Patent Application
20190066626
This disclosure relates generally to the field of circuit design, and more particularly to automatically generating circuit elements driven by clocks with different frequencies.
Hardware circuits such as FPGAs or ASICs may be configured to perform various kinds of calculations. When compiling hardware description code to generate output data for programming or configuring a circuit, there may be clock and throughput requirements on interface circuitry through which data is transferred into and out of different hardware blocks. For example, FPGAs may include dedicated interface circuitry between programmable functional blocks. It may be desirable, however, to run circuit blocks at a higher frequency than an interface block.
Hardware devices, including programmable hardware elements, may be implemented using interface elements that receive input data and operate at one clock frequency and functional elements that operate at another clock frequency. Clock and throughput requirements on the interface circuitry of hardware components may set clocking limits that are lower than what functional components may tolerate. For example, data input may be limited to a clock speed of 40 MHz, while functional blocks may process data at 80 MHz, 120 MHz, or other frequencies.
It may be advantageous to use more or fewer of some functional blocks or interface circuits in order to reduce cost, increase performance or reliability, reduce the requirements for additional parts, or other beneficial factors. Accordingly, it may be advantageous to generate hardware utilizing more than a single clock frequency. For example, interface circuitry may be run at one frequency while functional components may be run at a higher frequency. In some embodiments, functional components may have more clock cycles in which to process incoming data than the number of clock cycles available at the interface.
Generating instructions which indicate different clock frequencies for separate components may be difficult or time consuming; generating these instructions automatically may provide significant benefits in time savings, increased productivity, increased performance of hardware, or other benefits.
While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and are herein described in detail. It should be understood, however, that the drawings and detailed description thereto are not intended to limit the disclosure to the particular form disclosed, but on the contrary, the intention is to cover all modifications, equivalents and alternatives falling within the spirit and scope of the present disclosure, including the appended claims. Particular features, structures, or characteristics may be combined in any suitable manner consistent with this disclosure.
It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting. As used herein, the singular forms “a”, “an”, and “the” include singular and plural referents unless the content clearly dictates otherwise. Furthermore, the words “can” and “may” are used throughout this application in a permissive sense (i.e., having the potential to, being able to), not in a mandatory sense (i.e., must). The term “include,” and derivations thereof, mean “including, but not limited to.” The term “coupled” means directly or indirectly connected.
Within this disclosure, different entities (which may variously be referred to as “units,” “circuits,” other components, etc.) may be described or claimed as “configured” to perform one or more tasks or operations. This formulation—[entity] configured to [perform one or more tasks]—is used herein to refer to structure (i.e., something physical, such as an electronic circuit). More specifically, this formulation is used to indicate that this structure is arranged to perform the one or more tasks during operation. A structure can be said to be “configured to” perform some task even if the structure is not currently being operated. A “mobile device configured to generate a hash value” is intended to cover, for example, a mobile device that performs this function during operation, even if the device in question is not currently being used (e.g., when its battery is not connected to it). Thus, an entity described or recited as “configured to” perform some task refers to something physical, such as a device, circuit, memory storing program instructions executable to implement the task, etc. This phrase is not used herein to refer to something intangible.
The term “configured to” is not intended to mean “configurable to.” An unprogrammed mobile computing device, for example, would not be considered to be “configured to” perform some specific function, although it may be “configurable to” perform that function. After appropriate programming, the mobile computing device may then be configured to perform that function.
Reciting in the appended claims that a structure is “configured to” perform one or more tasks is expressly intended not to invoke 35 U.S.C. § 112(f) for that claim element. Accordingly, none of the claims in this application as filed are intended to be interpreted as having means-plus-function elements. Should Applicant wish to invoke Section 112(f) during prosecution, it will recite claim elements using the “means for” [performing a function] construct.
As used herein, the term “based on” is used to describe one or more factors that affect a determination. This term does not foreclose the possibility that additional factors may affect the determination. That is, a determination may be solely based on specified factors or based on the specified factors as well as other, unspecified factors. Consider the phrase “determine A based on B.” This phrase specifies that B is a factor is used to determine A or that affects the determination of A. This phrase does not foreclose that the determination of A may also be based on some other factor, such as C. This phrase is also intended to cover an embodiment in which A is determined based solely on B. As used herein, the phrase “based on” is synonymous with the phrase “based at least in part on.”
In various disclosed embodiments, circuitry may be created with different clock speeds used for different components of the system. Clock and throughput parameters (e.g., based on capabilities of interface circuitry) may set limits that are lower than other components are capable of operating. For example, data input at an interface may be limited to a clock speed of 40 MHz, while functional blocks configured to receive data from the interface may be configured to process data at 80 MHz, 120 MHz, or other frequencies.
