 108. The chiplet mesh network 124 can be implemented in CPI, or another chiplet-to-chiplet protocol. The chiplet mesh network 124 generally enables a pipeline of chiplets where one chiplet serves as the interface to the pipeline while other chiplets in the pipeline interface only with themselves.
Additionally, dedicated device interfaces, such as an SPI interface or one or more standard memory interfaces, such as the memory interface 116 (such as, for example, synchronous memory interfaces, such as DDR5, DDR6), can also be used to interconnect chiplets. Connection of a chiplet system or individual chiplets to external devices such as a larger system can be through a desired interface, for example, a PCIe interface. Such an external interface may be implemented, in some examples, through a host interface chiplet 112, which in the depicted examples, provides a PCIe interface external to the chiplet system 110. Such dedicated interfaces are generally employed when a convention or standard in the industry has converged on such an interface. The illustrated example of a Double Data Rate (DDR) interface 116 connecting the memory controller chiplet 114 to a dynamic random access memory (DRAM) memory device is an example of such an industry convention.
Of the variety of possible support chiplets, the memory controller chiplet 114 is likely present in the chiplet system 110 due to the near omnipresent use of storage for computer processing as well as sophisticated state-of-the-art memory devices. Thus, using a memory device chiplet 118 and memory controller chiplet 114 produced by others gives chiplet system designers access to robust products by sophisticated producers. Generally, the memory controller chiplet 114 provides a memory device specific interface to read, write, or erase data. Often, the memory controller chiplet 114 can provide additional features, such as error detection, error correction, maintenance operations, or atomic operation execution. For some types of memory, maintenance operations tend to be specific to the memory device chiplet 118, such as garbage collection in NAND flash or storage class memories, temperature adjustments (e.g., cross temperature management) in NAND flash memories. In an example, the maintenance operations can include logical-to-physical (L2P) mapping or management to provide a level of indirection between the physical and logical representation of data. In some flash memory configurations, for example, “managed NAND” devices, some or all of such management operations can be under control of a dedicated NAND memory controller coupled to multiple NAND memory die. In other types of memory, for example DRAM, some memory operations, such as refresh, may be controlled by a host processor or by a memory controller at some times, and at other times controlled by the DRAM memory device itself, or by logic associated with one or more DRAM devices, such as an interface chip (in some examples, a buffer). Such an interface/buffer may be utilized in some examples to redistribute and change the clock rate of signals between an interface and individual memory devices. In some examples, such an interface/buffer may incorporate additional control functionality.
Atomic operations are a data manipulation that, for example, may be performed by the memory controller chiplet 114. In other chiplet systems, the atomic operations may be performed by other chiplets. For example, an atomic operation of “increment” can be specified in a command by the application chiplet 106, the command including a memory address and possibly an increment value. Upon receiving the command, the memory controller chiplet 114 retrieves a number from the specified memory address, increments the number by the amount specified in the command, and stores the result. Upon a successful completion, the memory controller chiplet 114 provides an indication of a command success to the application chiplet 106. Atomic operations avoid transmitting the data across the chiplet network 122, resulting in lower latency execution of such commands.
Atomic operations can be classified as built-in atomics or programmable (e.g., custom) atomics. Built-in atomics are a finite set of operations that are immutably implemented in hardware. Programmable atomics are small programs that can run on a programmable atomic unit (PAU) (e.g., a custom atomic unit (CAU)) of the memory controller chiplet 114.
The memory device chiplet 118 can be, or include any combination of, volatile memory devices or non-volatile memories. Examples of volatile memory devices include, but are not limited to, random access memory (RAM) —such as DRAM, synchronous DRAM (SDRAM), graphics double data rate type 6 SDRAM (GDDR6 SDRAM), among others. Examples of non-volatile memory devices include, but are not limited to, negative-and-(NAND)-type flash memory, storage class memory (e.g., phase-change memory or memristor based technologies), ferroelectric RAM (FeRAM), among others. The illustrated example includes the memory device chiplet 118 as a chiplet, however, the memory device chiplet 118 can reside elsewhere, such as in a different package on the peripheral board 104. For many applications, multiple memory device chiplets may be provided. In some examples, these memory device chiplets may each implement one or multiple storage technologies. In some examples, a memory chiplet may include, multiple stacked memory die of different technologies, for example one or more SRAM devices stacked or otherwise in communication with one or more DRAM devices. Memory controller chiplet 114 may also serve to coordinate operations between multiple memory chiplets in the chiplet system 110; for example, to utilize one or more memory chiplets in one or more levels of cache storage, and to use one or more additional memory chiplets as main memory. Chiplet system 110 may also include multiple memory controllers, as may be used to provide memory control functionality for separate processors, sensors, networks, etc. A chiplet architecture, such as in the chiplet system 110 offers particular advantages in allowing adaptation to different memory storage technologies, and different memory interfaces, through updated chiplet configurations, without requiring redesign of the remainder of the system structure.
The off-die memory controller 206 is directly coupled to the off-die memory 230 (e.g., via a bus or other communication connection) to provide write operations and read operations to and from the off-die memory 230. In the depicted example, the off-die memory controller 206 is also coupled for output to the atomic and merge operations 220, and for input to the cache controller 204 (e.g., a memory side cache controller). In an example, the off-die memory controller 206 can be coupled to the off-die memory 230 using an SPI bus.
In an example, the off-die memory controller 206 (e.g., a memory controller for off-die memory) can include or comprise a portion of the SPI controller 232. The SPI controller 232 can be coupled to an SPI bus and configured to manage communication between the memory controller chiplet 114 and one or more other chiplets, such as other chiplets in the chiplet network 122 or the off-die memory 230. In an example, the memory controller chiplet 114 can use the SPI controller 232 to carry out initialization routines with various chiplets coupled to the memory controller chiplet 114. Once initialized, the memory controller chiplet 114 can continue to use SPI-based communications with the chiplets or can change to using other protocols or busses.
