An approach to clock frequency modulation of a clock source with a fixed frequency
The modulated clock system addresses the inefficiency of multiple PLLs by using a single clock source with delay elements and phase multiplexers to adjust clock phases, effectively generating varied frequencies for different components without additional hardware, thus optimizing space and resource usage.
Patent Information
- Application Number
- DE102013020610
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2013-02-06
- Filing Date
- 2013-12-15
- Publication Date
- 2024-11-21
- Estimated Expiration
- 2033-12-15
AI Technical Summary
Existing integrated circuits face inefficiencies in providing clock signals to components requiring different frequencies, as maintaining separate phase locked loops (PLLs) for each device consumes significant space and resources.
A modulated clock system that uses a single clock source with a calibrator, output clock, and updater to generate clock signals at varying frequencies without additional PLLs, utilizing delay elements and phase multiplexers to adjust clock phases for different components.
This approach reduces the need for multiple PLLs, optimizing space usage and efficiently generating clock signals at different frequencies for various components within a computer system.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
BACKGROUND OF THE INVENTIONField of the invention
[0001] The present invention relates generally to integrated circuits and, more particularly, to an approach for clock frequency modulation of a fixed frequency clock source. Description of the state of the art
[0002] In US 2011 / 0221497 A1, a delay associated with each of two signals along a respective transmission path is precisely measured using a delay measurement circuit fabricated in-situ on a device on which circuitry for transmitting the two signals is fabricated. Thus, the measured delay associated with each of the two signals is subject to the same manufacturing-dependent attributes that reflect the actual circuitry through which the two signals are transmitted during operation of the device. The offset between the two signals is quantified as the difference in the measured delays. Coarse and fine delay modules are defined within the transmission path of each of the two signals. Based on the measured offset between the two signals, the coarse and fine delay modules are adjusted accordingly to compensate for the offset.The corresponding settings for the coarse and fine delay modules can be stored in non-volatile memory elements.
[0003] A computer clock signal is a special type of signal that oscillates between a high state and a low state and is used like a metronome to coordinate the actions of various circuits within the computer. A clock generator in the computer generates the clock signal. Although more complex arrangements can exist, a typical clock signal takes the form of a square wave with a 50% duty cycle (i.e., the proportion of time a device spends in an active state as a proportion of the total time under consideration), usually at a fixed, constant frequency. Circuits that use the clock signal for synchronization become active on the rising edge, the falling edge, or, in the case of double the data rate, both the rising and falling edges of the clock signal.
[0004] During operation, components within the computer, such as processing units, may require slightly different clock frequencies than other components within the computer that also require clock signals. One approach to meeting the requirement for different clock signals is to provide a dedicated clock signal for each device (e.g., each processing unit or other device) that requires a clock signal. For example, an integrated circuit may be provided with different phase-locked loops (PLLs) for each device that requires a different clock frequency, even if the required frequencies of the devices are only slightly different.For example, if the required clock frequency for one processing unit is 100.0 MHz, while the required clock frequency for the other processing unit is slightly different at 100.5 MHz, these two processing units typically require separate PLLs. This allows each device to operate at the required frequency for that device. Unfortunately, maintaining dedicated clocks (e.g., PLLs) tends to consume a relatively large amount of space on a semiconductor circuit. Therefore, this approach is typically undesirable.
[0005] As the foregoing shows, what is needed in this area of technology is an optimized approach to providing clock signals to different components within a computer. OVERVIEW OF THE INVENTION
[0006] An embodiment of the present technique includes an output clock according to claim 1.