In some embodiments, hardware (e.g., a programmable hardware device being programmed) may have constraints that include, without limitation: a maximum or minimum amount of area for circuitry, clock speed of functional components, or number of functional blocks. It may be advantageous to use more or fewer of some functional blocks or interface circuits in order to reduce cost, increase performance or reliability, reduce the requirements for additional parts, etc.
Accordingly, it may be advantageous to automatically generate hardware with different portions that utilize multiple different clock frequencies. For example, interface circuitry may be run at one frequency while functional components may be run at a higher frequency. In some embodiments, this may provide functional components with more clock cycles in which to process incoming data than the number of clock cycles available at the interface.
When compiling a program for implementing on hardware (e.g. on an FPGA or ASIC) certain program statements may specify that a specific number of copies of the same function that may be run in parallel, in some embodiments. A parallel “for” loop is one example of such a statement, where different iterations of the loop may be performed independently. These program statements may be referred to as an indication of “execution parallelism” for the function. The compiler may choose to use a number of different functional hardware blocks to implement copies of the function in parallel.
In some embodiments, execution parallelism is related to performing calculations or processing data at the same time. As one non-limiting example, for a function f(y)=2y, and a list of numbers on which the function is to be performed, performing the multiplication in parallel may include multiplying each number of the list by 2 at the same time, e.g. using a separate processor or calculating circuit for each number. Execution parallelism may include executing program instructions at substantially the same time, at overlapping times, beginning at substantially the same time, or other timings with at least partial overlap in execution. Execution parallelism may include executing the same instructions in parallel and/or executing different instructions in parallel.
Non-limiting examples of code that may indicate execution parallelism include, but are not limited to: function, subprogram, procedure, basic block, for/generate statement in VHDL, Verilog, and SystemVerilog; “Parallel FOR loop” structure in LabVIEW; ParFor in C, Matlab, Python, Fortran; “thread” in python; “replicate” structures in LabVIEW; Async/Await functions in C#; task parallel library tasks in C#; “for/generate” in VHDL; thread functions in Java, and so on.
In some embodiments, a compiling program may detect a parallel statement in the program for implementing on hardware; the parallel statement may specify a parallelism parameter, e.g., indicating that up to K instances of a function can be correctly performed in parallel. In some embodiments, the compiler may have default preferences, or preferences specified by a user, including the frequency of a base clock or other clocks. The compiler may automatically generate a functional block with a clock speed which is a multiple of the base clock, and then create a number, e.g. N, of these functional blocks and implement the K parallel functions onto these N functional blocks. In some embodiments, the compiler may then create all interface and data routing circuitry required to operate the functional blocks. In some embodiments, when N is smaller than K, ones of the functional blocks are configured to perform the function multiple times serially in order to perform all K instances of the function.
In some embodiments, the compiler may choose a number of functional blocks for parallel operation based on constraints relating to data input clock speed, functional block clock speed, number of functional blocks, total number of circuit components, or other constraints. Constraints may be entered by a user, stored in a library for hardware being configured, etc. In some embodiments, the compiler may generate multiple configurations that use different numbers of parallel functional blocks (and/or different clock speeds) and select one or more of the configurations according to one or more criteria. Criteria may include power, performance, and/or resource utilization of each configuration, for example. In some embodiments, resources utilized may include the amount of available programmable elements in a programmable hardware device, the amount of memory available to an element, the number of processing elements used, the number of data busses required, etc.
In some embodiments, the compiler may output instructions for instantiating hardware in a programmable hardware device (or ASIC). In some embodiments, a device executing the compiler may also instantiate the hardware in a programmable hardware device (or ASIC) or may interface with other hardware or software for instantiating the hardware.
Exemplary Compiler System configured to Generate Multi-Clock Circuitry
Turning now to
In some embodiments, program code 110 specifies hardware to be used for instantiation. In other embodiments, the hardware to be used may be determined by compiler 120 (e.g., based on functionality specified in program code 110), or may be specified as an input to compiler 120. Program code 110 may be stored on a non-transitory computer readable storage medium.
Compiler 210 in the illustrated embodiment receives program code 110 as an input, and is configured to output hardware description 170. In the illustrated embodiment, compiler 210 includes a set of modules including a parallelism detector 130, interface circuitry generator 140, functional circuitry determiner 150, and functional circuitry generator 160. These modules may operate sequentially or in parallel. Operations may be performed on program code 110, on the outputs from other modules, or combinations thereof.