In an example, the SPI controller 232 or bus controller can be provided on the host interface chiplet 112, and the host interface chiplet 112 can use a PCIe interface to communicate outside of the chiplet system 110. A memory controller, such as the memory controller chiplet 114, can be an SPI memory device or SPI secondary device. The memory controller can, in turn, be configured to use another memory interface such as the memory interface 116.
In the example configuration, the cache controller 204 is directly coupled to the cache 202, and may be coupled to the network communication interface 208 for input (such as incoming read or write requests), and coupled for output to the off-die memory controller 206.
The network communication interface 208 includes a packet decoder 210, network input queues 212, a packet encoder 214, and network output queues 216 to support a packet-based chiplet network 122, such as CPI. The chiplet network 122 can provide packet routing between and among processors, memory controllers, hybrid threading processors, configurable processing circuits, or communication interfaces. In such a packet-based communication system, each packet typically includes destination and source addressing, along with any data payload or instruction. In an example, the chiplet network 122 can be implemented as a collection of crossbar switches having a folded clos configuration, or a mesh network providing for additional connections, depending upon the configuration.
In various examples, the chiplet network 122 can be part of an asynchronous switching fabric. Here, a data packet can be routed along any of various paths, such that the arrival of any selected data packet at an addressed destination can occur at any of multiple different times, depending upon the routing. Additionally, the chiplet network 122 can be implemented at least in part as a synchronous communication network, such as a synchronous mesh communication network. Both configurations of communication networks are contemplated for use in accordance with the present disclosure.
The memory controller chiplet 218 can receive a packet having, for example, a source address, a read request, and a physical address. In response, the off-die memory controller 206 or the cache controller 204 will read the data from the specified physical address (which can be in the off-die memory 230 or in the cache 202), and assemble a response packet to the source address containing the requested data. Similarly, the memory controller chiplet 218 can receive a packet having a source address, a write request, and a physical address. In response, the memory controller chiplet 218 will write the data to the specified physical address (which can be in the off-die memory 230 or in the cache 202), and assemble a response packet to the source address containing an acknowledgement that the data was stored to a memory.
Thus, the memory controller chiplet 218 can receive read and write requests via the chiplet network 122 and process the requests using the cache controller 204 interfacing with the cache 202, if possible. If the request cannot be handled by the cache controller 204, then the off-die memory controller 206 handles the request by communication with the off-die memory 230, the atomic and merge operations 220, or both. As noted above, one or more levels of cache may also be implemented in off-die memory 230 and in some such examples may be accessed directly by the cache controller 204. Data read by the off-die memory controller 206 can be cached in the cache 202 by the cache controller 204 for later use.
The atomic and merge operations 250 are coupled to receive (as input) the output of the off-die memory controller 220, and to provide output to the cache 210, the network communication interface 225, or directly to the chiplet network 280. The reset or memory hazard clear unit 224, write merge unit 222, and the built-in (e.g., predetermined) built-in atomic unit 226 can each be implemented as state machines with other combinational logic circuitry (such as adders, shifters, comparators, AND gates, OR gates, XOR gates, or any suitable combination thereof) or other logic circuitry. These components can also include one or more registers or buffers to store operand or other data. The PAU 228 can be implemented as one or more processor cores or control circuitry, and various state machines with other combinational logic circuitry or other logic circuitry, and can also include one or more registers, buffers, or memories to store addresses, executable instructions, operand and other data, or can be implemented as a processor.
The write merge unit 222 receives read data and request data, and merges the request data and read data to create a single unit having the read data and the source address to be used in the response or return data packet). The write merge unit 222 provides the merged data to the write port of the cache 202 (or, equivalently, to the cache controller 204 to write to the cache 202). Optionally, the write merge unit 222 provides the merged data to the network communication interface 208 to encode and prepare a response or return data packet for transmission on the chiplet network 122.
When the request data is for a built-in atomic operation, the built-in atomic unit 226 receives the request and reads data, either from the write merge unit 222 or directly from the off-die memory controller 206. The atomic operation is performed, and using the write merge unit 222, the resulting data is written to the cache 202, or provided to the network communication interface 208 to encode and prepare a response or return data packet for transmission on the chiplet network 122.
The built-in atomic unit 226 handles predefined atomic operations such as fetch-and-increment or compare-and-swap. In an example, these operations perform a simple read-modify-write operation to a single memory location of 32-bytes or less in size. Atomic memory operations are initiated from a request packet transmitted over the chiplet network 122. The request packet has a physical address, atomic operator type, operand size, and optionally up to 32-bytes of data. The atomic operation performs the read-modify-write to a cache memory line of the cache 202, filling the cache memory if necessary. The atomic operator response can be a simple completion response, or a response with up to 32-bytes of data. Example atomic memory operators include fetch-and-AND, fetch-and-OR, fetch-and-XOR, fetch-and-add, fetch-and-subtract, fetch-and-increment, fetch-and-decrement, fetch-and-minimum, fetch-and-maximum, fetch-and-swap, and compare-and-swap. In various example embodiments, 32-bit and 64-bit operations are supported, along with operations on 16 or 32 bytes of data. Methods disclosed herein are also compatible with hardware supporting larger or smaller operations and more or less data.
Built-in atomic operations can also involve requests for a “standard” atomic operation on the requested data, such as a comparatively simple, single cycle, integer atomics-such as fetch-and-increment or compare-and-swap-which will occur with the same throughput as a regular memory read or write operation not involving an atomic operation. For these operations, the cache controller 204 may generally reserve a cache line in the cache 202 by setting a hazard bit (in hardware), so that the cache line cannot be read by another process while it is in transition. The data is obtained from either the off-die memory 230 or the cache 202, and is provided to the built-in atomic unit 226 to perform the requested atomic operation. Following the atomic operation, in addition to providing the resulting data to the packet encoder 214 to encode outgoing data packets for transmission on the chiplet network 122, the built-in atomic unit 226 provides the resulting data to the write merge unit 222, which will also write the resulting data to the cache 202. Following the writing of the resulting data to the cache 202, any corresponding hazard bit which was set will be cleared by the memory hazard clear unit 224.