[0007] Advantageously, the output clock of the present technique can generate an output clock signal that is slightly faster or slightly slower than an existing source clock signal, and this is accomplished without the need for an additional phase-locked loop (PLL) for each device requiring a different clock frequency. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] In order to indicate how the above-mentioned features of the present invention may be understood in detail, a more specific description of the invention, previously briefly summarized, will be given with reference to embodiments, some of which are illustrated in the accompanying drawings. It should be noted, however, that the accompanying drawings represent only typical implementations of this invention and are therefore not to be considered limiting of its scope, for the invention is capable of other equally effective implementations. Fig. 1 is a block diagram illustrating a computer system configured to implement one or more aspects of the present invention. Fig. 2 is a block diagram showing a parallel processing subsystem according to an embodiment of the present invention. Fig. Figure 3 is a block diagram of a computer system with a conventional system. Fig. 4 is a block diagram of a computer system having a modulated clock system 401 according to an embodiment of the present invention. Fig. Figure 5 is a block diagram of a portion of the modulated clock device of Fig. 4 according to an embodiment of the present invention. Fig. 6 is a view showing the illustrative mathematics for clock frequency modulation performed by an output clock according to an embodiment of the present invention. Fig. Figure 7 is another block diagram of a portion of the modulated clock device of Fig. 4 according to an embodiment of the present invention. Fig. Figure 8 is a more detailed block diagram of the modulated clock device according to Fig. 4 according to an embodiment of the present invention. Fig. 9 is a flowchart of method steps for generating an output clock signal according to an embodiment of the present invention. Fig. 10 is a flowchart of method steps for updating a phase of an output clock signal according to an embodiment of the present invention. Fig. 11 is a flowchart of method steps for automatically calibrating a modulated clock device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0009] In the following description, numerous specific details are set forth in order to provide a more thorough understanding of the present invention. However, those skilled in the art will recognize that the present invention may be practiced without one or more of these specific details. In other instances, well-known features are not described to avoid obscuring the present invention. System overview
[0010] Fig. 1 is a block diagram illustrating a computer system 100 configured to implement one or more aspects of the present invention. Computer system 100 includes a central processing unit (CPU) 102 and a system memory 104 including a device driver 103. Computer system 100 includes a central processing unit (CPU) 102 and a system memory 104 including a device driver 103. CPU 102 and system memory 104 communicate via an interconnect path, which may include a memory bridge 105. Memory bridge 105, which may be, for example, a northbridge chip, is connected to an input / output (I / O) bridge 107 via a bus or other communications path 106 (for example, a HyperTransport interconnect).The I / O bridge 107, which may be a southbridge chip, for example, receives user input from one or more user input devices 108 (e.g., keyboard, mouse) and forwards the input to the CPU 102 via path 106 and the memory bridge 105.
[0011] As also shown, a parallel processing subsystem 112 is connected to the memory bridge 105 via a bus or other communication path 113 (e.g., a Peripheral Component Interconnect (PCI) Express, an Accelerated Graphics Port (AGP), or a HyperTransport connection); in one embodiment, the parallel processing subsystem 112 is a graphics subsystem that provides pixels to a display device 110 (e.g., a conventional cathode ray tube (CRT)-based monitor or a liquid crystal display). A system disk 114 is also connected to the I / O bridge 107. A switch 116 provides connections between the I / O bridge 107 and other components, such as a network adapter 118 and various add-in cards 120 and 121.Other components (not explicitly shown), including a universal serial bus (USB) or other port connections, compact disk (CD) drives, digital video disk (DVD) drives, movie recorders, and the like, may also be connected to the I / O bridge 107. Communication paths connecting the various components in . Fig. 1 interconnection may be implemented using any suitable protocol, such as PCI, PCI Express (PCIe), AGP, HyperTransport, or one or more other bus or point-to-point communication protocols, and connections between different devices may use different protocols, as is known in the art. A component is typically a part of a device or device and is hardware or a combination of hardware and software.
[0012] As explained in more detail below with reference to Fig. 2-11, the parallel processing subsystem includes parallel processing units (PPUs) configured to execute a software application, such as device driver 103, using circuitry enabling camera control. These packet types are specified by the communication protocol used by communication path 113. In situations where a new packet type is introduced in the communication protocol (e.g., due to an extension of the communication protocol), parallel processing subsystem 112 may be configured to generate packets based on the new packet type and exchange data with CPU 102 (or other processing units) via communication path 113 using the new packet types.
[0013] In one implementation, parallel processing subsystem 112 includes circuitry optimized for graphics and video processing, including, for example, video output circuitry, and forms a graphics processing unit (GPU). In another embodiment, parallel processing subsystem 112 includes circuitry optimized for general-purpose processing while maintaining the underlying computational architecture described in more detail below. In yet another embodiment, parallel processing subsystem 112 may be integrated with one or more other system elements, such as memory bridge 105, CPU 102, and I / O bridge 107, to form a system-on-a-chip (SoC).
[0014] It should be noted that the system shown herein is illustrative in nature and that changes and modifications are possible. The interconnect topology, including the number and arrangement of bridges, the number of CPUs 102, and the number of parallel processing subsystems 112, may be modified as needed. For example, in some embodiments, system memory 104 is connected to CPU 102 directly rather than via a bridge, and other devices communicate with system memory 104 via memory bridge 105 and CPU 102. In other alternative topologies, parallel processing subsystem 112 is connected to I / O bridge 107 or directly to CPU 102 rather than memory bridge 105. In still other embodiments, I / O bridge 107 and memory bridge 105 may be integrated into a single chip. Large embodiments may include two or more CPUs 102 and two or more parallel processing subsystems 112.The specific components shown herein are optional; for example, any number of add-in cards or peripheral devices may be supported. In some embodiments, switch 116 is omitted, and network adapter 118 and add-in cards 120, 121 are connected directly to I / O bridge 107.
