Patent Grant
10216487
The field of invention pertains generally to image processing, and, more specifically, to a virtual image processing instruction set architecture (ISA) and memory model and exemplary target hardware having a two-dimensional shift array structure.
Image processing typically involves the processing of pixel values that are organized into an array. Here, a spatially organized two dimensional array captures the two dimensional nature of images (additional dimensions may include time (e.g., a sequence of two dimensional images) and data type (e.g., colors). In a typical scenario, the arrayed pixel values are provided by a camera that has generated a still image or a sequence of frames to capture images of motion. Traditional image processors typically fall on either side of two extremes.
A first extreme performs image processing tasks as software programs executing on a general purpose processor or general purpose-like processor (e.g., a general purpose processor with vector instruction enhancements). Although the first extreme typically provides a highly versatile application software development platform, its use of finer grained data structures combined with the associated overhead (e.g., instruction fetch and decode, handling of on-chip and off-chip data, speculative execution) ultimately results in larger amounts of energy being consumed per unit of data during execution of the program code.
A second, opposite extreme applies fixed function hardwired circuitry to much larger blocks of data. The use of larger (as opposed to finer grained) blocks of data applied directly to custom designed circuits greatly reduces power consumption per unit of data. However, the use of custom designed fixed function circuitry generally results in a limited set of tasks that the processor is able to perform. As such, the widely versatile programming environment (that is associated with the first extreme) is lacking in the second extreme.
A technology platform that provides for both highly versatile application software development opportunities combined with improved power efficiency per unit of data remains a desirable yet missing solution.
A method is described that includes instantiating, within an application software development environment, a virtual processor having an instruction set architecture and memory model that contemplate first and second regions of reserved memory. The first reserved region is to keep data of an input image array. The second reserved region is to keep data of an output image array. The method also includes simulating execution of a memory load instruction of the instruction set architecture by automatically targeting the first reserved region and identifying desired input data with first and second coordinates relative to the virtual processor's position within an orthogonal coordinate system and expressed in the instruction format of the memory load instruction.
The following description and accompanying drawings are used to illustrate embodiments of the invention. In the drawings:
i. Introduction
The description below describes numerous embodiments concerning a new image processing technology platform that provides a widely versatile application software development environment that can target a hardware architecture that uses larger blocks of data (e.g., line buffers and sheets as described further below) to provide for improved power efficiency.
1.0 Application Software Development Environment
a. Application and Structure of Kernels
For example, as observed in
As such, the larger overall image processing sequence may take the form of a image processing pipeline or a directed acyclic graph (DAG) and the development environment may be equipped to actually present the developer with a representation of the program code being developed as such. Kernels may be developed by a developer individually and/or may be provided by an entity that supplies any underlying technology (such as the actual signal processor hardware and/or a design thereof) and/or by a third party (e.g., a vendor of kernel software written for the development environment). As such, it is expected that a nominal development environment will include a “library” of kernels that developers are free to “hook-up” in various ways to effect the overall flow of their larger development effort. Some basic kernels that are expected to be part of such a library may include kernels to provide any one or more of the following basic image processing tasks: convolutions, denoising, color space conversions, edge and corner detection, sharpening, white balance, gamma correction, tone mapping, matrix multiply, image registration, pyramid construction, wavelet transformation, block-wise discrete cosine and Fourier transformations.
As will be described in more detail below, in various embodiments, in the actual underlying hardware an array of execution lanes and corresponding threads operate in unison (e.g., in a Single Instruction Multiple Data (SIMD) like fashion) to generate output image data for a portion of a “line group” of the frame currently being processed. A line group is a contiguous, sizable section of an image frame. In various embodiments, the developer may be conscious the hardware operates on line groups, or, the development environment may present an abstraction in which there is a separate processor and thread for, e.g., every pixel in the output frame (e.g., every pixel in an output frame generated by its own dedicated processor and thread). Regardless, in various embodiments, the developer understands the kernel to include an individual thread for each output pixel (whether the output array is visualized is an entire output frame or a section thereof).
As will be described in more detail below, in an embodiment the processors 205 that are presented to the developer in the virtual environment have an instruction set architecture (ISA) that, not only supports standard (e.g., RISC) opcodes, but also include specially formatted data access instructions that permit the developer to easily visualize the pixel by pixel processing that is being performed. The ability to easily define/visualize any input array location in combination with an entire ISA of traditional mathematical and program control opcodes allows for an extremely versatile programming environment that essentially permits an application program developer to define, ideally, any desired function to be performed on any sized image surface. For example, ideally, any mathematical operation can be readily programmed to be applied to any stencil size.
With respect to the data access instructions, in an embodiment the ISA of the virtual processors (“virtual ISA”) include a special data load instruction and a special data store instruction. The data load instruction is able to read from any location within an input array of image data. The data store instruction is able to write to any location within the output array of image data. The latter instruction allows for easily dedicating multiple instances of the same processor to different output pixel locations (each processor writes to a different pixel in the output array). As such, for example, stencil size itself (e.g., expressed as a width of pixels and a height of pixels) can be made an easily programmable feature. Visualization of the processing operations is further simplified with each of the special load and store instructions having a special instruction format whereby target array locations are specified simplistically as X and Y coordinates.
Regardless, by instantiating a separate processor for each of multiple locations in the output array, the processors can execute their respective threads in parallel so that, e.g., the respective values for all locations in the output array are produced concurrently. It is noteworthy that many image processing routines typically perform the same operations on different pixels of the same output image. As such, in one embodiment of the development environment, each processor is presumed to be identical and executes the same thread program code. Thus, the virtualized environment can be viewed as a type of two-dimensional (2D), SIMD processor composed of a 2D array of, e.g., identical processors each executing identical code in lock-step.
As is known in the art, an output pixel value is often determined by processing the pixels of an input array that include and surround the corresponding output pixel location. For example, as can be seen from
As can be seen from the respective program code from the pair of threads 301, 302, each virtual processor at least includes internal registers R1 and R2 and at least supports the following instructions: 1) a LOAD instruction from the input array into R1; 2) a LOAD instruction from the input array into R2; 3) an ADD instruction that adds the contents of R1 and R2 and places the resultant in R2; 4) a DIV instruction that divides the value within R2 by immediate operand 9; and, 5) a STORE instruction the stores the contents of R2 into the output array location that the thread is dedicated to. Again, although only two output array locations and only two threads and corresponding processors are depicted in
b. Virtual Processor Memory Model
In various embodiments, a pertinent feature of the virtual processors is their memory model. As is understood in the art, a processor reads data from memory, operates on that data and writes new data back into memory. A memory model is the perspective or view that a processor has of the manner in which data is organized in memory.
As observed in
The partitions as depicted in
With respect to the scratch pad region 402, it is not uncommon to temporarily store intermediate information over the course of execution of a sophisticated image processing algorithm (e.g., and then read the information back and use it a later time). Additionally, it is not uncommon for such information to be different across threads (different input values may effect different intermediate values). The memory model therefore includes per processor private scratchpad regions 402 for the storage of such intermediate information by each virtual processor's corresponding thread. In an embodiment, the scratch pad region for a particular processor is accessed 409 by that processor through a typical (e.g., linear) random access memory address and is a read/write region of memory (i.e., a virtual processor is able to both read information from private memory as well as write information into private memory). Embodiments of the virtual processor ISA instruction formal for accessing the scratchpad region are discussed in more detail further below.
The input array portion 403 contains the set of input data that is called into 408 the set of threads in order to produce output data. In a typical situation the input array corresponds to an image (e.g., a frame) or section of an image that each thread is operating on or within. The input image may be a true input such as the pixel information provided by a camera, or, some form of intermediate image such as the information provided by a previous kernel in a larger overall image processing sequence. Virtual processors typically do not compete for same input data items because they operate on different pixel locations of the input image data during a same cycle.