The PAU 228 enables high performance (high throughput and low latency) for programmable atomic operations (also referred to as “custom atomic operations”), comparable to the performance of built-in atomic operations. Rather than executing multiple memory accesses, in response to an atomic operation request designating a programmable atomic operation and a memory address, circuitry in the memory controller chiplet 218 transfers the atomic operation request to PAU 228 and sets a hazard bit stored in a memory hazard register corresponding to the memory address of the memory line used in the atomic operation, to ensure that no other operation (read, write, or atomic) is performed on that memory line, which hazard bit is then cleared upon completion of the atomic operation. Additional, direct data paths provided for the PAU 228 executing the programmable atomic operations allow for additional write operations without any limitations imposed by the bandwidth of the communication networks and without increasing any congestion of the communication networks.
The PAU 228 includes a multi-threaded processor, for example, such as a RISC-V ISA based multi-threaded processor having one or more processor cores, and further having an extended instruction set for executing programmable atomic operations. When provided with the extended instruction set for executing programmable atomic operations, the PAU 228 can be embodied as one or more hybrid threading processors. In some example embodiments, the PAU 228 provides barrel-style, round-robin instantaneous thread switching to maintain a high instruction-per-clock rate.
Programmable atomic operations can be performed by the PAU 228 involving requests for a programmable atomic operation on the requested data. A user can prepare programming code to provide such programmable atomic operations. For example, the programmable atomic operations can be comparatively simple, multi-cycle operations such as floating-point addition, or comparatively complex, multi-instruction operations such as a Bloom filter insert. The programmable atomic operations can be the same as or different than the predetermined atomic operations, insofar as they are defined by the user rather than a system vendor. For these operations, the cache controller 204 can reserve a cache line in the cache 202, by setting a hazard bit (in hardware), so that cache line cannot be read by another process while it is in transition. The data is obtained from either the off-die memory 230 or the cache 202, and is provided to the PAU 228 to perform the requested programmable atomic operation. Following the atomic operation, the PAU 228 will provide the resulting data to the network communication interface 208 to directly encode outgoing data packets having the resulting data for transmission on the chiplet network 122. In addition, the PAU 228 will provide the resulting data to the cache controller 204, which will also write the resulting data to the cache 202. Following the writing of the resulting data to the cache 202, any corresponding hazard bit which was set will be cleared by the cache controller 204.
In selected examples, the approach taken for programmable atomic operations is to provide multiple, generic, custom atomic request types that can be sent through the chiplet network 122 to the memory controller chiplet 218 from an originating source such as a processor or other system component. The cache controller 204 and/or off-die memory controller 206 identify the request as a custom atomic and forward the request to the PAU 228. In a representative embodiment, the PAU 228: (1) is a programmable processing element capable of efficiently performing a user defined atomic operation; (2) can perform load and stores to memory, arithmetic and logical operations and control flow decisions; and (3) leverages the RISC-V ISA with a set of new, specialized instructions to facilitate interacting with the controllers to atomically perform the user-defined operation. In desirable examples, the RISC-V ISA contains a full set of instructions that support high level language operators and data types. The PAU 228 can leverage the RISC-V ISA, but will commonly support a more limited set of instructions and limited register file size to reduce the die size of the unit when included within the memory controller chiplet 218.
As mentioned above, prior to the writing of the read data to the cache 202, the set hazard bit for the reserved cache line is to be cleared, by the memory hazard clear unit 224. Accordingly, when the request and read data is received by the write merge unit 222, a reset or clear signal can be transmitted by the memory hazard clear unit 224 to the cache 202 to reset the set memory hazard bit for the reserved cache line. Also, resetting this hazard bit will also release a pending read or write request involving the designated (or reserved) cache line, providing the pending read or write request to an inbound request multiplexer for selection and processing.
In an example, a chiplet system can be configured to include or use a serial peripheral interface (SPI) bus or interface. An SPI bus generally includes a multiple-wire serial communication interface that enables a synchronous data link between two devices, a primary device and a secondary device. It provides support for a low to medium bandwidth network connection between the devices supporting the SPI. In a chiplet system with an SPI bus, the primary device can be a first chiplet, and the secondary device, or secondary devices, can be one or more other chiplets coupled to the first chiplet using the SPI bus.
The wires of the SPI bus typically include two control channels and two data channels. The control channels of the SPI bus include a chip select (CS) channel, and a serial clock channel (SCLK). In some examples, more than one secondary device can be connected to the primary device, however, only one secondary device can generally be accessed at any given time. Selection of a particular secondary device from among multiple secondary devices can be accomplished using the chip select or CS channel. A CS signal is outputted from the primary device and allows for activation and deactivation of a secondary device by the primary device. In an example, every secondary device can use its own separate CS channel for activation. In other examples discussed herein, a single multiple-drop CS channel can be used for global activation or deactivation of multiple secondary devices, and unique secondary device identifiers can be used to select a particular secondary device to receive a command from the primary device.
The primary device can also provide a serial clock signal. The clock signal can be used to clock the shifting of serial data into and out of the primary and secondary devices. Use of this clock allows SPI architecture to operate in a primary/secondary full duplex mode, wherein data can be transferred simultaneously from one device to another.
The data channels of the SPI bus can include a Serial Data In (SDI) line and a Serial Data Out (SDO) line. The SI line is a data communication line that outputs data from the primary device to the secondary device. The SO line is a second data communication line used to output data from the secondary device to the primary device. Both data channels are active when the chip select channel is activated for the specified secondary device, unless the specified secondary device is configured to operate in an always-on manner.