[0015] Fig. 2 is a block diagram illustrating a parallel processing subsystem 112 according to an embodiment of the present invention. As shown, the parallel processing subsystem 112 includes one or more parallel processing units (PPUs) 202, each of which is connected to a local parallel processing (PP) memory 204. Generally, a parallel processing subsystem includes a number U of PPUs, where U ≥ 1. (Herein, multiple instances of the same objects are denoted by reference numerals indicating the object, and numbers in parentheses identify the instance when necessary.) The PPUs 202 and the parallel processing memories 204 may be implemented using one or more integrated circuit devices, such as programmable processors, application-specific integrated circuits (ASICs), or memory devices, or in any other technically feasible manner.
[0016] It is back on Fig. 1; in some embodiments, some or all of the PPUs 202 in the parallel processing subsystem 112 are graphics processors with image generation pipelines that may be configured to perform various tasks related to generating pixel data from graphics data provided by the CPU 102 and / or system memory 104 via the memory bridge 105 and bus 113, to interact with the local parallel processing memory 204 (which may be used as graphics memory and may include, for example, a conventional block buffer or frame buffer), to store and update pixel data, to provide pixel data to the display device 110, and the like. In some embodiments, the parallel processing subsystem 112 may include one or more PPUs 202 operating as graphics processors and may include one or more other PPUs 202 used for general-purpose computations.The PPUs may be identical or different, and each PPU may have its own dedicated parallel processing memory or no dedicated parallel processing memory. One or more PPUs 202 may output data to the display device 110, or each PPU 202 may output data to one or more display devices 110.
[0017] During operation, the CPU 102 is the overarching processor of the camera system 100 and controls and coordinates the operation of other system components. In particular, the CPU 102 issues instructions that control the operation of the PPUs 202. In some embodiments, the CPU 102 writes a stream of instructions for each PPU 202 into a shift buffer (in Fig. 1 or Fig. 2), which may be located in system memory 104, parallel processing memory 204, or other memory location accessible by both CPU 102 and PPU 202. PPU 202 reads the instruction stream from the shift buffer and then executes instructions asynchronously relative to the operation of CPU 102.
[0018] Let us now return to Fig. 2; each PPU 202 includes an I / O unit 205 that communicates with the rest of the camera system 100 via communication path 113, which is connected to the memory bridge 105 (or, in an alternative embodiment, directly to the CPU 102). The connection of the PPU 202 to the rest of the camera system 100 can also be different. In some embodiments, the parallel processing subsystem 112 is implemented as an add-on card that can be inserted into an expansion slot of the camera system 100. In other embodiments, a PPU 202 can be integrated into a single chip with a bus bridge, such as the memory bridge 105 or the I / O bridge 107. In still other embodiments, some or all elements of the PPU 202 can be integrated into a single chip with the CPU 102.
[0019] In one embodiment, communication path 113 is a PCIe connection in which dedicated lanes are assigned to each PPU 202, as is known in the art. Other communication paths may also be used. As previously stated, the reverse-flow interconnect may also be used to implement communication path 113 as well as any other communication path in computer system 100, CPU 102, or PPU 202. An I / O unit 205 generates packets (or other signals) for transmission over communication path 113 and also receives any incoming packets (or other signals) from communication path 113, and routes the incoming packets to appropriate components of PPU 202.For example, instructions related to processing tasks may be routed to a main interface 206, while instructions related to memory operations (e.g., reading from or writing to parallel processing memory 204) may be routed to a memory intersection unit 210. The main interface 206 reads each shift buffer and outputs the work specified by the shift buffer to a front-end area 212.
[0020] Each PPU 202 advantageously implements a highly parallel processing architecture. As shown in detail, the PPU 202(0) includes a processing cluster array 230 containing a number C of general purpose processing clusters (GPCs) 208, where C ≥ 1. Each GPC 208 is capable of executing a large number (e.g., hundreds or thousands) of threads concurrently, where each thread is an instance of a program. In various applications, different GPCs 208 may be reserved for different types of programs or for executing different types of computations. The allocation of GPCs 208 may vary depending on the workload encountered for each type of program or computation.
[0021] The GPCs 208 receive processing tasks to be executed via a work distribution unit 200, which receives commands defining processing tasks from a front-end unit 212. The front-end unit 212 ensures that the GPCs 208 are configured to an appropriate state before the processing specified by the shift buffers begins.
[0022] For example, if the PPU 202 is used for graphics processing, the processing workload for each patch is divided into approximately equal-sized tasks to enable distribution of the tiling and shading processing among multiple GPCs 208. A work distribution unit 200 is optionally configured to generate tasks at a frequency that allows tasks to be issued to multiple GPCs 208 for processing. In contrast, in conventional systems, processing is typically performed by a single processing unit while the other processing units remain idle, waiting for the single processing unit to complete its tasks before beginning their processing tasks. In some embodiments of the present invention, portions of the GPCs 208 are configured to perform different types of processing.For example, a first part may be configured to perform vertex shading and topology generation. A second part may be configured to perform tessellation and geometry shading. A third part may be configured to perform pixel shading in screen space to create a generated image. Intermediate data generated by the GPCs 208 may be stored in buffers so that the intermediate data can be exchanged among the GPCs 208 for further processing.