In an embodiment, a novel memory addressing scheme is used to define which particular input values are to be called in from the input array 403. Specifically, a “position relative” addressing scheme is used that defines the desired input data with X, Y coordinates rather than a traditional linear memory address. As such, the load instruction of the virtual processors' ISA includes an instruction format that identifies a specific memory location within the input array with an X component and a Y component. As such, a two-dimensional coordinate system is used to address memory for input values read from the input array 403.
The use of a position relative memory addressing approach permits the region of an image that a virtual processor is operating on to be more readily identifiable to a developer. As mentioned above, the ability to easily define/visualize any input array location in combination with an entire ISA of traditional mathematical and program control opcodes allows for an extremely versatile programming environment that essentially permits an application program developer to readily define, ideally, any desired function to be performed on any sized image surface. Various instruction format embodiments for instructions that adopt a position relative addressing scheme, as well as embodiments of other features of the supported ISA, are described in more detail further below.
The output array 404 contains the output image data that the threads are responsible for generating. The output image data may be final image data such as the actual image data that is presented on a display that follows the overall image processing sequence, or, may be intermediate image data that a subsequent kernel of the overall image processing sequence uses as its input image data information. Again, typically virtual processors do not compete for same output data items because they write to different pixel locations of the output image data during a same cycle.
In an embodiment, the position relative addressing scheme is also used for writes to the output array. As such, the ISA for each virtual processor includes a store instruction whose instruction format defines a targeted write location in memory as a two-dimensional X, Y coordinate rather than a traditional random access memory address. More details concerning embodiments of the position relative instructions of the virtual ISA are provided further below.
In an embodiment, as index values are typically used to define a desired look-up table entry, the look-up table information region is accessed using a normal linear accessing scheme. In an embodiment the look-up region of memory is read only (i.e., the processor can not change information in a look-up table and is only permitted to read information from it). For simplicity
In these sequences, as observed in
c. Virtual Processor ISA
As alluded to above in multiple instances, the virtual processor ISA may include a number of pertinent features. Some of these described at length immediately below.
In various embodiment the instruction format of each virtual processor's ISA uses a relative positioning approach to define an X, Y coordinate for each of the following: 1) a LOAD instruction that reads input image data from the input array memory region; 2) a STORE instruction that writes output image data to the output array; and, 3) an atomic update to the statistics region of memory.
The ability to easily define any input array location in combination with an entire ISA of traditional data access, mathematical and program control opcodes allows for an extremely versatile programming environment that essentially permits an application program developer to define, ideally, any desired function to be performed on any sized image surface. For example, ideally, any mathematical operation can be readily programmed to be applied to any stencil size.
In an embodiment, instructions for loads/stores from/to the input/output arrays have the following format
[OPCODE] LINEGROUP_(name)[(((X*XS+X0)/XD);((Y*YS+Y0)/YD);Z]
where [OPCODE] is the specific type of operation (LOAD from the input array, STORE to the output array) and LINEGROUP_(name) is the name assigned to a particular section of a particular image (e.g., a line group for a frame of image data) within the input or output array memory region. Here, because different line groups are operated on separately, the different linegroups are given different names so they can be uniquely identified/accessed (e.g., LINEGROUP_1, LINEGROUP_2, etc.). Line groups of same name may exist in both the input array memory region and the output array memory region. The origin of any line group may be, e.g., its lower left hand corner within its appropriate memory region.
In the case of instructions that perform updates on the atomic statistics table, in an embodiment, the instruction format takes on the following similar structure
[OPCODE] STATS_(name)[(((X*XS+X0)/XD);((Y*YS+Y0)/YD);Z]
with the notable difference being that the input operand information defines a position within a particular statistics table (STATS_(name)) rather than a particular line group within the input or output array. As with line groups, different names are given to different statistics tables so that a thread can uniquely operate on different statistics tables over the course of its operation. The [OPCODE] specifies the particular atomic act to be performed (e.g., STAT_ADD; STAT_MUL; STAT_MIN; STAT_MAX).
For either input/output array accesses or atomic statistics table accesses, the Z operand of the instruction defines which channel of a named line group or stats table is targeted by the instruction. Here, typically, a single image will have multiple channels. For example, video images typically have a red channel (R), a green channel (G) and a blue channel (B) for a same frame of the video stream. In a sense, a complete image can be viewed as separate R, G and B channel images stacked on top of each other. The Z operand defines which one of these is targeted by the instruction (e.g., Z=0 corresponds to the red channel, Z=1 corresponds to the blue channel and Z=2 corresponds to the green channel). Each line group and statistics table is therefore structured to include the content of each channel for the particular image being processed.
The (X*XS+X0)/XD operand defines the X location within a named line group or stats table that is targeted by the instruction and the (Y*YS+Y0)/YD operand defines the Y location within a named line group or stats table that is targeted by the instruction. The XS and XD terms for the X location and the YS and YD terms for the Y location are used for scaling between input and output images having different pixel densities. Scaling is described in more detail further below.
In a simplest case, there is no scaling between input and output images and the X and Y components of the instruction format simply take the form of X+X0 and Y+Y0 where X0 and Y0 are positional offsets relative to the position of the thread. A thread is viewed as being assigned to the position within the output array line group that its output value is written to. A corresponding, same position is readily identifiable in the input array line group and any stats table.
As an example, if the thread is assigned a specific X, Y location in an output array LINEGROUP_1, the instruction
LOAD LINEGROUP_1[(X−1);(Y−1);Z]
would load from LINEGROUP_1 a value that is to the left one pixel location and down one pixel location from the same X,Y position within the input array.
A simple blur kernel that averages together the pixel values for the X,Y location along with its left and right neighbors may therefore be written in pseudo-code as depicted in
For example, referring to
As such, the XS and YS multiplicand terms in the relative positioning instruction format are used to implement down sampling. If the blur pseudo code of
R1<=LOAD LINEGROUP_1[((3X)−1);3(Y);0]
R2<=LOAD LINEGROUP_1[3(X);3(Y);0]
R3<=LOAD LINEGROUP_1[((3X)+1);3(Y);0]
R2<=ADD R1,R2
R2<=ADD R2,R3
R2<=DIV R2,3
STORE LINEGROUP_1[(X);(Y);(0)];R2.
By contrast, in the case of 1:3 up-sampling (e.g., image 502 is the input image and image 501 is the output image) the XD and YD divisors would be used to create three output pixels for every input pixel along either axis. As such, the blur code would be rewritten as:
R1<=LOAD LINEGROUP_1[(X−1)/3;(Y)/3;0]
R2<=LOAD LINEGROUP_1[(X)/3;(Y)/3;0]
R3<=LOAD LINEGROUP_1[(X+1)/3;(Y)/3;0]
R2<=ADD R1,R2
R2<=ADD R2,R3
R2<=DIV R2,3
STORE LINEGROUP_1[(X);(Y);(0)];R2
In various embodiments the instruction format for instructions that access the private, constant and look-up portions of memory include an operand that also takes the form of a*b+c where a is a base position, b is a scaling term and c is an offset. Here, however, a linear addressing approach is taken where the a*b+c term essentially corresponds to a linear index that is applied to the targeted table. Each of these instructions also include in the opcode and an identifier of the memory region being accessed. For example, an instruction that performs a look-up from the look-up table memory region may be expressed as
LOAD LKUP_(name)[(A*B+C)].
where LOAD is the opcode that identifies a load operation and LKUP_(name) specifies the name of the look-up table in the look-up table memory region being accessed. Again, multiple look-up tables may be used by a thread and therefore a naming scheme is used to identify the appropriate one of the more than one that exist in the look-up table memory region.
A similar format with similarly minded opcode may be utilized for instructions that target the constant and the private memory regions (e.g., LOAD CNST_(name)[(A*B+C)]; LOAD PRVT_(name)[(A*B+C)]. In an embodiment, look-up table and the constant table accesses are read-only (a processor can not change the data that has been placed there). As such no STORE instructions exist for these memory regions. In an embodiment the private region of memory is read/write. As such a store instruction exists for that memory region (e.g., STORE PRVT[(A*B+C)].