In an example, initialization of I/O channels of multiple chiplets in the chiplet system 110 can include or use an SPI bus. Using the SPI bus, I/O channels for all chiplets of a multi-chiplet I/O interface can be interoperable within a single system. The SPI-based techniques discussed herein can be used for both hardware-based and software-based initialization mechanisms to guarantee interoperability of the interconnected chiplet I/O channels.
A peripheral device or secondary device, such as the SPI memory device 302, can include a secondary device controller 314 with multiple interface pins, including a chip select (CS) pin to receive a select signal 318, a clock (SCLK) pin to receive a clock signal 320, a controller output peripheral input (COPI or MOSI) pin to receive a secondary device input signal 322 from the primary device 312, and a controller input peripheral output (CIPO or MISO) pin to provide a secondary device output signal 316 to the primary device 312. Data transfer between the primary device 312 and the SPI memory device 302 or secondary device can take place serially and synchronously using the secondary device output signal 316 and the secondary device input signal 322.
In an example, the SPI memory device 302 can include a device identifier 310. The device identifier 310 can include a memory address or register that can be statically or dynamically assigned to a particular value, such as in response to a Static ID input. That is, the device identifier 310 can be statically or dynamically strapped to a value that uniquely identifies the SPI memory device 302, such as when the SPI memory device 302 is used together with other secondary devices in the SPI system 300.
In an example, the primary device 312 can enable the secondary device controller 314 by setting the chip select signal 318 low. In an example, the secondary device controller 314 can be enabled using a statically-tied chip select port at the SPI memory device 302, such as when a CS signal is unused or unavailable. After the secondary device controller 314 is enabled, the primary device 312 can send the clock signal 320 and a corresponding data signal, or secondary device input signal 322. Each bit transmitted in the secondary device input signal 322 or secondary device output signal 316 can be synchronous to either a rising edge or falling edge of the clock signal 320. In an example, the primary device 312 can provide data on the secondary device input signal 322 latched on a rising clock edge and the SPI memory device 302 can provide data on the secondary device output signal 316 on a falling edge. In this example, a first rising edge of the clock signal 320 can correspond to the first bit of the secondary device input signal 322, and subsequent rising clock edges of the clock signal 320 can correspond to subsequent bits of the secondary device input signal 322. Similarly, each bit output on the secondary device output signal 316 can transition on a falling edge of the clock signal 320. Other signal timing schemes can similarly be used.
In an example, communication between the primary device 312 and the secondary device controller 314 can begin when the primary device 312 sets the select signal 318 low. The primary device 312 can subsequently send the clock signal 320 and begin transmitting a message using the secondary device input signal 322. In an example, a message can include a one-byte command followed by a memory address of one or more whole bytes, and further followed by data of one or more whole bytes. The secondary device controller 314 can respond by sending a synchronous message using the secondary device output signal 316. Due to the nature of conventional SPI, the secondary device controller 314 can be configured to output information on the secondary device output signal 316 until a specified or expected time at which the primary device 312 expects or is configured to receive a response.
In an example, the primary device 312 can send a write register command or a read register command in a message to the secondary device controller 314. The write register command or read register command can configure the secondary device controller 314 to access memory registers 324. Data transfer to and from the secondary device controller 314 and registers 324 can use a register bus 334, such as can have a dedicated control wire. The registers 324 can include, for example, a status register to indicate device operation status and/or a special mode enable register such as a block writing lock register or BWL register 326 such as to prevent certain portions of memory from being written. In an example, the registers 324 can include a one time programmable (OTP) enable register, or OTP register 328, such as to enable reading from or writing to an OTP portion of memory, and/or a parameter page (PP) enable register, or PP register 330, such as to enable reading from or writing to a parameter page of memory. In an example, the secondary device controller 314 can be configured to access the registers 324 when it performs other internal operations.
In an example, access to the registers 324 can permit a user or client to control functional aspects of the SPI memory device 302, such as an output buffer drive strength, a desired number of clock cycles of latency for outputting data, an address cycle format to require whole bytes or to use a minimum number of addresses, and/or whether to enable or disable error correcting codes (ECC). Certain registers can hold, for example, error status information, such as can be reset upon the issuance of a register write command, while other registers can enable a user to control timing based on varying clock signal 320 frequencies. In an example, a register can be configured to enable the SPI memory device 302 to switch between different modes and interfaces, such as between SPI NAND and NAND user modes and interfaces.
In an example, to perform an operation at a specified memory address, the secondary device controller 314 can send a memory address signal using a bus to row and column decoders (not depicted). The secondary device controller 314 can control activation of the row and column decoders using a control wire. Depending on the operation, the secondary device controller 314 can, for example, load data bytes into a cache register 304 using a cache bus 332. In an example, a NAND array 308 can receive data, such as one page at a time, through a data register 306, such as in coordination with the cache register 304.
In an example, a first SPI peripheral or secondary device, such as the SPI memory device 302 or other secondary device or chiplet, can be configured to respond to a read request from a controller or other chiplet, such as the primary device 312, within a specified number of clock cycles. The SPI bus, such as can be coupled to one or multiple secondary devices or chiplets, can be impeded from carrying out other data communication until the first SPI peripheral or secondary device sends a complete response to the primary device 312. The present inventors have recognized, among other things, that a solution to this bus occupation issue can include using a secondary device status field in an SPI message that is sent from the first SPI secondary device. For example, the secondary device status field can be included in the secondary device output signal 316 that is communicated to the primary device 312. The secondary device status field can indicate that the sending secondary device is or is not ready to send a response. When the secondary device status field indicates the secondary device is not ready to send a response, then the primary device 312 can be configured to issue a later or deferred request to the secondary device. Systems and methods discussed herein can thus be used to enable long latency SPI operations to proceed in the background while an SPI bus is used for other communications or transactions, such as between the primary device 312 and one or more other secondary devices. In an example, systems and methods discussed herein can be used during initialization routines for chiplet systems, for example, using SPI operations and hardware for initial communications before transitioning to other protocols or busses.