[0023] The memory interface 214 includes a number D of partition units 215, each directly connected to a portion of the parallel processing memory 204, where D ≥ 1. As shown, the number of partition units 215 is equal to the number of DRAMs 220. In other implementations, the number of partition units 215 is not equal to the number of memory devices. The dynamic random access memories (DRAMs) 220 may be replaced by other suitable memory devices and may be of generally conventional design. Render targets, such as block buffers or image buffers or texture maps, may be stored in the DRAMs 220, enabling the partition units 215 to write portions of each render target in parallel to efficiently utilize the available bandwidth of the parallel processing memory 204.
[0024] Each of the GPCs 208 can process data to be written to one of the DRAMs 220 within the parallel processing memory 204. The intersection unit 210 is configured to route the output of each GPC 208 to the input of a partition unit 215 or another GPC 208 for further processing. The GPCs 208 communicate with the memory interface 214 via the intersection unit 210 to read from or write to various external memory devices. In one embodiment, the intersection unit 210 has a connection to the memory interface 214 to communicate with the I / O unit 205 and also has a connection to the local parallel processing memory 204, thereby enabling the processing cores in the different GPCs 208 to communicate with the system memory 104 or other memory not located locally in the PPU 202. In the Fig. In the embodiment shown in Figure 2, the intersection unit 210 is directly connected to the I / O unit 205. The intersection unit 210 may use virtual channels to separate traffic flows between the GPCs 208 and the partition units 215.
[0025] The GPCs 208, in turn, may be programmed to perform processing tasks relating to a wide variety of applications, including, but not limited to, linear and non-linear data transformations, filtering of video and / or audio data, modeling operations (e.g., applying physical laws to determine the position, velocity, and other attributes of objects), image generation operations (e.g., tessellation shading programs, vertex shading, geometry shading, and / or pixel shading programs), etc.The PPUs 202 may transfer data from the system memory 104 and / or the local parallel processing memories 204 to an internal (on-chip) memory, process the data, and write the resulting data back to the system memory 104 and / or the local parallel processing memories 204, where such data may be accessed by other system components, including the CPU 102 or another parallel processing subsystem 112.
[0026] A PPU 202 may be equipped with any amount of local parallel processing memory 204, including no local memory, and may use local memory and system memory in any combination. For example, in one embodiment of a unified memory architecture (UMA), a PPU 202 may be a graphics processor. In such embodiments, little or no dedicated graphics (parallel processing) memory is provided, and the PPU 202 would use system memory exclusively or almost exclusively. In UMA embodiments, a PPU 202 may be integrated into a bridge chip or a processor chip, or may be provided as a discrete chip with a high-speed interconnect (e.g., PCIe) connecting the PPU 202 to system memory via a bridge chip or other communication device.
[0027] As previously stated, any number of PPUs 202 may be included in a parallel processing subsystem 112. For example, multiple PPUs 202 may be provided on a single add-in card, or multiple add-in cards may be connected to communication path 113, or one or more PPUs 202 may be integrated into a bridge chip. The PPUs 202 in a multi-PPU system may be identical or different from one another. For example, different PPUs 202 may have a different number of processing cores, a different amount of local parallel processing memory, etc. When multiple PPUs 202 are present, these PPUs may operate in parallel to process data at a higher throughput than would be possible with a single PPU 202.Systems having one or more PPUs 202 may be implemented in a variety of configurations and form factors, including desktop computers, mobile or handheld personal computers, service provider computers, workstation computers, game consoles, embedded systems, and the like.
[0028] An embodiment of the invention may be implemented as a program product for use in a computer system, such as the computer system 100 of Fig. 1, for example. One or more programs of the program product define functions of the embodiments (including the methods described herein) and may be embodied on a variety of computer-readable storage media. Illustrative computer-readable storage media include, without limitation: (i) non-writable storage media (e.g., read-only memory devices in a computer, such as CD-ROM disks readable by a CD-ROM drive, flash memory, ROM chips, or other types of non-volatile semiconductor memory) on which information is permanently stored; and (ii) writable storage media (e.g., floppy disks in a floppy disk drive, or a hard disk drive or other type of random access semiconductor memory) on which alterable information is stored. Overview of a clock system
[0029] Fig. 3 is a block diagram of a conventional computer system 300 having a clock system 301. In computers, synchronization is essential to the function of the computer. The clock system 301 synchronizes tasks performed by the conventional computer system 300. Such tasks include, for example, loading data prior to processing the data, among a variety of other computer tasks. All components or devices of the computer do not necessarily operate with a single clock signal. Different devices in the conventional computer system 300 require operation at different clock frequencies for the different devices. For purposes of explanation, components of the conventional computer system 300 that require slightly different clock frequencies are the PPUs 202, including PPU 202(0), PPU 202(1),..., and PPU 202(U-1), where U ≥ 1.