In various embodiments, each virtual processor includes general purpose registers that can contain integer, floating point or fixed point values. Additionally, the general purpose registers may contain data values of configurable bit width such as 8, 16 or 32 bit values. Thus, the image data at each pixel location in an input array or output array can have a data size of 8, 16 or 32 bits. Here, a virtual processor can be configured for an execution mode that establishes the bit size and the numerical format of the values within the general purpose registers. Instructions may also specify immediate operands (which are input operands whose input values are expressed directly in the instruction itself rather being found in a specified register). Immediate operands can also have configurable 8, 16 or 32 bit widths.
In an extended embodiment, each virtual processor is also capable of operating in a scalar mode or a SIMD mode internal to itself. That is, the data within a specific array location may be viewed as a scalar value or as a vector having multiple elements. For example a first configuration may establish scalar operation of 8 bits where each image array position holds a scalar 8 bit value. By contrast another configuration may establish parallel/SIMD operation of 32 bits where each image array location is assumed to hold four 8 bit values for a total data size of 32 bits per array location.
In various embodiments each virtual processor also includes registers to hold predicate values. A single predicate value is often only one bit in length and expresses a resultant from an opcode that performs a true/false or greater than/less than test on existing data. Predicate values are used, e.g., to determine branch directions through the code during execution (and therefore are used as operands in conditional branch instructions). Predicate values can also be expressed as an immediate operand in an instruction.
In various embodiments each virtual processor includes registers to hold scalar values. Here, scalar values are stored into and read from the partition space of the memory model that is reserved for constants (as discussed above with respect to
In various embodiments each virtual processor is designed as a RISC-like instruction set whose supported arithmetic instruction opcodes include any workable combination of the following: 1) ADD (addition of operands A and B); 2) SUB (subtraction of operands A and B); 3) MOV (move operand from one register to another register); 4) MUL (multiple operands A and B); 5) MAD (multiply operands A and B and add C to resultant); 6) ABS (return absolute value of operand A); 7) DIV (divide operand A by operand B); 8) SHL (shift operand A to the left); 9) SHR (shift operand A to the right); 10) MIN/MAX (return which of operands A and B is greater); 11) SEL (select specified bytes of operand A); 12) AND (return the logical AND of operands A and B); 13) OR (return the logical OR of operands A and B); 14) XOR (return the logical exclusive OR of operands A and B); 15) NOT (return the logical inverse of operand A).
The instruction set also includes standard predicate operations such as: 1) SEQ (returns a 1 if A equals B); 2) SNE (returns a 1 if A does not equal B); 3) SLT (returns a 1 if A is less than B); 4) SLE (returns a 1 if A is less than or equal to B). Control flow instructions are also included such as JMP (jump) and BRANCH each of which may include nominal variables or predicates as operands.
d. Application Software Development and Simulation Environment
Kernels within the library 602 may be provided by a third party vendor of kernels and/or a provider of any underlying technology (e.g., a vendor of a hardware platform that includes the targeted hardware image processor or a vendor of the targeted hardware image processor (e.g., provided as a design thereof or as actual hardware)).
In the case of custom developed kernels, in many situations the developer need only write the program code for a single thread 603. That is, the developer need only write program code that determines a single output pixel value by referencing input pixel values relative to the output pixel location (e.g., with the aforementioned position relative memory access instruction format). Upon satisfaction of the operation of the single thread 603, the development environment may then automatically instantiate multiple instances of the thread code on a respective virtual processor to effect a kernel on an array of processors that operate on an image surface area. The image surface area may be a section of an image frame (such as a line group).
In various embodiments, the custom thread program code is written in the object code of the virtual processor ISA (or a higher level language that is compiled down to the virtual processor ISA object code). Simulation of execution of the custom kernel's program code may be performed in a simulated runtime environment that includes a virtual processor accessing a memory organized according to the memory model. Here, software models (object oriented or otherwise) of a virtual processor 604 and a memory 605 that incorporates the model are instantiated.
The virtual processor model 604 then simulates execution of the thread code 603. Upon satisfaction of the performance of a thread, its larger kernel and any larger function that the kernel belongs to, the whole is compiled into the actual object code of the underlying hardware. The entirety of the simulation environment 601 may be implemented as software that runs on a computer system (e.g., a workstation) 606.
The following section describes embodiments of an exemplary target hardware architecture for application software written within an application software development environment having various features of the environment discussed above. It is pertinent to point out that the target hardware described below is only exemplary and conceivably many other alternative hardware architectures could be targeted by code written within the afore-described software development environment.
2.0 Hardware Architecture Embodiments
a. Image Processor Hardware Architecture and Operation
In an embodiment, program code is compiled and loaded onto a corresponding stencil processor 702 to perform the image processing operations earlier defined by a software developer (program code may also be loaded onto the stencil processor's associated sheet generator 703, e.g., depending on design and implementation). In at least some instances an image processing pipeline may be realized by loading a first kernel program for a first pipeline stage into a first stencil processor 702_1, loading a second kernel program for a second pipeline stage into a second stencil processor 702_2, etc. where the first kernel performs the functions of the first stage of the pipeline, the second kernel performs the functions of the second stage of the pipeline, etc. and additional control flow methods are installed to pass output image data from one stage of the pipeline to the next stage of the pipeline.
In other configurations, the image processor may be realized as a parallel machine having two or more stencil processors 702_1, 702_2 operating the same kernel program code. For example, a highly dense and high data rate stream of image data may be processed by spreading frames across multiple stencil processors each of which perform the same function.
In yet other configurations, essentially any DAG of kernels may be loaded onto the hardware processor by configuring respective stencil processors with their own respective kernel of program code and configuring appropriate control flow hooks into the hardware to direct output images from one kernel to the input of a next kernel in the DAG design.
As a general flow, frames of image data are received by a macro I/O unit 705 and passed to one or more of the line buffer units 701 on a frame by frame basis. A particular line buffer unit parses its frame of image data into a smaller region of image data, referred to as a “a line group”, and then passes the line group through the network 704 to a particular sheet generator. A complete or “full” singular line group may be composed, for example, with the data of multiple contiguous complete rows or columns of a frame (for simplicity the present specification will mainly refer to contiguous rows). The sheet generator further parses the line group of image data into a smaller region of image data, referred to as a “sheet”, and presents the sheet to its corresponding stencil processor.
In the case of an image processing pipeline or a DAG flow having a single input, generally, input frames are directed to the same line buffer unit 701_1 which parses the image data into line groups and directs the line groups to the sheet generator 703_1 whose corresponding stencil processor 702_1 is executing the code of the first kernel in the pipeline/DAG. Upon completion of operations by the stencil processor 702_1 on the line groups it processes, the sheet generator 703_1 sends output line groups to a “downstream” line buffer unit 701_2 (in some use cases the output line group may be sent back to the same line buffer unit 701_1 that earlier had sent the input line groups).
One or more “consumer” kernels that represent the next stage/operation in the pipeline/DAG executing on their own respective other sheet generator and stencil processor (e.g., sheet generator 703_2 and stencil processor 702_2) then receive from the downstream line buffer unit 701_2 the image data generated by the first stencil processor 702_1. In this manner, a “producer” kernel operating on a first stencil processor has its output data forwarded to a “consumer” kernel operating on a second stencil processor where the consumer kernel performs the next set of tasks after the producer kernel consistent with the design of the overall pipeline or DAG.
A stencil processor 702 is designed to simultaneously operate on multiple overlapping stencils of image data. The multiple overlapping stencils and internal hardware processing capacity of the stencil processor effectively determines the size of a sheet. Here, within a stencil processor 702, arrays of execution lanes operate in unison to simultaneously process the image data surface area covered by the multiple overlapping stencils.