In an example, when the first SPI secondary device responds with a deferred read response status, the primary device 312 can queue a later read request for the deferred information. Between receiving the deferred read response status information from the secondary device and issuing the later deferred read request, the primary device 312 can perform other operations using the same SPI bus. In an example, the primary device 312 can eventually issue a deferred read request, such as using a special request signal. In an example, the primary device 312 can immediately issue the deferred read request upon receipt of the deferred read response from a secondary device, or the primary device 312 can issue the deferred read request after a specified or variable blanking period.
In an example, in response to the primary device 312 issuing a deferred read request using the special request signal, the secondary device can optionally issue another deferral if the payload is not yet available. If the payload is available, then the secondary device can reply with a message that includes a successful secondary device status field and the payload.
The examples of
In an example, a message sent from a primary device to a secondary device can include a 2-bit Command field (e.g., C[1:0] in Table 1). The Command field can comprise a portion of the secondary device input signal 322 and can indicate a command or instruction from the primary device 312. In an example, the Command field can include information about whether the receiving device or secondary device is directed to perform a read operation or a write operation. In an example, the Command field can indicate whether a controller request (e.g., a read request) is an initial request or a deferred request.
In an example, a message can include a 7-bit Chiplet Identification field and parity bit (e.g., ID[6:0] and IDP in Table 1). The Chiplet Identification field can include an identifier that can be used, for example, to address a particular chiplet or particular secondary device in a system, such as in the chiplet system 110. In an example, a Chiplet Identification field can be used when peripheral or secondary device SPI chip select channels are unavailable or are unused in an SPI bus. In an example, a Chiplet Identification field can be used to reduce or eliminate a number of chip select channels or lines that are required or used in an SPI bus.
In an example, a 32-bit Address and parity bit (e.g., A[31:0] and AP in Table 1) can follow the Chiplet Identification message. The Address message can be used, for example, to locate a particular register, such as in the data register 306 or elsewhere in the chiplet system 110. In an example, a 64-bit Data message and parity bit (e.g., D[63:0] and DP in Table 1) can follow the Address message. The Data message can comprise a data payload such as for storage in, or retrieval from, the secondary device.
In an example, a message that includes a secondary device status field and parity bit (e.g., S[2:0] and SP in Table 1) can be provided from the secondary device to the primary device. The secondary device status field can, in an example, be a one, two, three, or more bit field or message. In the examples illustrated herein, the secondary device status field comprises a 3-bit message and a parity bit, however secondary device status fields or messages can be configured to have more or fewer bits depending on an amount of information to be exchanged. Various commands or information can be encoded in a secondary device status field, such as to indicate a successful operation, an unsuccessful or incomplete operation, or a fault.
The various message fields, usages, and message components discussed herein are examples only and should not be considered limiting. For example, other additional message fields can be used, or fewer message fields can be used in SPI deferred-response communications. In an example, the various fields can be arranged in different orders or sequences to similarly provide deferred-response communications. The various sizes or bit lengths of the fields or message components are provided as examples only.
Using the conventions provided in Table 1,
The example of
In the example of
In response to the Command, Chiplet Identification, and Address fields, the secondary device can prepare and communicate a response to the primary device 312, such as using the secondary device output signal 316. In the example of
The example of
In the example of
Following the blanking period and in response to the write instruction, the secondary device can return a message to the primary device 312 using the secondary device output signal 316. In the example of
The first message can include various fields, such as a command field with one or more bits configured to enable access to the second chiplet, and a chiplet identification field with one or more bits configured to indicate or specify the second chiplet from among one or more other chiplets in the system. In an example, the first message comprises a request for a particular data payload from the second chiplet. In an example, the first message comprises an initialization instruction for the second chiplet.
At block 604, the first method 600 can include receiving the first message at the second chiplet. The second chiplet can be configured with a static identifier that uniquely identifies the second chiplet from other secondary chiplets or peripherals that are coupled to the SPI interface. At decision block 606, the second chiplet can be configured to analyze a chiplet identification field of the first message to determine whether it includes or corresponds to the unique static identifier of the second chiplet.
If the information in the chiplet identification field of the first message corresponds to the identifier of the second chiplet, then the first method 600 continues at block 608. At block 608, the first method 600 includes using the second chiplet to receive a payload or instruction portion of the first message. The second chiplet can optionally perform an operation (e.g., a write operation or a read operation) associated with the first message.
If the information in the chiplet identification field of the first message does not correspond to the identifier of the second chiplet, then the first method 600 can continue at block 610. At block 610, the first method 600 includes, at the second chiplet, ignoring or disregarding the first message. In other words, if the second chiplet does not detect or receive a message that includes the unique identifier associated with the second chiplet, then the second chiplet is not required or instructed to act on a payload or instruction portion of the message.
The first method 600 can optionally include receiving the first message at one or more other chiplets, and each of the one or more other chiplets can have its own unique identifier. Each of the one or more other chiplets can be configured to analyze the first message to determine whether the first message includes an identifier corresponding to one of the chiplets. If a particular chiplet has an identifier that corresponds to the identifier field of the first message, then the particular chiplet can be configured to receive or respond to the payload portion of the first message.
The second method 700 begins at block 702 with providing each of multiple different secondary devices with respective static device identifiers. For example, instances of the device identifier 310 in each of the different secondary devices can be populated with a different fixed value.
At block 704, the second method 700 can include using a primary device on an SPI interface to provide a chip select signal to globally enable or disable communication with secondary devices on the same SPI interface. At block 706, the primary device can populate outgoing messages with identification fields that correspond to a particular destination device or particular one of the secondary devices.