[0030] The clock system 301 includes clocks 302, including clocks 302(0), clocks 302(1), ..., and clocks 302(U-1), where U ≥ 1. Each clock 302 is connected to a specific device, since in this example, each specific device operates at a different clock frequency. For example, clock 302(0) is connected to PPU 202(0), etc. Each clock 302 is configured to provide an output clock signal 303 to a specific device. For example, clock 302(0) is configured to provide an output clock signal 303(0) to PPU 202(0), etc. Each output clock signal (e.g., output clock signal 302(0)) is a continuous series of precise high and low pulses that have substantially the same length. A single clock cycle is the time that elapses from the beginning of one high pulse to the beginning of the next high pulse.
[0031] The circuitry in each clock generator 302, which respectively provides the output signal 302 to each associated device 202, may comprise electronic circuitry including a phase-locked loop (PLL). Typically, a reference clock signal (not shown) drives the phase-locked loop (PLL), which then drives a clock distribution network (not shown). The clock distribution network is usually balanced so that the clock signal arrives at each endpoint simultaneously. One of these endpoints is the feedback input of the PLL. The function of the PLL is to compare the distributed clock signal with the incoming reference clock signal and vary the phase and frequency of the output until the reference clock signal and the feedback signal match in phase and frequency.PLLs are widely used and can adjust clock signals in systems several feet apart, as well as clock signals located in small areas of individual chips.
[0032] However, maintaining a PLL for each clock 302 consumes a relatively large amount of space on an integrated circuit. Therefore, the approach outlined below reduces the number of PLLs and / or other devices required to generate clock signals at different frequencies. Overview of clock frequency modulation
[0033] Fig. Figure 4 is a block diagram of a computer system 400 with a modulated clock system 401 according to an embodiment of the present invention. For illustrative purposes, the modulated clock system 401 includes a clock source 410 (e.g., a single PLL) connected to modulated clock devices 402. Instead of the clocks 302 (e.g., PLL) of Fig. 2 comprises the modulated clock system 401 from Fig. 4, the modulated clock devices 402 and a single clock source 410. Therefore, the modulated clock system 401 significantly reduces the number of PLLs and / or other devices required to generate clock signals for the various devices of the computer system 400.
[0034] Also for purposes of explanation, the components of computer system 400 that require slightly different clock frequencies are the PPUs 202, including PPU 202(0), PPU 202(1), ..., and PPU 202(U-1), where U ≥ 1. However, the present approach is not limited thereto. For example, a single PPU (e.g., PPU 202(0)) may include multiple devices, each requiring operation at different clock frequencies. Alternatively, devices requiring different frequencies may not be part of a PPU of a GPU, but may be part of a CPU or, for example, another device.
[0035] The modulated clock devices 402 include a modulated clock device 402(0), a modulated clock device 402(1), ..., and a modulated clock device 402(U-1), where U ≥ 1. Each modulated clock device 402 is connected to a specific device, since in this example, each specific device operates at a different clock frequency. For example, the modulated clock device 402(0) is connected to the PPU 202(0), etc. Each modulated clock device 402 is configured to send an output clock signal 403 to a specific device. Each modulated clock device 402 may supply a different clock signal to a different device (e.g., the PPU 202(0), the PPU 202(1), etc.). For example, the clock generator 302(0) is configured to send an output clock signal 302(0) to the PPU 202(0), etc. Similarly, the modulated clock device 402(1) is configured to send another output clock signal 403(1) to the PPU 202(1), etc.
[0036] Each modulated clock device 402 includes a calibration device 404, an output clock 406, and an update device 408. For example, the modulated clock device 402(0) includes a calibration device 404(0), an output clock 406(0), and an update device 408(0). The modulated clock device 402(1) includes a calibration device 404(1), an output clock 406(1), and an update device 408(1). The modulated clock device 402(U-1) includes a calibration device 404(U-1), an output clock 406(U-1), and an update device 408(U-1), where U ≥ 1.
[0037] In some cases, the calibration device 404(0) may be challenging in terms of the physical area that the calibration device 404(0) occupies on an integrated circuit and also in terms of the processing required to perform calibration operations. Consequently, in an alternative embodiment (not shown), the modulated clock system 401 is configured to include a modulated clock device 402(0) comprising (which will be described below with reference to Fig. 5-8): a calibration device 404, a sequence of delay elements, a plurality of output phase multiplexers 516, and a plurality of update devices 408.