As will be described in more detail below, in various embodiments, sheets of image data are loaded into a two-dimensional register array structure within the stencil processor 702. The use of sheets and the two-dimensional register array structure is believed to effectively provide for power consumption improvements by moving a large amount of data into a large amount of register space as, e.g., a single load operation with processing tasks performed directly on the data immediately thereafter by an execution lane array. Additionally, the use of an execution lane array and corresponding register array provide for different stencil sizes that are easily programmable/configurable.
Because of the vertical overlapping stencils 802 within the stencil processor, as observed in
With the line group 803 of the input image data having been defined by the line buffer unit and passed to the sheet generator unit, the sheet generator unit further parses the line group into finer sheets that are more precisely fitted to the hardware limitations of the stencil processor. More specifically, as will be described in more detail further below, in an embodiment, each stencil processor consists of a two dimensional shift register array. The two dimensional shift register array essentially shifts image data “beneath” an array of execution lanes where the pattern of the shifting causes each execution lane to operate on data within its own respective stencil (that is, each execution lane processes on its own stencil of information to generate an output for that stencil). In an embodiment, sheets are surface areas of input image data that “fill” or are otherwise loaded into the two dimensional shift register array.
Thus, as observed in
As observed in
Note that there is some overlap between the data of the first sheet 804 and the data of the second sheet 805 owing to the border regions of stencils that surround an output pixel location. The overlap could be handled simply by the sheet generator re-transmitting the overlapping data twice. In alternate implementations, to feed a next sheet to the stencil processor, the sheet generator may proceed to only send new data to the stencil processor and the stencil processor reuses the overlapping data from the previous sheet.
b. Stencil Processor Design and Operation
The I/O unit 904 is responsible for loading “input” sheets of data received from the sheet generator into the data computation unit 901 and storing “output” sheets of data from the stencil processor into the sheet generator. In an embodiment the loading of sheet data into the data computation unit 901 entails parsing a received sheet into rows/columns of image data and loading the rows/columns of image data into the two dimensional shift register structure 906 or respective random access memories 907 of the rows/columns of the execution lane array (described in more detail below). If the sheet is initially loaded into memories 907, the individual execution lanes within the execution lane array 905 may then load sheet data into the two-dimensional shift register structure 906 from the random access memories 907 when appropriate (e.g., as a load instruction just prior to operation on the sheet's data). Upon completion of the loading of a sheet of data into the register structure 906 (whether directly from a sheet generator or from memories 907), the execution lanes of the execution lane array 905 operate on the data and eventually “write back” finished data as a sheet directly back to the sheet generator, or, into the random access memories 907. If the later the I/O unit 904 fetches the data from the random access memories 907 to form an output sheet which is then forwarded to the sheet generator.
The scalar processor 902 includes a program controller 909 that reads the instructions of the stencil processor's program code from scalar memory 903 and issues the instructions to the execution lanes in the execution lane array 905. In an embodiment, a single same instruction is broadcast to all execution lanes within the array 905 to effect a SIMD-like behavior from the data computation unit 901. In an embodiment, the instruction format of the instructions read from scalar memory 903 and issued to the execution lanes of the execution lane array 905 includes a very-long-instruction-word (VLIW) type format that includes more than one opcode per instruction. In a further embodiment, the VLIW format includes both an ALU opcode that directs a mathematical function performed by each execution lane's ALU (which, as described below, in an embodiment may specify more than one traditional ALU operation) and a memory opcode (that directs a memory operation for a specific execution lane or set of execution lanes).
The term “execution lane” refers to a set of one or more execution units capable of executing an instruction (e.g., logic circuitry that can execute an instruction). An execution lane can, in various embodiments, include more processor-like functionality beyond just execution units, however. For example, besides one or more execution units, an execution lane may also include logic circuitry that decodes a received instruction, or, in the case of more MIMD-like designs, logic circuitry that fetches and decodes an instruction. With respect to MIMD-like approaches, although a centralized program control approach has largely been described herein, a more distributed approach may be implemented in various alternative embodiments (e.g., including program code and a program controller within each execution lane of the array 905).
The combination of an execution lane array 905, program controller 909 and two dimensional shift register structure 906 provides a widely adaptable/configurable hardware platform for a broad range of programmable functions. For example, application software developers are able to program kernels having a wide range of different functional capability as well as dimension (e.g., stencil size) given that the individual execution lanes are able to perform a wide variety of functions and are able to readily access input image data proximate to any output array location.
Apart from acting as a data store for image data being operated on by the execution lane array 905, the random access memories 907 may also keep one or more look-up tables such as any look-up tables held in the look-up table component of the virtual processing memory described above in Section 1.0. In various embodiments one or more scalar look-up tables may also be instantiated within the scalar memory 903. The one or more scalar look-up tables may be any scalar look-up tables held in the scalar look-up table component of the memory model described above in Section 1.0.
A scalar look-up involves passing the same data value from the same look-up table from the same index to each of the execution lanes within the execution lane array 905. In various embodiments, the VLIW instruction format described above is expanded to also include a scalar opcode that directs a look-up operation performed by the scalar processor into a scalar look-up table. The index that is specified for use with the opcode may be an immediate operand or fetched from some other data storage location. Regardless, in an embodiment, a look-up from a scalar look-up table within scalar memory essentially involves broadcasting the same data value to all execution lanes within the execution lane array 905 during the same clock cycle. Additional details concerning the use and operation of look-up tables is provided further below.
A field 954 for one or more immediate operands is also included. Which of the instructions 951, 952, 953 use which immediate operand information may be identified in the instruction format. Each of instructions 951, 952, 953 also include their own respective input operand and resultant information (e.g., local registers for ALU operations and a local register and a memory address for memory access instructions). In an embodiment, the scalar instruction 951 is executed by the scalar processor before the execution lanes within the execution lane array execute either of the other to instructions 952, 953. That is, the execution of the VLIW word includes a first cycle upon which the scalar instruction 951 is executed followed by a second cycle upon with the other instructions 952, 953 may be executed (note that in various embodiments instructions 952 and 953 may be executed in parallel).
In an embodiment, the scalar instructions executed by the scalar processor include commands issued to the sheet generator to load/store sheets from/into the memories or 2D shift register of the data computation unit. Here, the sheet generator's operation can be dependent on the operation of the line buffer unit or other variables that prevent pre-runtime comprehension of the number of cycles it will take the sheet generator to complete any command issued by the scalar processor. As such, in an embodiment, any VLIW word whose scalar instruction 951 corresponds to or otherwise causes a command to be issued to the sheet generator also includes no-operation (NOOP) instructions in the other two instruction field 952, 953. The program code then enters a loop of NOOP instructions for instruction fields 952, 953 until the sheet generator completes its load/store to/from the data computation unit. Here, upon issuing a command to the sheet generator, the scalar processor may set a bit of an interlock register that the sheet generator resets upon completion of the command. During the NOOP loop the scalar processor monitors the bit of the interlock bit. When the scalar processor detects that the sheet generator has completed its command normal execution begins again.