As used herein, the term “module” refers to circuitry configured to perform specified operations or to physical non-transitory computer readable media that stores information (e.g., program instructions) that instructs other circuitry (e.g., a processor) to perform specified operations. Such circuitry may implemented in multiple ways, including as a hardwired circuit or as a memory having program instructions stored therein that are executable by one or more processors to perform the operations. The hardware circuit may include, for example, custom very-large-scale integration (VLSI) circuits or gate arrays, off-the-shelf semiconductors such as logic chips, transistors, or other discrete components. A module may also be implemented in programmable hardware devices such as field programmable gate arrays, programmable array logic, programmable logic devices, or the like. A module may also be any suitable form of non-transitory computer readable media storing program instructions executable to perform specified operations.
Parallelism detector 130, in some embodiments, is configured to process program code 110 to identify code indicative of execution parallelism for a function. In some embodiments, the indication includes at least a first number of instances of the function that operate on different sets of input data. Examples of code that may indicate parallelism include, but are not limited to: function, subprogram, procedure, basic block, for/generate statement in VHDL, Verilog, and SystemVerilog; “Parallel FOR loop” structure in LabVIEW; ParFor in C, Matlab, Python, Fortran; “thread” in python; “replicate” structures in LabVIEW; Async/Await functions in C#; task parallel library tasks in C#; “for/generate” in VHDL; thread functions in Java, and so on.
In some embodiments, parallelism detector 130 may process only a portion of program code 110 or parallelism detector 130 may process program code 110 multiple times. In some embodiments, parallelism detector 130 may output information related to processing program code 110 before, during, or after processing. Output from parallelism detector 130 may be displayed, saved to a non-transitory computer-readable medium, etc.
Interface circuitry generator 140, in some embodiments, is configured to process program code 110 and/or outputs from parallelism detector 130. In some embodiments, the processing performed by interface circuitry generator 140 is based at least in part on parameters inputs from a user or otherwise specified. For example, interface circuitry generator 140 may generate interface circuitry based on the amount, type, speed, size, frequency or other characteristics of input data that to be interfaced. In some embodiments, characteristics of the data to be used as input data to interface circuitry may be specified in program code 110, may be user input, or may be received from compiler 120.
In some embodiments, interface circuitry may include wiring, connections, functional elements, modules, or other devices or configurations for receiving data from a data source and sending data to a functional block for processing. Interface circuitry may include circuitry for receiving data at one clock speed and sending data at another clock speed, circuitry for distributing data to various functional blocks, circuitry for aggregating data from various functional blocks, or other circuitry for interfacing between data sources and functional elements.
A multi-clock buffer is one example of interfacing circuitry that is configured to receive data at one clock speed and distribute data at another clock speed. In some embodiments, a multi-clock buffer may include a first-in first-out (FIFO) storage in which the data received first is the first data to be sent out. In some embodiments, input data may be received with a number of channels, and the data may be sent out with fewer channels. For example, if a multi-clock FIFO uses two clock signals, one with twice the frequency of the other, the portion of the multi-clock FIFO using a faster clock may send data using half as many channels.
A commutator is another example of interfacing circuitry that is configured to receive input data and switch between distributing the input data to different ones of the functional blocks. In some embodiments, a commutator may distribute data in a “round robin” fashion, where each of the associated functional blocks receives data in turn, one after the other. In other embodiments, a commutator may distribute data according to a different pattern, including distributing data to multiple functional blocks at once, distributing data to some functional blocks more or less often than others, or other manners of distributing data.
A de-commutator may also be used for interfacing functional blocks with data. In some embodiments, a de-commutator performs a reverse function of a commutator, e.g. a de-commutator may receive data from multiple functional blocks and transmit data to a single output. For example, a de-commutator may receive a processed result from each of a number of functional blocks and may send the data to a multi-clock FIFO at different times or after combining the data to a single output. In some embodiments, a single output may include multiple channels of data.
Functional circuitry determiner 150, in the illustrated embodiment, is configured to process program code 110 and/or outputs of other modules. Functional circuitry determiner 150 may perform calculations to determine the number, type, or configuration of functional blocks needed to implement the functionality specified by program code 110. In some embodiments, determining the number of functional blocks may be based on the rate of data input, the performance characteristics of the functional blocks, the clock frequencies of interface circuits or functional blocks, the number of instances of a function to be run, program code 110, other information relevant to processing data, or combinations or subsets thereof. In some embodiments, functional circuitry determiner 150 determines, for a function with specified parallelism, to use the number of functional blocks specified by parallelism detector 130. For example, the parallelism detector may determine that K parallel instances of a function may be implemented and functional circuitry determiner 150 may determine to instantiate K parallel functional blocks. In some embodiments, functional circuitry determiner 150, may determine to include fewer functional blocks than K, e.g., based on resource limitations in an FPGA being programmed. In some other embodiments, functional circuitry determiner 150 may determine to include more functional blocks than K; for example, the determination may be based on lower voltage and/or power considerations.