[0038] The output clock 406(0) is described below with reference to Fig. 5 and Fig. 6. The updating device 408(0) is described below with reference to Fig. 7. The calibration device 404(0) is described in more detail below with reference to Fig. 8. The other devices with corresponding reference numerals are constructed in a similar manner. Output clock
[0039] Fig. 5 is a block diagram of a portion of the modulated clock device 402(0) of Fig. 4 according to an embodiment of the present invention. Output clock 406(0) is connected to clock source 410. Output clock 406(0) provides a digital way to modulate an input clock signal 411 of fixed-frequency clock source 410 (e.g., PLL) by inserting a sequence of delay devices (e.g., digital delay elements) behind clock source 410, with tap points after each delay stage. A tap point is a connection point (e.g., node) between a delay device and an output phase multiplexer (mux) 516(0).A delay device is a digital device configured to receive a clock signal (e.g., input clock signal 411 from clock source 410 or a delayed clock signal from a preceding digital delay stage) and provide a delayed clock signal that is delayed in phase relative to the received clock signal. This concept of delay is further explained below with reference to FIG. Fig. 6. In this example the Fig. 5, the output clock 406(0) includes 10 delay stages and 10 corresponding tap points. For example, a tap point 518 is located after a first delay stage. For the simplicity of Fig. 5, the other tap points have no reference numerals. In another example, the output clock 406(0) can contain any number of delays. In this example, the Fig. 5, the delay elements are connected in series. In another example (not shown), the delay elements can be arranged in parallel.
[0040] As in Fig. 5, the output clock generator 406(0) includes an output phase multiplexer (mux) 516 configured to receive the input clock signal 411 and each delayed clock signal as modified by each delay. The output phase mux 516(0) is further configured to receive the output phase selection 514(0). The output phase 514(0) determines the phase with which the output phase mux 516(0) must provide the output clock signal 403(0). The output phase selection is provided by an update device, which is described below with reference to Fig. 7. The output clock generator 406(0) is configured to modulate the input clock signal 411 to accommodate operating parameters of an electronic component configured to receive the output clock signal 403(0) of the output clock generator 406(0).
[0041] As previously with reference to Fig. 4, a particular component may require a different clock frequency compared to the frequency of the input clock signal 411 and / or other components. Clock modulation of the output clock 406(0) allows the system to provide a required clock frequency for a particular component without adding another PLL.
[0042] In Fig. 5, by using the output phase mux 516, the output clock 406(0) can modulate the input clock signal 411 by changing the tap point (e.g., by jumping forward to a later tap point, or by jumping back to an earlier tap point) at high speed. By continuously jumping from one tap to a later tap in the sequence of delay stages, the output clock 406(0) can provide an output clock signal 403(0) that is slower than the input clock signal 411. On the other hand, by continuously jumping back from one tap to an earlier tap, the output clock 406(0) can provide an output clock signal 403(0) that is faster than the input clock signal 411. These concepts of jumping forward or backward between taps will be explained in more detail below with reference to Fig. 6 explained.
[0043] Fig. Figure 6 is a diagram illustrating the exemplary mathematics of clock frequency modulation performed by an output clock 406(0) according to an embodiment of the present invention. In this example, assume that the output clock 406(0) needs to generate a clock signal having a frequency of 95.2 MHz (or 10.5 ns / cycle) and that the output clock receives an input clock signal 411 of 100 MHz (or 10.0 ns / cycle). A typical configuration would be to provide an additional PLL to generate the required 95.2 MHz, since otherwise there would be no way to obtain the required 95.2 MHz from the 100 MHz input clock signal 411. In the present system, on the other hand, instead of adding another PLL, the system is enabled to modulate the 100 MHz input clock signal to generate the required frequency of 95.2 MHz.
[0044] In the example from Fig. 6, the input clock signal is evenly divided into 10 phases: phase 0, phase 36, phase 72, phase 108, phase 144, phase 180, phase 216, phase 252, phase 288, phase 324, and phase 360 (or phase 0). Phase 0 is the phase of the original input clock signal 411. Phase 360 is shifted 360 degrees from phase 0, but for practical purposes can be considered equivalent to phase 0. Phase 36 is shifted 36 degrees from phase 0. Phase 72 is shifted 72 degrees from phase 0, and so on. The duration of each clock cycle of each of the phases is 10.0 ns / cycle (or 100 MHz), which is the duration of the cycle of the input clock signal 411. The time interval between two adjacent phases is 0.1 ns. The delay times do not necessarily have to be the same. In another embodiment (not shown), at least two of the delay times are not the same.
[0045] To generate a frequency of 95.2 MHz (or 10.5 ns / cycle) for the output clock signal 403(0), the output clock generator 406(0) shifts the output clock signal 403(0) at specified times according to an appropriate phase. One result is that some clock cycles of the output clock signal 403(0) are longer than the required duration (e.g., 10.5 ns), while some clock cycles are shorter than the required duration (e.g., 10.5 ns). Over a longer period of time, an average duration of the clock cycles is essentially equal to the sum of all clock cycles divided by the number of generated clock cycles. The average duration of the clock cycles simulates the required frequency (e.g., 10.5 ns) of the output clock signal 403(0).