The execution lane array 1005 and shift register structure 1006 are fixed in position relative to one another. However, the data within the shift register array 1006 shifts in a strategic and coordinated fashion to cause each execution lane in the execution lane array to process a different stencil within the data. As such, each execution lane determines the output image value for a different pixel in the output sheet being generated. From the architecture of
Some notable architectural features of the data computation unit 1001 include the shift register structure 1006 having wider dimensions than the execution lane array 1005. That is, there is a “halo” of registers 1009 outside the execution lane array 1005. Although the halo 1009 is shown to exist on two sides of the execution lane array, depending on implementation, the halo may exist on less (one) or more (three or four) sides of the execution lane array 1005. The halo 1005 serves to provide “spill-over” space for data that spills outside the bounds of the execution lane array 1005 as the data is shifting “beneath” the execution lanes 1005. As a simple case, a 5×5 stencil centered on the right edge of the execution lane array 1005 will need four halo register locations further to the right when the stencil's leftmost pixels are processed. For ease of drawing,
Additional spill-over room is provided by random access memories 1007 that are coupled to each row and/or each column in the array, or portions thereof (E.g., a random access memory may be assigned to a “region” of the execution lane array that spans 4 execution lanes row wise and 2 execution lanes column wise. For simplicity the remainder of the application will refer mainly to row and/or column based allocation schemes). Here, if an execution lane's kernel operations require it to process pixel values outside of the two-dimensional shift register array 1006 (which some image processing routines may require) the plane of image data is able to further spill-over, e.g., from the halo region 1009 into random access memory 1007. For example, consider a 6×6 stencil where the hardware includes a halo region of only four storage elements to the right of an execution lane on the right edge of the execution lane array. In this case, the data would need to be shifted further to the right off the right edge of the halo 1009 to fully process the stencil. Data that is shifted outside the halo region 1009 would then spill-over to random access memory 1007. Other applications of the random access memories 1007 and the stencil processor of
Each execution lane also has available, in a local register R2, the contents “beneath” it in the two dimensional shift array. Thus, R1 is a physical register of the execution lane while R2 is a physical register of the two dimensional shift register array. The execution lane includes an ALU that can operate on operands provided by R1 and/or R2. As will be described in more detail further below, in an embodiment the shift register is actually implemented with multiple (a “depth” of) storage/register elements per array location but the shifting activity is limited to one plane of storage elements (e.g., only one plane of storage elements can shift per cycle).
As observed initially in
As observed in
As observed in
In the example of
In an embodiment, the two dimensional shift register structure is implemented by permitting, during a single cycle, the contents of any of (only) one of registers R2 through R4 to be shifted “out” to one of its neighbor's register files through output multiplexer 1203, and, having the contents of any of (only) one of registers R2 through R4 replaced with content that is shifted “in” from a corresponding one if its neighbors through input multiplexers 1204 such that shifts between neighbors are in a same direction (e.g., all execution lanes shift left, all execution lanes shift right, etc.). Although it may be common for a same register to have its contents shifted out and replaced with content that is shifted in on a same cycle, the multiplexer arrangement 1203, 1204 permits for different shift source and shift target registers within a same register file during a same cycle.
As depicted in
Although in one embodiment the content of only one register is permitted to be shifted per execution lane per cycle, other embodiments may permit the content of more than one register to be shifted in/out. For example, the content of two registers may be shifted out/in during a same cycle if a second instance of the multiplexer circuitry 1203, 1204 observed in
If less than all the content of an execution lane's register files are shifted out during a shift sequence note that the content of the non shifted out registers of each execution lane remain in place (do not shift). As such, any non shifted content that is not replaced with shifted in content persists local to the execution lane across the shifting cycle. The memory unit (“M”) observed in each execution lane is used to load/store data from/to the random access memory space that is associated with the execution lane's row and/or column within the execution lane array. Here, the M unit acts as a standard M unit in that it is often used to load/store data that cannot be loaded/stored from/to the execution lane's own register space. In various embodiments, the primary operation of the M unit is to write data from a local register into memory, and, read data from memory and write it into a local register.
With respect to the ISA opcodes supported by the ALU unit of the hardware execution lane 1201, in various embodiments, the mathematical opcodes supported by the hardware ALU are integrally tied with (e.g., substantially the same as) the mathematical opcodes supported by a virtual execution lane (e.g., ADD, SUB, MOV, MUL, MAD, ABS, DIV, SHL, SHR, MIN/MAX, SEL, AND, OR, XOR, NOT). As described just above, memory access instructions can be executed by the execution lane 1201 to fetch/store data from/to their associated random access memory. Additionally the hardware execution lane 1201 supports shift op instructions (right, left, up, down) to shift data within the two dimensional shift register structure. As described above, program control instructions are largely executed by the scalar processor of the stencil processor.
d. Implementation Embodiments
The hardware design embodiments discussed above may be embodied within a semiconductor chip and/or as a description of a circuit design for eventual targeting toward a semiconductor manufacturing process. In the case of the later, such circuit descriptions may take of the form of higher/behavioral level circuit descriptions (e.g., a VHDL description) or lower level circuit description (e.g., a register transfer level (RTL) description, transistor level description or mask description) or various combinations thereof. Such circuit descriptions may be included, for instance, with one or more instances of the application software development environment described above in section 1.0 as target hardware for software written within the development environment. Circuit descriptions are typically embodied on a computer readable storage medium (such as a CD-ROM or other type of storage technology).
3.0 Concluding Statements
From the preceding sections is pertinent to recognize that the virtual environment described above in Section 1.0 may be instantiated on a computer system. Likewise an image processor as described above in Section 2.0 may be embodied in hardware on a computer system (e.g., as part of a handheld device's System on Chip (SOC) that processes data from the handheld device's camera).
It is pertinent to point out that the various image processor architecture features described above are not necessarily limited to image processing in the traditional sense and therefore may be applied to other applications that may (or may not) cause the image processor to be re-characterized. For example, if any of the various image processor architecture features described above were to be used in the creation and/or generation and/or rendering of animation as opposed to the processing of actual camera images, the image processor may be characterized as a graphics processing unit. Additionally, the image processor architectural features described above may be applied to other technical applications such as video processing, vision processing, image recognition and/or machine learning. Applied in this manner, the image processor may be integrated with (e.g., as a co-processor to) a more general purpose processor (e.g., that is or is part of a CPU of computing system), or, may be a stand alone processor within a computing system.
The hardware design embodiments discussed above may be embodied within a semiconductor chip and/or as a description of a circuit design for eventual targeting toward a semiconductor manufacturing process. In the case of the later, such circuit descriptions may take of the form of higher/behavioral level circuit descriptions (e.g., a VHDL description) or lower level circuit description (e.g., a register transfer level (RTL) description, transistor level description or mask description) or various combinations thereof. Circuit descriptions are typically embodied on a computer readable storage medium (such as a CD-ROM or other type of storage technology).
From the preceding sections is pertinent to recognize that an image processor as described above may be embodied in hardware on a computer system (e.g., as part of a handheld device's System on Chip (SOC) that processes data from the handheld device's camera). In cases where the image processor is embodied as a hardware circuit, note that the image data that is processed by the image processor may be received directly from a camera. Here, the image processor may be part of a discrete camera, or, part of a computing system having an integrated camera. In the case of the later the image data may be received directly from the camera or from the computing system's system memory (e.g., the camera sends its image data to system memory rather than the image processor). Note also that many of the features described in the preceding sections may be applicable to a graphics processor unit (which renders animation).
As observed in
An applications processor or multi-core processor 1350 may include one or more general purpose processing cores 1315 within its CPU 1301, one or more graphical processing units 1316, a memory management function 1317 (e.g., a memory controller), an I/O control function 1318 and an image processing unit 1319. The general purpose processing cores 1315 typically execute the operating system and application software of the computing system. The graphics processing units 1316 typically execute graphics intensive functions to, e.g., generate graphics information that is presented on the display 1303. The memory control function 1317 interfaces with the system memory 1302 to write/read data to/from system memory 1302. The power management control unit 1324 generally controls the power consumption of the system 1300.
The image processing unit 1319 may be implemented according to any of the image processing unit embodiments described at length above in the preceding sections. Alternatively or in combination, the IPU 1319 may be coupled to either or both of the GPU 1316 and CPU 1301 as a co-processor thereof. Additionally, in various embodiments, the GPU 1316 may be implemented with any of the image processor features described at length above.
Each of the touchscreen display 1303, the communication interfaces 1304-1307, the GPS interface 1308, the sensors 1309, the camera 1310, and the speaker/microphone codec 1313, 1314 all can be viewed as various forms of I/O (input and/or output) relative to the overall computing system including, where appropriate, an integrated peripheral device as well (e.g., the one or more cameras 1310). Depending on implementation, various ones of these I/O components may be integrated on the applications processor/multi-core processor 1350 or may be located off the die or outside the package of the applications processor/multi-core processor 1350.