Parameters for determining the functional circuitry may be provided as user input to compiler 120 or indicated in program code 110, in some embodiments. In some embodiments, determining the number of functional blocks may be based at least in part on a set of constraints, wherein the constraints include at least one of area, speed, or number of functional blocks. For example, compiler 120 may attempt to increase the number of functional blocks used in parallel to increase speed or decrease the number of functional blocks used in parallel to reduce area. Further, the circuitry being implemented may have a fixed limit on total number of functional blocks, functional blocks of a certain type, etc.
In some embodiments, functional circuitry determiner 150 is configured to determine the functional circuitry by generating multiple configurations that use different numbers of functional blocks for a given parallel function. Functional circuitry determiner 150 may then simulate, analyze, or otherwise evaluate the performance of the different configurations and may select one of the configurations. In some embodiments, selecting a generated circuit may be based on parameters input by a user, indicated by program code 110, set as defaults, or other manner of evaluating performance.
Functional circuitry generator 160, in some embodiments, is configured to generate instructions for instantiating hardware including the number of functional blocks determined by functional circuitry determiner 150. In some embodiments, functional circuitry generator 160 may generate data that specifies instantiation of functional circuitry configured to receive input data formatted in the manner output by interface circuitry generated by interface circuitry generator 140.
In some embodiments, functional circuitry may include circuits or modules configured to perform mathematical operations, for example addition, substraction, Fourier transformations, or other simple or complex calculations.
In the illustrated embodiment, compiler 120 is configured to output hardware description 170. In some embodiments, hardware description 170 includes data that specifies instantiating a hardware device (as a non-limiting example, a programmable hardware element). In some embodiments, hardware description 170 may be a text based description, a machine code description, a binary description, or other descriptions. Hardware description 170 may be saved as a file on a non-transitory computer-readable medium, may exist in a transitory computer-readable medium, may be transmitted to another device or computer, may be transmitted across a network, etc.
In some embodiments, the computing device that executes compiler 120 may also configure instantiated hardware 180. Instantiated hardware 180 may be instantiated based on hardware description 170. In some embodiments, instantiated hardware 180 may be instantiated without hardware description 170 being generated.
Instantiated hardware 180 may comprise various types of hardware. Non-limiting examples of hardware include programmable hardware elements such as field programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), graphics processing units (GPUs), multicore processors, central processing units (CPUs), co-processors, application-specific instruction-set processors (ASIP), soft IP blocks, hard macro blocks, IP blocks, module generator blocks, nodes on a heterogeneous network of compute blocks, a computer, a network of computers, embedded cores, or other computational resources.
In the illustrated embodiment, data input 202 operates at a slow clock frequency, relative to a clock frequency at which core 206 operates. Thus, core 206 is referred to as an “overclock” core. Data input 202 sends data (e.g., received from other functional blocks or an I/O port of hardware 200) to multi-clock FIFO 204 at the slow clock frequency, and multi-clock FIFO 204 sends data at the frequency of fast clock 205 to overclock core 206. In the illustrated embodiment, overclock core 206 processes data and sends the resulting data to multi-clock FIFO 207. Multi-clock FIFO 207 receives data at the frequency of fast clock 205 and sends data to other functional blocks or to an output port at the slow clock frequency of data output 208.
In the illustrated embodiment, a single functional block, overclock core 206 is implemented. In some embodiments, overclock core 206 may perform one or more functions, may perform various calculations, may perform a function at least twice for different input data, or may perform other computations. Overclock core 206 may have been generated by functional circuitry generator 160. In some embodiments, hardware 200 may be described by hardware description 170 and instantiated by compiler 120 or other manners of instantiating hardware, as described previously.
In some embodiments, data input and output may comprises multiple channels of data corresponding to multiple sources and/or targets.
Overclock cores 330, in the illustrated embodiment, are configured to operate at the frequency of fast clock 306. In some embodiments, operating at the faster clock rate may allow overclock cores 330 to perform calculations as fast as or faster than input data may be received. Four overclock cores 330 are illustrated in
In the illustrated embodiment, overclock cores 330 send output data to de-commutator 325. In some embodiments, overclock cores 330 may send output data consecutively in the same manner as input data was received. In some embodiments, overclock cores may send data according to the order in which processing was complete, according to a different order than data was received, or in other manners.