For example, a first secondary device can monitor the CS signal from the primary device (e.g., the global CS signal from block 704, such as provided to multiple different secondary devices on the same SPI interface) and, when the CS signal enables the secondary devices, the first secondary device can be configured to compare its unique identifier with identification information included in a message received from the primary device. In an example, the identification information can comprise a portion of a data payload received from the primary device. If the identification information matches the identifier of the first secondary device, then the first secondary device can be further enabled or configured to attend to the payload. If the identification information does not match the identifier of the first secondary device, then the first secondary device can disregard the payload or can remain in a partially disabled state. Other secondary devices on the same SPI interface can similarly receive and analyze the message to determine whether to remain in a listening state or to take some action in response to the message.
The example of
At block 804, the third method 800 can include statically tying a CS port of each secondary device to an enabled state. That is, a CS input port of each of the secondary devices can be tied to an always-on or always active state such that each secondary device is configured to monitor essentially all of the messages or data signals on the SPI bus from the primary device.
At block 806, the third method 800 can include using the primary device to prepare messages with an identification field that includes an identifier of a particular one of the multiple secondary devices on the SPI bus. Each secondary device can receive and independently analyze messages from the primary device, received using the SPI bus, to determine whether to attend to or disregard the payload from the data signal, such as based on whether the payload includes a particular device's unique static identifier.
In an example, a system that includes a primary device and one or more secondary devices coupled by an SPI interface can be configured to selectively use a static identifier-based communication protocol or a conventional chip select-based SPI protocol for communication. For example,
The example of the fourth method 900 can include or use multiple secondary devices coupled to an SPI bus or interface, and each of the secondary devices includes a CS port. In this example, the SPI bus includes a number of CS channels that corresponds to a number of secondary devices in the system. The example of the fourth method 900 illustrates generally how secondary devices configured with unique identifiers can be compatible with conventional SPI bus systems that use a CS signal to select a particular secondary device.
In the example of
At block 904, the fourth method 900 can include using a primary device to send messages using the SPI bus. The primary device can populate outgoing messages with an identification field that includes the fixed value. At block 906, the fourth method 900 can include using the primary device to provide a CS signal to select a particular secondary device from among the multiple available secondary devices on the SPI bus. That is, by fixing a device identifier 310 on each of the secondary devices to have a common value, the system can be configured for conventional SPI-based communication with multiple secondary devices where each secondary device is separately enabled or disabled by a dedicated or respective CS signal.
In alternative embodiments, the machine 1000 can operate as a standalone device or can be connected (e.g., networked) to other machines. In a networked deployment, the machine 1000 can operate in the capacity of a server machine, a client machine, or both in server-client network environments. In an example, the machine 1000 can act as a peer machine in peer-to-peer (P2P) (or other distributed) network environment. The machine 1000 can be a personal computer (PC), a tablet PC, a set-top box (STB), a personal digital assistant (PDA), a mobile telephone, a web appliance, a network router, switch or bridge, or any machine capable of executing instructions (sequential or otherwise) that specify actions to be taken by that machine. Further, while only a single machine is illustrated, the term “machine” shall also be taken to include any collection of machines that individually or jointly execute a set (or multiple sets) of instructions to perform any one or more of the methodologies discussed herein, such as cloud computing, software as a service (SaaS), other computer cluster configurations.
The machine 1000 (e.g., computer system) can include a hardware processor 1002 (e.g., a central processing unit (CPU), a graphics processing unit (GPU), a hardware processor core, or any combination thereof), a main memory 1008, a static memory 1010 (e.g., memory or storage for firmware, microcode, a basic-input-output (BIOS), unified extensible firmware interface (UEFI), etc.), and mass storage 1012 (e.g., hard drives, tape drives, flash storage, or other block devices) some or all of which can communicate with each other via an interlink 1018 (e.g., a bus, such as an SPI bus). The machine 1000 can further include a display device 1020, an alphanumeric input device 1022 (e.g., a keyboard), and a user interface (UI) navigation device 1024 (e.g., a mouse). In an example, the display device 1020, input device 1022, and navigation device 1024 can be a touch screen display. The machine 1000 can additionally include a mass storage 1012 (e.g., drive unit), a signal generation device 1028 (e.g., a speaker), a network interface device 1014, and one or more sensor(s) 1026, such as a global positioning system (GPS) sensor, compass, accelerometer, or other sensor. The machine 1000 can include an output controller 1030, such as a serial (e.g., universal serial bus (USB), parallel, or other wired or wireless (e.g., infrared (IR), near field communication (NFC), etc.) connection to communicate or control one or more peripheral devices (e.g., a printer, card reader, etc.).
Registers of the processor 1002, the main memory 1008, the static memory 1010, or the mass storage 1012 can be, or include, a machine-readable medium 1006 on which is stored one or more sets of data structures or instructions 1004 (e.g., software) embodying or utilized by any one or more of the techniques or functions described herein. The instructions 1004 can also reside, completely or at least partially, within any of registers of the processor 1002, the main memory 1008, the static memory 1010, or the mass storage 1012 during execution thereof by the machine 1000. In an example, one or any combination of the hardware processor 1002, the main memory 1008, the static memory 1010, or the mass storage 1012 can constitute the machine-readable medium 1006 or media. While the machine-readable medium 1006 is illustrated as a single medium, the term “machine readable medium” can include a single medium or multiple media (e.g., a centralized or distributed database, or associated caches and servers) configured to store the one or more instructions 1004. In an example, the various memory units or processor 1002 can be communicatively coupled using a bus such as an SPI bus.