In an embodiment one or more cameras 1310 includes a depth camera capable of measuring depth between the camera and an object in its field of view. Application software, operating system software, device driver software and/or firmware executing on a general purpose CPU core (or other functional block having an instruction execution pipeline to execute program code) of an applications processor or other processor may perform any of the functions described above.
Embodiments of the invention may include various processes as set forth above. The processes may be embodied in machine-executable instructions. The instructions can be used to cause a general-purpose or special-purpose processor to perform certain processes. Alternatively, these processes may be performed by specific hardware components that contain hardwired logic for performing the processes, or by any combination of programmed computer components and custom hardware components.
Elements of the present invention may also be provided as a machine-readable medium for storing the machine-executable instructions. The machine-readable medium may include, but is not limited to, floppy diskettes, optical disks, CD-ROMs, and magneto-optical disks, FLASH memory, ROMs, RAMs, EPROMs, EEPROMs, magnetic or optical cards, propagation media or other type of media/machine-readable medium suitable for storing electronic instructions. For example, the present invention may be downloaded as a computer program which may be transferred from a remote computer (e.g., a server) to a requesting computer (e.g., a client) by way of data signals embodied in a carrier wave or other propagation medium via a communication link (e.g., a modem or network connection).
In the foregoing specification, the invention has been described with reference to specific exemplary embodiments thereof. It will, however, be evident that various modifications and changes may be made thereto without departing from the broader spirit and scope of the invention as set forth in the appended claims. The specification and drawings are, accordingly, to be regarded in an illustrative rather than a restrictive sense.
This application is a continuation of and claims priority to U.S. patent application Ser. No. 14/694,890, filed on Apr. 23, 2015, the entire contents of which are hereby incorporated by reference.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4445177 | Bratt et al. | Apr 1984 | A |
| 4720781 | Chambers | Jan 1988 | A |
| 4720871 | Chambers | Jan 1988 | A |
| 4835712 | Derbin et al. | May 1989 | A |
| 4908644 | Shindo | Mar 1990 | A |
| 4935894 | Ternes et al. | Jun 1990 | A |
| 5253308 | Johnson | Oct 1993 | A |
| 5612693 | Craft et al. | Mar 1997 | A |
| 5751864 | Moriwake | May 1998 | A |
| 5820219 | Farkas | Sep 1998 | A |
| 5892962 | Cloutier | Apr 1999 | A |
| 6031573 | MacCormack et al. | Feb 2000 | A |
| 6049859 | Gliese et al. | Apr 2000 | A |
| 6366289 | Johns | Apr 2002 | B1 |
| 6587158 | Dale et al. | Jul 2003 | B1 |
| 6728862 | Wilson | Apr 2004 | B1 |
| 6745319 | Balmer et al. | Jun 2004 | B1 |
| 6970196 | Masatoshi et al. | Nov 2005 | B1 |
| 7010177 | Mattison | Mar 2006 | B1 |
| 7167890 | Lin et al. | Jan 2007 | B2 |
| 7200287 | Fukuda et al. | Apr 2007 | B2 |
| 7286717 | Nomizu | Oct 2007 | B2 |
| 7583851 | Kudo et al. | Sep 2009 | B2 |
| 7941634 | Georgi | May 2011 | B2 |
| 8023539 | Ueno | Sep 2011 | B2 |
| 8156284 | Vorbach et al. | Apr 2012 | B2 |
| 8321849 | Nickolls | Nov 2012 | B2 |
| 8436857 | Twilleager | Aug 2013 | B2 |
| 8508612 | Cote et al. | Aug 2013 | B2 |
| 8543843 | Cheng et al. | Sep 2013 | B1 |
| 8650384 | Lee | Feb 2014 | B2 |
| 8687223 | Yamamoto | Apr 2014 | B2 |
| 8749667 | Noraz et al. | Jun 2014 | B2 |
| 8786614 | Curry | Jul 2014 | B2 |
| 8797323 | Salvi et al. | Aug 2014 | B2 |
| 8823736 | Barringer et al. | Sep 2014 | B2 |
| 8970884 | Tsuji et al. | Mar 2015 | B2 |
| 8976195 | Lindhom et al. | Mar 2015 | B1 |
| 9256831 | van Lunteren | Feb 2016 | B2 |
| 20030005365 | Wilson | Jan 2003 | A1 |
| 20050270412 | Kamon | Dec 2005 | A1 |
| 20060044576 | Tabata | Mar 2006 | A1 |
| 20060228007 | Knee | Oct 2006 | A1 |
| 20070047828 | Ishii | Mar 2007 | A1 |
| 20070080969 | Yamaura | Apr 2007 | A1 |
| 20070156729 | Shaylor | Jul 2007 | A1 |
| 20080111823 | Fan | May 2008 | A1 |
| 20080244222 | Supalov | Oct 2008 | A1 |
| 20080266302 | Andre et al. | Oct 2008 | A1 |
| 20090002390 | Kuno | Jan 2009 | A1 |
| 20090228677 | Liege | Sep 2009 | A1 |
| 20090300621 | Mantor et al. | Dec 2009 | A1 |
| 20090317009 | Ren | Dec 2009 | A1 |
| 20100005243 | Yagi | Jan 2010 | A1 |
| 20100122105 | Arsian et al. | May 2010 | A1 |
| 20100182042 | Law et al. | Jul 2010 | A1 |
| 20100188538 | Sugawa | Jul 2010 | A1 |
| 20110055495 | Wolford | Mar 2011 | A1 |
| 20110087867 | Jacobson | Apr 2011 | A1 |
| 20110125768 | Shibao | May 2011 | A1 |
| 20110153925 | Bains et al. | Jun 2011 | A1 |
| 20120131309 | Johnson | May 2012 | A1 |
| 20120320070 | Arvo | Dec 2012 | A1 |
| 20130027416 | Vaithiananthan et al. | Jan 2013 | A1 |
| 20130202051 | Zhou | Aug 2013 | A1 |
| 20130241941 | Donaldson | Sep 2013 | A1 |
| 20130243329 | Oro Garcia et al. | Sep 2013 | A1 |
| 20130314428 | Chen | Nov 2013 | A1 |
| 20130318544 | Kuroda | Nov 2013 | A1 |
| 20140019486 | Majurndar | Jan 2014 | A1 |
| 20140028876 | Mills | Jan 2014 | A1 |
| 20140136816 | Krig | May 2014 | A1 |
| 20140270478 | Chen et al. | Sep 2014 | A1 |
| 20140282611 | Campbell | Sep 2014 | A1 |
| 20150106596 | Varback et al. | Apr 2015 | A1 |
| 20160162402 | Woolley, Jr. | Jun 2016 | A1 |
| 20160210720 | Taylor et al. | Jul 2016 | A1 |
| 20160219225 | Zhu | Jul 2016 | A1 |
| Number | Date | Country |
|---|---|---|
| 0293701 | Dec 1988 | EP |
| 2013090070 | May 2013 | JP |
| 9409595 | Apr 1994 | WO |
| 2007071883 | Jun 2007 | WO |
| Entry |
|---|
| International Preliminary Report on Patentability issued in International Application No. PCT/US2016/026029, dated Nov. 2, 2017, 9 pages. |
| “ChimeraTM: The NVIDIA Computational Photography Architecture” Whitepaper, NVIDIA Corporation 2013, 15 pages. |
| “Multioutput Scaler Reference Design” Altera Corporation, Application Note AN-648-1.0, Aug. 2012, 18 pages. |
| Adams, et al. “The Frankencamera: An Experimental Platform for Computational Photography”, ACM Transactions on Graphics, vol. 29, No. 4, Article 29, Publication Jul. 2010, 12 pages. |
| Adobe Systems,“Adobe Photoshop CS6,” released and on sale in May 2012, p. 1-2. |