In the illustrated embodiment, de-commutator 325 receives data from overclock cores 330 according to fast clock 306 and send data to multi-clock FIFO 340. De-commutator 325 may send data in the same order as received, in a different order, or may perform other operations on the received data, in some embodiments. Multi-clock FIFO 340 receives data at the frequency of fast clock 360, in the illustrated embodiment, and sends it to data output 350 at the frequency of slow clock 306. In some embodiments, multi-clock FIFO 340 may have more channels in the data output than in the data input, for example if fast clock 306 has a frequency twice that of slow clock 304, multi-clock FIFO 340 may output twice as many channels of data.
Turning now to
In the example of
Commutator 420 may operate at the frequency of 80 MHz clock 406 and receive four channels of data from multi-clock FIFO 410. In the illustrated embodiment, there are four instances of overclock core 430, each of which may correspond to a channel of data coming from multi-clock FIFO 410. In some embodiments, commutator 420 may split the four channels of data and send each channel to a core. In some embodiments, commutator 420 may perform other actions, including sending data to more than one core, or some data to the same core, or other manners of distributing data.
In the illustrated embodiment, overclock cores 430 process data, send the processed data to de-commutator 425, and then to multi-clock FIFO 440, and data output 450 in a similar manner as previously described. De-commutator 425 may receive data from each instance of overclock core 430 and combine the data into four channel data at 80 MHz to send to multi-clock FIFO 440. Multi-clock FIFO 440 may then output the data as eight channel data at 40 MHz.
In the illustrated embodiment, parallelism detector 130 may have detected that eight instances of a function may be indicated. In the illustrated embodiment, 80 MHz clock 406 has twice the slow frequency of 40 MHz clock 404; accordingly, four blocks of overclock core 430 may be specified or instantiated.
In some embodiments, functional circuitry generator 160 may generate multiple configurations of functional circuitry for functional circuitry determiner 150 to use in determining the number of functional blocks. For example, in
In some embodiments, the number of functional blocks multiplied by the ratio of the fast clock to the slow clock may equal the number of instances of the function to be performed. For example, in embodiments where a function is indicated to be run eight times and a slow clock frequency may be 40 MHz, a fast clock may be 160 MHz with two functional blocks.
In some embodiments, functional circuitry determiner 150 may select between configurations, e.g.
In some embodiments, the area usable for functional circuitry may be a constraint, e.g. there may be 1000 area units available, and each functional block may require 100 area units, and interface circuitry may require 200 area units. There may be several ways to configure functional blocks and interface circuitry that will meet the area constraint, for example 400 area units of functional blocks and 200 area units of interface circuitry, 600 area units of functional blocks and 400 units of interface circuitry, etc. In some embodiments, additional constraints (e.g. constraints input by a user) may be used to choose between configurations. For example, the processing speed may be constrained such that at least 6 functional blocks may be required to process data at a rate that meets the constraint. As another example, power consumption may be specified as a constraint or may be indicated as a design parameter that is more important than other parameters (e.g., more important than performance for a specific task, so long as minimum performance requirements are met). In some embodiments, user input may not indicate constraints, but may indicate preferences, for example, the preference that configurations with lower power consumption may be chosen before configurations with faster processing. Other constraints or preferences based on characteristics of the circuitry may be specified, and other configurations may be possible; specific examples given here (e.g. 1000 area units) are not intended to limit the scope of the present disclosure.
At 610 in the illustrated embodiment, an indication of execution parallelism for a function is detected within a program. In some embodiments, the indication includes at least a first number of independent instances of the function that operate on different sets of input data. In some embodiments, the program may be written in a hardware description language. The program may comprise a text-based format, may be machine code, may be another format of program information, or combinations thereof. In some embodiments, detection is performed by a computer system without user input. In some embodiments, a user may specify characteristics of the program or the format of the program, or may select a subset of the program to be processed.
In some embodiments, different sets of data may include data that is similar in format, meaning, or content. In some embodiments, different sets of data may include data that comprises the same values or sets of values. Different functional blocks may process different sets of data sequentially or in parallel. In some embodiments, functional blocks may process similar quantities of input data or some functional blocks may process more or less data than other functional blocks.