[0046] In this example the Fig. 6, half of the clock cycles of the output clock signal 402(0) each have a duration of 10 ns, while half of the clock cycles each have a duration of 11.0 ns. Over a longer period of time, the average clock cycle duration is approximately 10.5 ns, which is equal to (10.0 ns + 11.0 ns) / 2. Therefore, the average clock cycle duration simulates the required frequency of 95.2 MHz for the output clock signal 403(0).
[0047] As in Fig. 6, the output clock 406(0) does not shift the tap point for the first clock cycle of 10.0 ns. Then, at an appropriate time before the end of the second clock cycle, the output clock 406(0) shifts the tap point from phase 0 to phase 36 so that the second cycle has a duration of 11.0 ns. The appropriate time for the shift is when both phases (e.g., phase 0 and phase 36) are either high or when both phases are low, so that no dirty shift occurs. Next, the output clock 406(0) does not perform a shift for the third cycle of 10.0 ns. Then, at an appropriate time before the end of the fourth clock cycle, the output clock 406(0) shifts the tap point from phase 36 to phase 72 so that the fourth cycle has a duration of 11.0 ns.Again, the appropriate time for the shift is when both phases are either high or when both phases (e.g., phase 36 and phase 72) are low, so that a dirty shift does not occur. Next, the output clock 406(0) does not perform a shift for the fifth cycle of 10.0 ns. Then, at an appropriate time before the end of the sixth clock cycle, the output clock 406(0) shifts the tap point from phase 72 to phase 108, so that the sixth cycle has a duration of 11.0 ns. Again, the appropriate time for the shift is when both phases (e.g., phase 72 and phase 708) are either high or when both phases are low, so that a dirty shift does not occur.The output clock 406(0) proceeds to perform a sequence of shifts in a similar manner to simulate an average clock cycle, which in this example is 10.5 ns / cycle (or 95.2 MHz).
[0048] Note that the previous example shows how the output clock 406(0) can slow down the clock frequency (e.g., reducing 100 MHz in the input signal to 95.2 MHz in the output signal). Similarly, the output clock 406(0) can make a clock frequency faster (e.g., increasing a 100 MHz input signal to a 105 MHz output signal). To increase a clock frequency, instead of performing a sequence of shifts to the subsequent phase, the output clock 406(0) is configured to perform a sequence of shifts to a previous phase. As an example (not shown), for the first clock cycle of 10.0 ns, the output clock 406(0) performs no shift. Then, at an appropriate time before the end of the second clock cycle, the output clock 406(0) shifts the tap point from phase 0 to phase 324 so that the second cycle has a duration of 9.0 ns.The appropriate time for the shift is when both phases (e.g., phase 0 and phase 324) are either high or both phases are low, so that a dirty shift does not occur. Next, the output clock 406(0) does not perform a shift for the third cycle of 10.0 ns. Then, at an appropriate time before the end of the fourth clock cycle, the output clock 406(0) shifts the tap point from phase 324 to phase 288, so that the fourth cycle has a duration of 9.0 ns. Again, the appropriate time for the shift is when both phases are either high or both phases (e.g., phase 324 and phase 288) are low, so that a dirty shift does not occur. Next, the output clock 406(0) does not perform a shift for the fifth cycle of 10.0 ns.Next, at an appropriate time before the end of the sixth clock cycle, the output clock 406(0) shifts the tap point from phase 288 to phase 252 so that the sixth cycle has a duration of 9.0 ns. Again, the appropriate time for the shift is when both phases (e.g., phase 288 and phase 252) are high, or when both phases are low, so that a dirty shift does not occur. The output clock 406(0) continues to perform a sequence of shifts in a similar manner to simulate an average clock cycle, which in this example is 9.5 ns / cycle (or 105 MHz).
[0049] A method for the modulated clock device 402(0) is described below with reference to Fig. 9 is further explained. Clock update device
[0050] Fig. Figure 7 is another block diagram of a portion of the modulated clock device 402(0) of Fig. 4 according to an embodiment of the present invention. One purpose of updater 408(0) is to shift tap point 518(0) to the appropriate phase at the appropriate time. Each update must occur shortly before output clock 4060(0) generates the next output phase, since there is a non-zero time interval between each phase update and the generation of the next output phase.
[0051] The update device 408(0) is connected to the output clock generator 406(0), which is connected to the clock source 410. The update device 408(0) includes an update phase mux 718(0) having inputs to receive the phases of the input clock signal 411. In this example, the update phase mux 718(0) receives the original input clock signal 411 and 10 delayed phases of the input clock signal 411. An output phase fraction counter 726(0) is connected to the update phase mux 718(0) and the output phase mux 516(0). The output phase subcounter 726(0) is configured to receive an update clock signal 703(0) from the update phase mux 718(0) and send the output phases 514(0) to the output phase mux 516(0). An update phase device 722(0) is connected to the output phase subcounter 726(0) and the update phase mux 718(0).The update phase device 722(0) is configured to receive the output phases 514(0) from the output phase subcounter 726(0) and to provide an update phase 724(0) to the update phase mux 718(0).