| Ahn, et al., “Evaluating the Image Stream Architecture.” In ACM SIGARCH Computer Architecture News, vol. 32, No. 2, IEEE Computer Society, Jun. 2004, 14 pages. |
| Arasan “Silicon Hive Announces New Image Signal Processor”, Eindhoven, The Netherlands, May 16, 2006, http://www.design-reuse.com/news/13362/silicon-hive-image-ginal-processor.html, 3 pages. |
| Balfour, et al., “An Energy-Efficient Processor Architecture for Embedded Systems” IEEE Computer Architecture Letters 7, No. 1 p. 29-32, May 2008. |
| Barry, et al., “Always-On Vision Processing Unit for Mobile Applications”, IEEE Micro, Mar./Apr. 2015, pp. 56-66. |
| Bolotoff, Paul V., “Alpha—The History in Facts and Comments” http://alasir.com/articles/a;pha_history/alpha_21164_21164pc.html, Last modification date Apr. 22, 2007,5 pages. |
| Bushey, et al., “Flexible Function-Level Acceleration of Embedded Vision Applications Using the Pipelined Vision Processor.” In Signals, Systems and Computers, 2013 Asilomar Conference, pp. 1447-1452, IEEE, 2013. |
| Cardells-Tormo et al., “Area-efficient 2-D Shift-variant Convolvers for FPGA-based Digital Image Processing,” IEEE Workshop on Signal Processing Systems Design and Implementation, Nov. 2, 2005, 5 pages. |
| CEVA-MM3101: An Imaging—Optimized DSP Core Swings for an Embedded Vision Home Run, http://bdti.com/insideDSP/2012.01/24CEVA Jan. 19, 2012, 3 pages. |
| Chao et al., “Pyramid Architecture for 3840 x 2160 Quad Full High Definition 30 Frames/s Video Acquisition,” IEEE Transactions on Circuits and Systems for Video Technology, Nov. 1, 2010, 1 0 pages. |
| Chen et al., “CRISP: Coarse-Grained Reconfigurable Image Stream Processor for Digital Still Cameras and Camcorders,” IEEE Transactions on Circuits and Systems for Video Technology, Sep. 2008, 14 pages. |
| Chen, et al., “DaDianNao: A Machine-Learning Supercomputer.” 47m Annual IEEE/ACM International Symposium, pp. 609-622, IEEE, Dec. 2014. |
| Chenyun, et al., “A Paradigm Shift in Local Interconnect Technology Design in the Era of Nanoscale Multigate and Gate-All-Around Devices,” IEEE Electron Device Letters, vol. 36, No. 3, pp. 274-276, Mar. 2015. |
| Clearspeed Whitepaper: CSX Processor Architecture, VJ'Nw.clearspeecl.com, 2007, 16 pages. |
| Codrescu, et al., “Hexagon DSP: An Architecture Optimized for Mobile Multimedia and Communications.” IEEE Micro vol. 34, Issue 2, pp. 34-43, Mar. 2014. |
| Dally, William J., “Computer architecture is all about interconnect.” Proceedings of sm International Symposium High-Perf. Comp. Architecture, Feb. 2002, 11 pages. |
| De Dinechin, et al., “A Clustered Manycore Processor Architecture for Embedded and Accelerated Applications.” In High Performance Extreme Computing Conference (HPEC), IEEE, pp. 1-6, Sep. 2013. |
| DeVito, et al., “Terra: A Multi-Stage Language for High-Performance Computing”, PLD1'13, Jun. 16-22, 2013, Seattle, Washington, 11 pages. |
| DeVito, et al.,“First-class Runtime Generation of High-Performance Types using Exotypes”, PLD1'14, Jun. 9-11, ACM, 12 pages. |
| Dykes et al. “Communication and Computation Patters of Large Scale Image Convolutions on Parallel Architectures,” Parallel Processing Symposium, Proceedings, Eighth International Cancum, Mexico, Apr. 26-29, 1994, 6 pages. |
| Eichenberger, et al., “Optimizing Compiler for the Cell Processor” PACT, Tuesday, Sep. 20, 2005, Part 1, pp. 1-16. |
| Eichenberger, et al., “Optimizing Compiler for the Cell Processor” PACT, Tuesday, Sep. 20, 2005, Part 2, pp. 17-32. |
| EyeQ2TM, “Vision System on a Chip”, Mobileye, Sep. 2007, 4 pages. |
| Farabet, et al., “Hardware Accelerated Convolutional Neural Networks for Synthetic Vision Systems.” Proceedings of 2010 IEEE International Symposium, pp. 257-260, Jun. 2010. |
| Farabet, et al., “Neuflow: A Runtime Reconfigurable Dataflow Processor for Vision.” IEEE Computer Society Conference, pp. 109-116, Jun. 2011. |
| Galal, et al., “FPU Generator for Design Space Exploration.” 21 si IEEE Symposium on Computer Arithmetic (ARITH), Apr. 2013, 10 pages. |
| Gentile, et al., “Image Processing Chain for Digital Still Cameras Based on SIMPil Architecture.” ICCP International Conference Workshops, pp. 215-222, IEEE, Jun. 2005. |
| Goldstein, et al., “PipeRench: A Coprocessor for Streaming Multimedia Acceleration”, Carnegie Mellow University, Research Showcase @CMU, Appears in the 26th Annual International Symposium on Computer Architecture, May 1999, Atlanta, Georgia, 14 pages. |
| Gupta, et al., “A VLSI Architecture for Updating Raster-Scan Displays”, Computer Graphics, vol. 15, No. 3, Aug. 1981, pp. 71-78. |
| Hameed, et al., “Understanding Sources of Inefficiency in General-Purpose Chips.” ACM SIGARCH Computer Architecture News, vol. 38, No. 3, pp. 37-47, 2010. |
| Hanrahan, Pat, “Domain-Specific Languages for Heterogeneous GPU Computing”, NVIDIA Technology Conference, Oct. 2, 2009, 30 pages. |
| Hegarty, et al., “Darkroom: Compiling High-Level Image Processing Code into Hardware Pipelines”, Proceedings of ACM SIGGRAPH, Jul. 2014, 11 pages. |
| Henretty, et al., “A Stencil Compiler for Short-Vector SIMD Architectures”, ACM, ICS'13, Jun. 10-14, 2013, Eugene, Oregon, pp. 13-24. |
| Horowitz, Mark, “Computing's Energy Problem: (and what we can do about it)”, IEEE, International Solid-State Circuits Conference 2014, 46 pages. |
| Kapasi, et al. “The Imagine Stream Processor”, IEEE, International Conference on Computer Design: VLSI in Computers and Processors (ICCD'02), 2002, 17 pages. |
| Khailany, et al., “A Programmable 512 GOPS Stream Processor for Signal, Image, and Video Processing”, IEEE Journal of Solid-State Circuits, vol. 43, No. 1, pp. 202-213, Jan. 2008. |
| Khawam, et al., “The Reconfigurable Instruction Cell Array”, IEEE Transactions on Very Large Scale Integration (VLSI) Systems, vol. 16, No. 1, Jan. 2008, pp. 75-85. |
| Khronos, SPIR 2.0 “Provisional Specification to Enable Compiler Innovation on Diverse Parallel Architectures”, Aug. 11, 2014, 2 pages. |
| Levinthal, “Parallel Computers for Graphics Applications”, ACM, 1987, pp. 193-198. |
| Levinthal, et al., “Chap-A SIMD Graphics Processor”, Computer Graphics, vol. 18, No. 3, Jul. 1984, pp. 77-82. |
| M.J. Duff, “Clip 4: A Large Scale Integrated Circuit Array Parallel Processor,” Proc. IEEE lnt'l Joint Conf. Pattern Recognition, . . . Jan. 2004, pp. 728-733. |
| Mandal, et al., “An Embedded Vision Engine (EVE) for Automotive Vision Processing.” In Circuits and Systems (ISCAS), 2014 IEEE Symposium, pp. 49-52, IEEE, 2014. |