The term “machine readable medium” can include any medium that is capable of storing, encoding, or carrying instructions for execution by the machine 1000 and that cause the machine 1000 to perform any one or more of the techniques of the present disclosure, or that is capable of storing, encoding or carrying data structures used by or associated with such instructions. Non-limiting machine readable medium examples can include solid-state memories, optical media, magnetic media, and signals (e.g., radio frequency signals, other photon based signals, sound signals, etc.). In an example, a non-transitory machine readable medium comprises a machine readable medium with a plurality of particles having invariant (e.g., rest) mass, and thus are compositions of matter. Accordingly, non-transitory machine-readable media are machine readable media that do not include transitory propagating signals. Specific examples of non-transitory machine readable media can include: non-volatile memory, such as semiconductor memory devices (e.g., electrically programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM)) and flash memory devices; magnetic disks, such as internal hard disks and removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks.
In an example, information stored or otherwise provided on the machine-readable medium 1006 can be representative of the instructions 1004, such as instructions 1004 themselves or a format from which the instructions 1004 can be derived. This format from which the instructions 1004 can be derived can include source code, encoded instructions (e.g., in compressed or encrypted form), packaged instructions (e.g., split into multiple packages), or the like. The information representative of the instructions 1004 in the machine-readable medium 1006 can be processed by processing circuitry into the instructions to implement any of the operations discussed herein. For example, deriving the instructions 1004 from the information (e.g., processing by the processing circuitry) can include: compiling (e.g., from source code, object code, etc.), interpreting, loading, organizing (e.g., dynamically or statically linking), encoding, decoding, encrypting, unencrypting, packaging, unpackaging, or otherwise manipulating the information into the instructions 1004.
In an example, the derivation of the instructions 1004 can include assembly, compilation, or interpretation of the information (e.g., by the processing circuitry) to create the instructions 1004 from some intermediate or preprocessed format provided by the machine-readable medium 1006. The information, when provided in multiple parts, can be combined, unpacked, and modified to create the instructions 1004. For example, the information can be in multiple compressed source code packages (or object code, or binary executable code, etc.) on one or several remote servers. The source code packages can be encrypted when in transit over a network and decrypted, uncompressed, assembled (e.g., linked) if necessary, and compiled or interpreted (e.g., into a library, stand-alone executable etc.) at a local machine, and executed by the local machine.
The instructions 1004 can be further transmitted or received over a communication network 1016 using a transmission medium via the network interface device 1014 utilizing any one of a number of transfer protocols (e.g., frame relay, internet protocol (IP), transmission control protocol (TCP), user datagram protocol (UDP), hypertext transfer protocol (HTTP), etc.). Example communication networks can include a local area network (LAN), a wide area network (WAN), a packet data network (e.g., the Internet), mobile telephone networks (e.g., cellular networks), plain old telephone (POTS) networks, and wireless data networks (e.g., Institute of Electrical and Electronics Engineers (IEEE) 802.11 family of standards known as Wi-Fi®, IEEE 802.16 family of standards known as WiMax®), IEEE 802.15.4 family of standards, peer-to-peer (P2P) networks, among others. In an example, the network interface device 1014 can include one or more physical jacks (e.g., Ethernet, coaxial, or phone jacks) or one or more antennas to connect to the communication network 1016. In an example, the network interface device 1014 can include a plurality of antennas to wirelessly communicate using at least one of single-input multiple-output (SIMO), multiple-input multiple-output (MIMO), or multiple-input single-output (MISO) techniques. The term “transmission medium” shall be taken to include any intangible medium that is capable of storing, encoding or carrying instructions for execution by the machine 1000, and includes digital or analog communications signals or other intangible medium to facilitate communication of such software. A transmission medium is a machine readable medium.
The following are examples or devices and methods in accordance with the teachings herein. In an example, one or more of the following examples can be used with an SPI system to reduce or eliminate a number of chip select channels while also providing backwards compatibility with conventional SPI systems that include or use dedicated chip select channels for each secondary device.
Example 1 can include an apparatus that includes a primary chiplet coupled to multiple secondary chiplets using a serial peripheral interface (SPI) interface configured for an SPI protocol. In Example 1, the primary chiplet can be configured as an SPI primary device, and the primary chiplet can be configured to send a first message using a first communication channel in the SPI interface, wherein the first message comprises a first command field indicating an instruction for a first selected chiplet of the multiple secondary chiplets, the first selected chiplet configured as an SPI secondary device, and a first chiplet identification field indicating the first selected chiplet of the multiple secondary chiplets.
Example 2 can include or use the features of Example 1, wherein the primary chiplet is configured to send a second message using the first communication channel in the SPI interface, wherein the second message comprises a second command field indicating an instruction for a second selected chiplet of the multiple secondary chiplets, the second selected chiplet configured as an SPI secondary device, and a second chiplet identification field indicating the second selected chiplet, wherein the first and second chiplet identification fields comprise different information.
Example 3 can include or use the features of Example 2, wherein the primary chiplet is configured to send the first message and the second message without updating a status of a chip select signal on a chip select communication channel in the SPI interface. That is, in some examples, the primary chiplet can be configured to maintain a chip select signal in a fixed or static state, or a chip select communication channel in the interface can be unused.
Example 4 can include or use features of any of the preceding examples and can further comprise the first selected chiplet, wherein the first selected chiplet is decoupled from, or is configured to disregard information on, a chip select communication channel in the SPI interface.
Example 5 can include or use features of any of the preceding examples, wherein each of the multiple secondary chiplets is configured as a respective SPI secondary device, and wherein each of the multiple secondary chiplets is coupled to the primary chiplet using the same chip select communication channel in the SPI interface.
Example 6 can include or use features of any of the preceding examples, wherein in response to receiving the first message and before preparing a responsive message to the first message, the first selected chiplet is configured to determine whether the first chiplet identification field of the first message corresponds to an identification of the first selected chiplet.
Example 7 can include or use the features of Example 6, wherein in response to receiving the first message from the primary chiplet, each of the multiple secondary chiplets is configured to determine whether to prepare a response to the first message based on information from the first chiplet identification field.