At 620 in the illustrated embodiment, interface circuitry configured to provide the sets of input data to one or more functional blocks to perform the function is instantiated in a hardware element. In some embodiments, the interface circuitry is configured to receive data based on a first clock signal that has a first frequency. The first frequency may be specified by a user, detected in a program, or set as a default by a processing system, for example. In some embodiments, instantiation is performed without user input, or may require a user to input characteristics of the hardware to be instantiated. In some embodiments, the hardware element may be a programmable hardware element; in other embodiments it may be an ASIC.
At 630 in the illustrated embodiment, a second number of functional blocks configured to perform the function is instantiated in the hardware element. In some embodiments, the first number is different than the second number. At least one of the instantiated functional blocks may be configured to operate based on a second clock signal that has a second frequency that is greater than the first frequency, in some embodiments. At least one of the instantiated functional blocks may be configured to retrieve data from the interface circuitry based on the second clock signal. In some embodiments, the second clock signal may have a frequency that is a multiple of the first frequency. In some embodiments, the first number and/or first frequency is equal to the second number or second frequency. In some embodiments, the ratio of the second clock signal to the first clock signal may correspond to the number of functional blocks and the number of instances of the function. The ratio of the second frequency to the first frequency may correspond to the function performed by the functional blocks, the amount of input data to be processed, or other characteristics of the data, functional blocks, or interface circuitry.
In various embodiments, the disclosed techniques allow a compiler to automatically select an amount of parallelism for multi-clock circuitry based on various parameters, which may improve performance, reduce area, reduce design time, etc. while satisfying constraints (e.g., constraints imposed by hardware, performance requirements, cost, or other design constraints).
The method illustrated in
At 710 in the illustrated embodiment, an indication of execution parallelism for a function in a program is detected. The indication indicates a first number of independent instances of the function, in the illustrated embodiment. In some embodiments, the first number of instances of the function may operate on different sets of input data, or they may operate on the same set of input data. Input data may be acquired from one or more sources over one or more periods of time. In some embodiments, detection is performed by a computer system without user input. In some embodiments, a user may specify characteristics which indicate execution parallelism, or may select a set of functions which may indicate execution parallelism.
At 720 in the illustrated embodiment, hardware description data is generated that specifies configuration of interface circuitry configured to provide sets of input data to one or more functional blocks to perform the function from 710. In some embodiments, the set of instructions is written in a hardware description language. In the illustrated embodiment, the instructions indicate that the interface circuitry is configured to receive data based on a first clock signal that has a first frequency. In some embodiments, the instructions may indicate that data may be received from multiple sources, or may be received from one or more channels. The instruction may indicate that data input may be performed using additional circuitry or components, or may be provided directly to the interface circuitry. In some embodiments, the set of instructions may indicate a value for the first frequency, or other information may be used to indicate the first frequency. At 730 in the illustrated embodiment, hardware description data is generated that describes a second number of functional blocks configured to perform the function of 710. In some embodiments, the set of instructions is written in a hardware description language. In the illustrated embodiment, the instruction may indicate that the first number of instances of the function is different than the second number of functional blocks. At least one of the functional blocks is configured to operate based on a second clock signal that has a second frequency that is different than the first frequency and retrieve data from the interface circuitry based on the second clock signal, in the illustrated embodiment. In some embodiments, the instructions may indicate that the first number of instances of the function is equal to than the second number of functional blocks, or that first frequency is the same as the second frequency. In some embodiments, at least one of the functional blocks may be configured to perform the function at least twice on different input data.
In some embodiments, all of the functional blocks may be configured to operate based on the second clock signal, or some of the functional blocks may be configured to operate based on the second clock signal and other functional blocks may be configured to operate based on another clock signal. In some embodiments, the second number of functional blocks may be related to the first and second frequencies and the first number of instances of the function. For example, the product of the second number of functional blocks and the ratio of the second frequency to the first frequency may be equal to the first number of instances of the function. In some embodiments, the second number of functional blocks may be related to the frequencies and functions in other ways. In some embodiments, data may be retrieved from the interface circuitry using a different number of channels than data was input into the interface circuitry.
At 740 in the illustrated embodiment, the generated hardware description data is stored in a non-transitory computer-readable medium. Non-transitory computer-readable media may include optical disks, magnetic disk drives, solid state drives, magnetic tape, paper punch cards, or other media.
In some embodiments, the generated instructions may be used to instantiate hardware. Instantiated hardware may be generated based entirely on the instructions, or additional information may be used to instantiate hardware in some embodiments. For example, user input may specify frequencies of operation or numbers, functional blocks, or other circuitry. In some embodiments, the instructions may indicate to instantiate multiple sets of hardware and evaluate them for performance.