| Mcintosh-Smith, “A Next-Generation Many-Core Processor With Reliability, Fault Tolerance and Adaptive Power Management Features Optimized for Embedded and High Performance Computing Applications” in Proceedings of the High Performance Embedded Computinq Conference (HPEC) 2008, pp. 1-2, 5. |
| Mody et al. “High Performance and Flexible Imaging Sub-System,” Advanes in Computing, Communication and Informatics (ICACCI), 2014 International Conference on, pp. 545-548, IEEE, 2014. |
| Molnar, et al., “PixelFlow: High-Speed Rendering Using Image Composition” proceedings of Siggraph 92, Computer Graphics, 26, Jul. 2, 1992, 10 pages. |
| Moloney, David, “1 TOPS/W Software Programmable Media Processor.” Hot Chips 23 Symposium (HCS), IEEE, Aug. 2011, 24 pages. |
| Moloney, et al., “Myriad 2: Eye of the Computational Vision Storm”, Hot Chips 26 Symposium (HCS), IEEE, Aug. 2014, 18 pages. |
| MPPA—Manycore, Product Family Overview http://www.kalrav.eu/img/pdf/flyer_mppa_manycore-4.pdf_February_2014, 2 pages. |
| Nightingale, Edmund B., et al., “Helios: Heterogeneous Multiprocessing with Satellite Kernels,” SOSP '09, Oct. 11-14, 2009, Big Sky, Montana, U.S.A., (Oct. 11, 2009), 14 pages. |
| NVIDIA—NVIDIA's Next Generation CUDA™ Compute Architecture: Kepler™,GK110/210, 2014, 23 pages,http://international.download.nvidia.com/pdf/kepler/NVIDIA-Kepler-GK110-GK210-Architecture-Whitepaper.pdf. |
| NVIDIA Tegra K1—A New Era in Mobile Computing—Whitepapers—Jan. 2014, 26 pages. |
| NVIDIA Tegra X1—NVIDIA's New Mobile Superchip—Whitepapers—Jan. 2015, 41 pages. |
| NVIDIA, “PTX:Parallel Thread Execution ISA Version 1.4”, Aug. 23, 2010, 135 pages. |
| Olofsson, et al., “A 25 GFLOPS/Watt Software Programmable Floating Point Accelerator” High Performance Embedded Computing Conference, 2010, 2 pages. |
| Oosterhout, Optimized Pixel Template Image Correlator, Master Thesis, Aug. 19, 1992, 74 pages. |
| Parker, Richard, “Embedding Computer Vision in Electronic Devices: How New Hardware and Software Choices Have Changed the Designer's Approach”, Future Technology Magazine, pp. 22-23, Feb. 2013. |
| PCT/US2016/026029, International Search Report and Written Opinion, dated Jul. 5, 2016, 12 pages. |
| Pelc, Oscar, “Multimedia Support in the i.MX31 and i.MC31L Applications Processors,” Freescale Semiconductor, Inc. Feb. 2006, http://cache.freescale.com/files/32bit/doc/white_paper/IMX31MULTIWP.pdf 12 pages. |
| Pham, et al., “NeuFlow: Dataflow Vision Processing System-On-A-Chip.” IEEE 55m International Midwest Symposium, Aug. 2012, 4 pages. |
| Qadeer, et al., “Convolution Engine: Balancing Efficiency & Flexibility in Specialized Computing”, ISCA '13 Tel-Aviv, Israel, ACM 2013, pp. 24-35. |
| Ragan-Kelley et al., “Halide: A Language and Compiler for Optimizing Parallelism, Locality, and Recomputation in Image Processing Pipelines,” Jun. 16, 2013, 12 pages. |
| S.F. Reddaway,“DAP—A Distributed Processor Array”, ISCA '73 Proceedings of the First Ann. Symposium on Computer Architecture, pp. 61-65. |
| SCP2200: Image Cognition Processors Family [Product Brief]*, https://www.element14.com/community/docs/DOC-50990 Oct. 26, 2012 2 pages. |
| SCP2200: Image Cognition Processors*, http://www.element14.com.community/docs/DOC-50978/1/scp2200-image-cognition-processors Oct. 25, 2012, 2 pages. |
| Sedaghati, et al., “SIVEC: A Vector Instruction Extension for High Performance Stencil Computation,” Parallel Architectures and Compilation Techniques, Oct. 10, 2011, 12 pages |
| SemiWiki.com—New CEVA-ZM4 Vision IP Does Point clouds and and More: Published Feb. 27, 2015, http://www.semiwiki.com/forum/content/4354-new-ceva-xm4-vision-ip-does-point-clouds-more.html. |
| Shacham, et al. “Smart Memories Polymorphic Chip Multiprocessor”, Proceedings of the 46th Design Automation Conference (OAC), San Francisco, CA, Jul. 2009. |
| Shacham, et al., “Rethinking Digital Design: Why Design Must Change”, IEEE micro Magazine, Nov./Dec. 2010. |
| Shacham, Ofer, “Chip Multiprocessor Generator: Automatic Generation of Custom and Heterogeneous Complete Platforms” Dissertation—Stanford University, May 2011, 190 paqes. Part 1. |
| Shacham, Ofer, “Chip Multiprocessor Generator: Automatic Generation of Custom and Heterogeneous Complete Platforms” Dissertation—Stanford University, May 2011, 190 pages. Part 2. |
| Silicon Hive: “Silicon System Design of Media and Communications Applications”, Jun. 13, 2007, 22 pages. |
| Spampinato, et al., “A Basic Linear Algebra Compiler”, ACM, CG0'14, Feb. 15-19, 2014, Orlando Fl, pp. 23-32. |
| SPIR, The SPIR Specification, Version 2.0, Revision Date Jun. 5, 2014, 40 pages. |
| Stein et al. “A Computer Vision System on a Chip: A Case Study From the Automotive Domain,” IEEE Computer Society Conference, p. 130, 2005. |
| Stojanov, et al., “Abstracting Vector Architectures in Library Generators: Case Study Convolution Filters”, ARRAY 14, ACM, Jun. 11, 2014, UK, pp. 14-19. |
| Stream Processors, Inc. Announces Storm-1 Family of Data-Parallel Digital Signal Processors, ISSCC 2007, Feb. 12, 2007, 3 pages. http://www.businesswire.com/news/home/20070212005230/en/Stream-Processors-Announces-Storm-1-Family-Data-Parallel-Digital. |
| Stream Processors, Inc., Company History—Foundational Work in Stream Processing initiated in 1995, https://en.wikipedia.org/wiki/stream_processors,_inc, 5 pages. |
| Tanabe, et al., “Visconti: multi-VLIW image Recognition Processor Based on Configurable Processor [obstacle detection applications]”, Custom Integrated Circuits Conference, IEEE, 2003, http://ieeexplore.ieee.org/document/1249387/?anumber=1249387&tag=1. |
| Van der Wal, et al. “The Acadia Vision Processor,” IEEE proceedings of the International Workshop on OCmputer Application for Machine Perception, Sep. 2000, http://citeseerx.ist.pdr.edu/viewdoc/download?doi=10.1.1.32.3830&rep=rep1&type=pdf 10 pages. |
| Wahib et al., “Scalable Kernel Fusion for Memory-Bound GPU Applications,” SC14: International Conference for High Performance Computing, Networking, Storage and Analysis, Nov. 16, 2014, 12 pages. |
| Yu et al., “Optimizing Data Intensive Window-based Image Processing on reconfigurable Hardware Boards,” Proc. IEEE Workshop on Signal Processing System Design and Implementation, Nov. 2, 2005, 6 pages. |
| Zhou, Minhua, et al., “Parallel Tools in HEVC for High-Throughput Processing, Applications of Digital Processing,” XXXV, Proc. of SPI, vol. 8499, (Jan. 1, 2012), pp. 1-13. |
| Number | Date | Country | |
|---|---|---|---|
| 20170242943 A1 | Aug 2017 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 14694890 | Apr 2015 | US |
| Child | 15591984 | US |