Example 8 can include or use features of any of the preceding examples, wherein the first message further comprises a register address field of the first message, the register address field including one or more bits set to indicate an address of a memory register on the first selected chiplet.
Example 9 can include or use the features of Example 8, wherein the command field of the first message indicates a write instruction for the first selected chiplet, and the first message further comprises a data field, and data in the data field is configured to be stored in a location corresponding to the register address field of the first message.
Example 10 can include or use the features of Example 9, wherein in the first message, the data field immediately follows the first chiplet identification field, and the first chiplet identification field immediately follows the command field.
Example 11 can include or use features of Example 8, wherein the command field of the first message indicates a read instruction for the first selected chiplet.
Example 12 can include a system comprising a primary chiplet coupled to multiple secondary chiplets, and the primary chiplet can be configured to provide a clock signal, and the system can comprise a first selected chiplet of the multiple secondary chiplets, the first selected chiplet configured to communicate with the primary chiplet using an interface configured to use a serial peripheral interface (SPI) protocol. In Example 12, the primary chiplet can be configured to perform operations comprising sending the clock signal to the first selected chiplet, and sending a first message to the first selected chiplet using the interface, wherein the first message comprises a command field including one or more bits to enable the primary chiplet to access memory registers on the first selected chiplet, and a register address field including one or more bits to address a particular memory register on the first selected chiplet, and a first chiplet identification field indicating the first selected chiplet of the multiple secondary chiplets.
Example 13 can include or use the features of Example 12, wherein the primary chiplet is configured to send a chip select signal to the multiple secondary chiplets, and wherein each of the multiple secondary chiplets is configured to determine whether to respond to the first message based on information in the first chiplet identification field.
Example 14 can include or use features of any of Examples 12 or 13, wherein the primary chiplet is configured to concurrently send a first chip select signal to the first selected chiplet, and a different second chip select signal to each of the other secondary chiplets.
Example 15 can include or use features of any of Examples 12-14, wherein the first selected chiplet comprises a chip select interface provided in an always-active state.
Example 16 can include or use features of any of Examples 12-15, and can further comprise a second selected chiplet of the multiple secondary chiplets, the second selected chiplet configured to communicate with the primary chiplet using the interface, wherein the primary chiplet is further configured to perform operations that include sending a second message to the second selected chiplet using the interface, wherein the second message comprises a second chiplet identification field and the second chiplet identification field corresponds exclusively to the second selected chiplet.
Example 17 can include or use features of any of Examples 12-16, wherein the first selected chiplet is decoupled from, or is configured to disregard information on, a chip select communication channel in the interface coupling the primary chiplet and the multiple secondary chiplets.
Example 18 can include a method for data communication between a primary device and one or more secondary devices. Example 18 can include, at a primary chiplet configured to be communicatively coupled to a secondary chiplet using a serial peripheral interface (SPI) bus, sending a first message to the secondary chiplet, wherein the first message comprises a command field including one or more bits configured to enable access to the secondary chiplet and a first chiplet identification field configured to distinguish the secondary chiplet from other secondary chiplets that are coupled to the same bus.
Example 19 can include or use the features of Example 18, and can further include sending a second message from the primary chiplet to a different chiplet using the bus, wherein the second message comprises a second chiplet identification field configured to distinguish the different chiplet from each of the other secondary chiplets.
Example 20 can include or use features of any of Examples 18 or 19, and can further include, at the secondary chiplet, conditioning a response on information in the first chiplet identification field and on information on a chip select portion of the SPI bus.
Each of the above Examples can be combined or used together in various ways to carry out communications over a synchronous interface, such as over an SPI interface or analogous interface.
The above detailed description includes references to the accompanying drawings, which form a part of the detailed description. The drawings show, by way of illustration, specific embodiments in which the invention can be practiced. These embodiments are also referred to herein as “examples”. Such examples can include elements in addition to those shown or described. However, the present inventors also contemplate examples in which only those elements shown or described are provided. Moreover, the present inventors also contemplate examples using any combination or permutation of those elements shown or described (or one or more aspects thereof), either with respect to a particular example (or one or more aspects thereof), or with respect to other examples (or one or more aspects thereof) shown or described herein.
In this document, the terms “a” or “an” are used, as is common in patent documents, to include one or more than one, independent of any other instances or usages of “at least one” or “one or more.” In this document, the term “or” is used to refer to a nonexclusive or, such that “A or B” can include “A but not B.” “B but not A,” and “A and B,” unless otherwise indicated. In the appended claims, the terms “including” and “in which” are used as the plain-English equivalents of the respective terms “comprising” and “wherein”. Also, in the following claims, the terms “including” and “comprising” are open-ended, that is, a system, device, article, or process that includes elements in addition to those listed after such a term in a claim are still deemed to fall within the scope of that claim. Moreover, in the following claims, the terms “first,” “second,” and “third,” etc. are used merely as labels, and are not intended to impose numerical requirements on their objects.
The above description is intended to be illustrative, and not restrictive. For example, the above-described examples (or one or more aspects thereof) can be used in combination with each other. Other embodiments can be used, such as by one of ordinary skill in the art upon reviewing the above description. It is submitted with the understanding that it will not be used to interpret or limit the scope or meaning of the claims. Also, in the above Detailed Description, various features can be grouped together to streamline the disclosure. This should not be interpreted as intending that an unclaimed disclosed feature is essential to any claim. Rather, inventive subject matter can lie in less than all features of a particular disclosed embodiment. Thus, the following claims are hereby incorporated into the Detailed Description, with each claim standing on its own as a separate embodiment, and it is contemplated that such embodiments can be combined with each other in various combinations or permutations. The scope of the invention should be determined with reference to the appended claims, along with the full scope of equivalents to which such claims are entitled.