Referring now to
Interconnect 810 may include various interconnects, buses, MUX's, controllers, etc., and may be configured to facilitate communication between various elements of device 800. In some embodiments, portions of interconnect 810 may be configured to implement various different communication protocols. In other embodiments, interconnect 810 may implement a single communication protocol and elements coupled to interconnect 810 may convert from the single communication protocol to other communication protocols internally.
In the illustrated embodiment, processor 820 includes bus interface unit (BIU) 825, cache 830, and cores 835 and 840. In various embodiments, processor 820 may include various numbers of processors, processor cores and/or caches. For example, processor 820 may include 1, 2, or 4 processor cores, or any other suitable number. In one embodiment, cache 830 is a set associative L2 cache. In some embodiments, cores 835 and/or 840 may include internal instruction and/or data caches. In some embodiments, a coherency unit (not shown) in interconnect 810, cache 830, or elsewhere in device 800 may be configured to maintain coherency between various caches of device 800. BIU 825 may be configured to manage communication between processor 820 and other elements of device 800. Processor cores such as cores 835 and 840 may be configured to execute instructions of a particular instruction set architecture (ISA) which may include operating system instructions and user application instructions.
Cache/memory controller 845 may be configured to manage transfer of data between interconnect 810 and one or more caches and/or memories, including cache/memory 846. For example, cache/memory controller 845 may be coupled to an L3 cache, which may in turn be coupled to a system memory. In other embodiments, cache/memory controller 845 may be directly coupled to a memory. In some embodiments, cache/memory controller 845 may include one or more internal caches.
In the illustrated embodiment, cache/memory 846 contains code 848. In some embodiments, code 848 may be used to configure the computing system 800. In other embodiments, code 848 may include instructions for processor 820 to execute, such as instructions relating to the control of any of the systems or devices discussed above, or code 848 may include information directing the usage of I/O Bridge 850. Code 848 may include other information not described here, including but not limited to data, configurations for other components of computing system 800, or instructions to be executed by computing system 800.
As used herein, the term “coupled to” may indicate one or more connections between elements, and a coupling may include intervening elements. For example, in
Graphics/display unit 860 may include one or more processors and/or one or more graphics processing units (GPU's). Graphics/display unit 860 may receive graphics-oriented instructions, such as OPENGL® or DIRECT3D® instructions, for example. Graphics/display unit 860 may execute specialized GPU instructions or perform other operations based on the received graphics-oriented instructions. Graphics/display unit 860 may generally be configured to process large blocks of data in parallel and may build images in a frame buffer for output to a display. Graphics/display unit 860 may include transform, lighting, triangle, and/or rendering engines in one or more graphics processing pipelines. Graphics/display unit 860 may output pixel information for display images.
Graphics/display unit 860 may be configured to read data from a frame buffer and provide a stream of pixel values for display. Graphics/display unit 860 may be configured as a display pipeline in some embodiments. Additionally, Graphics/display unit 860 may be configured to blend multiple frames to produce an output frame. Further, Graphics/display unit 860 may include one or more interfaces (e.g., MIPI® or embedded display port (eDP)) for coupling to a user display (e.g., a touchscreen or an external display).
I/O bridge 850 may include various elements configured to implement: universal serial bus (USB) communications, security, audio, and/or low-power always-on functionality, for example. I/O bridge 850 may also include interfaces such as pulse-width modulation (PWM), general-purpose input/output (GPIO), serial peripheral interface (SPI), and/or inter-integrated circuit (I2C), for example. Various types of peripherals and devices may be coupled to device 800 via I/O bridge 850. In the illustrated embodiment, I/O Bridge 850 is coupled to storage device 852.
In some embodiments, storage device 852 may be a hard disk drive or solid state drive. Storage device 852 may be a tape drive, magnetic drive, removable media drive, etc. in some embodiments. In the illustrated embodiment, storage device 852 includes data 854.
Data 854 may include instructions for executing a compiler, instructions for instantiating hardware, a hardware description, program code, or other forms of data.
In some embodiments, any of various operations discussed herein may be performed by executing program instructions stored on a non-transitory computer readable medium. Such program instructions may be executed using system 800, for example. In these embodiments, the non-transitory computer-readable memory medium may be configured so that it stores program instructions and/or data, where the program instructions, if executed by a computer system, cause the computer system to perform a method, e.g., any of a method embodiments described herein, or, any combination of the method embodiments described herein, or, any subset of any of the method embodiments described herein, or, any combination of such subsets.
The present disclosure has described various exemplary circuits in detail above. It is intended that the present disclosure cover not only embodiments that include such circuitry, but also a computer-readable storage medium that includes design information that specifies such circuitry. Accordingly, the present disclosure is intended to support claims that cover not only an apparatus that includes the disclosed circuitry, but also a storage medium that specifies the circuitry in a format that is recognized by a fabrication system configured to produce hardware (e.g., an integrated circuit) that includes the disclosed circuitry. Claims to such a storage medium are intended to cover, for example, an entity that produces a circuit design, but does not itself fabricate the design.
Non-transitory computer-readable medium 910, may comprise any of various appropriate types of memory devices or storage devices. Medium 910 may be an installation medium, e.g., a CD-ROM, floppy disks, or tape device; a computer system memory or random access memory such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; a non-volatile memory such as a Flash, magnetic media, e.g., a hard drive, or optical storage; registers, or other similar types of memory elements, etc. Medium 910 may include other types of non-transitory memory as well or combinations thereof. Medium 910 may include two or more memory mediums which may reside in different locations, e.g., in different computer systems that are connected over a network.
Design information 915 may be specified using any of various appropriate computer languages, including hardware description languages such as, without limitation: VHDL, Verilog, SystemC, SystemVerilog, RHDL, M, MyHDL, etc. Design information 915 may be usable by semiconductor fabrication system 920 to fabricate at least a portion of integrated circuit 930. The format of design information 915 may be recognized by at least one semiconductor fabrication system 920. In some embodiments, design information 915 may also include one or more cell libraries which specify the synthesis and/or layout of integrated circuit 930. In some embodiments, the design information is specified in whole or in part in the form of a netlist that specifies cell library elements and their connectivity.
Semiconductor fabrication system 920 may include any of various appropriate elements configured to fabricate integrated circuits. This may include, for example, elements for depositing semiconductor materials (e.g., on a wafer, which may include masking), removing materials, altering the shape of deposited materials, modifying materials (e.g., by doping materials or modifying dielectric constants using ultraviolet processing), etc. Semiconductor fabrication system 920 may also be configured to perform various testing of fabricated circuits for correct operation.
In various embodiments, integrated circuit 930 is configured to operate according to a circuit design specified by design information 915, which may include performing any of the functionality described herein. For example, integrated circuit 930 may include any of various elements shown in
Medium 940 and design information 945 may have features similar to medium 910 and design information 915, as discussed above. Hardware description languages used to design ASICs may be similar or different than those used to program programmable hardware elements. Programmable hardware element 960 may be a field-programmable gate array (FPGA), programmable logic array (PLA), complex programmable logic device (CPLD) etc. Programmable hardware element 960 may include logic blocks, hard blocks for common functions, configurable clocking structures, memories, fuses, etc. A given programmable hardware element 960 may be programmed differently at different times, e.g., by adjusting the functionality of the logic blocks, interconnections between circuit elements, etc.
In various embodiments, programmable hardware element 960 is configured, after being programmed, to operate according to the circuit design specified by design information 945, which may include performing any of the functionality described herein. For example, programmable hardware element 960 may implement any of various elements shown in
As used herein, the term “implement the circuit according to the design” includes both fabricating an integrated circuit according to the design and programming a programmable hardware element according to the design. Semiconductor fabrication system 920 and programming device 950 are examples of computing systems configured to implement circuits according to design information. Speaking generally, implementing a circuit according to a design may include other ways to implement hardware circuits in addition to the techniques discussed with references to
As used herein, a phrase of the form “design information that specifies a design of a circuit configured to . . . ” does not imply that the circuit in question must be fabricated in order for the element to be met. Rather, this phrase indicates that the design information describes a circuit that, upon being fabricated, will be configured to perform the indicated actions or will include the specified components.
Although specific embodiments have been described above, these embodiments are not intended to limit the scope of the present disclosure, even where only a single embodiment is described with respect to a particular feature. Examples of features provided in the disclosure are intended to be illustrative rather than restrictive unless stated otherwise. The above description is intended to cover such alternatives, modifications, and equivalents as would be apparent to a person skilled in the art having the benefit of this disclosure.
The scope of the present disclosure includes any feature or combination of features disclosed herein (either explicitly or implicitly), or any generalization thereof, whether or not it mitigates any or all of the problems addressed herein. Accordingly, new claims may be formulated during prosecution of this application (or an application claiming priority thereto) to any such combination of features. In particular, with reference to the appended claims, features from dependent claims may be combined with those of the independent claims and features from respective independent claims may be combined in any appropriate manner and not merely in the specific combinations enumerated in the appended claims.