[0052] The output phase subcounter 726(0) monitors (e.g., counts) the update clock signal 703(0) to provide the current output phase 514(0). The output phase subcounter 726(0) calculates a current tap point for the output phase mux 516(0). The output phase subcounter 726(0) is a fractional accumulator / subtractor modulo the number of active taps. For example, let N equal the number of taps for 360 degrees of phase of the input clock signal 411. Let UP equal the current output phase 514(0). Let OP equal the current update phase 724(0). Then, for illustrative purposes, the update phase device 722(0) can calculate the update phase UP according to the following exemplary equation: UP=[OP+int(34∗(N−1))]mod N
[0053] For example, if OP is 4 and N is 10, then UP is 0, which is equal to [4+int(3*(10-1) / 4)] mod 10. If OP is 5 and N is 10, then UP is 1, which is equal to [5+int(3*(10-1) / 4)] mod 10, and so on.
[0054] Equation 1 is an example of how the update phaser 723(0) may shift the update phase UP from the output phase OP such that an edge of the update clock signal 703(0) is substantially always between a negative edge and a positive edge of the output clock signal 403(0). This shift of the update phase ensures that there are substantially no signal errors (e.g., dirty circuits) on the output clock signal 403(0) due to the change of taps in the output phase mux 516(0). The present technique is not limited to Equation 1. The update phaser 722(0) may be configured to perform a shift in many alternative ways, all of which are within the scope of the present technique.
[0055] In this way, the frequency of the output clock signal 403(0) remains constant over a longer period at a fixed ratio to the input clock signal 411. The value to be added or subtracted in the output phase subcounter 726(0) is linked to the number of stages required for a 360-degree phase to ensure constant frequency operation over a longer period. For example, let us return to the example of Fig. 6; the output phase fraction counter 726(0) is configured to count a shift of 1.0 ns every other cycle to achieve a long-term frequency of 95.2 MHz (or 10.5 ns / cycle).
[0056] A method for the update device 408(0) is described below with reference to Fig. 10 is explained in more detail. Calibration device
[0057] Fig. Figure 8 is a more detailed block diagram of the modulated clock device 402(0) of Fig. 4 according to an embodiment of the present invention. One purpose of the calibration device 404(0) is to detect the position of the 360-degree phase (or the zero-degree phase) of the input clock signal 411. The exact location of the 360-degree phase may vary depending on the frequency of the input clock signal 411 and variations in the process, voltage, and temperature (PVT) of the modulated clock device 402(0).
[0058] Calibration device 404(0) is connected to clock source 410. In this example, calibration device 404(0) includes a sequence of XNOR gates and a sequence of low-pass filters 830(0). Each XNOR gate is configured to receive input clock signal 411 and one of the phases of input clock signal 411. For example, a first XNOR gate receives input clock signal 411 and the input clock signal with a phase of 36 degrees. A second XNOR gate receives input clock signal 411 and the input clock signal with a phase of 72 degrees. A third XNOR gate receives input clock signal 411 and the input clock signal with a phase of 108 degrees. A fourth XNOR gate receives input clock signal 411 and the input clock signal with a phase of 144 degrees. A fifth XNOR gate receives the input clock signal 411 and the input clock signal at a phase of 180 degrees.A sixth XNOR gate receives the input clock signal 411 and the input clock signal with a phase of 216 degrees. A seventh XNOR gate receives the input clock signal 411 and the input clock signal with a phase of 252 degrees. An eighth XNOR gate receives the input clock signal 411 and the input clock signal with a phase of 288 degrees. A ninth XNOR gate receives the input clock signal 411 and the input clock signal with a phase of 324 degrees. A tenth XNOR gate receives the input clock signal 411 and the input clock signal with a phase of 360 degrees.
[0059] In the example from Fig. 8, one output of each XNOR gate is connected to one of the low-pass filters 830(0) with a total of 10 low-pass filters. For simplicity, only a single low-pass filter is shown in Fig. 8. Each XNOR gate operates as a comparator for the input clock signal 411 and the particular phase of the input clock signal 411 received at the respective XNOR gates. Each XNOR gate outputs a high state (e.g., 1) when both the input clock signal 411 and the particular phase of the input clock signal 411 are substantially equal (e.g., when both inputs are 1 or when both inputs are 0). Each XNOR gate outputs a low state (e.g., 0) when one but not both inputs are high (e.g., when one input is 1 while the other output is 0).
[0060] Since the delay circuit sequence is of limited length (for example in Fig. 8 10 delays), the tap point for the output clock signal 403(0) must be fed back from a certain point. This feedback must be performed for both the acceleration and deceleration of the output clock signal 402(0). To accomplish this feedback, the calibration device 404(0) is configured to detect the specific delay at which the input clock signal 411 has been delayed by a phase of 360 degrees (or an integer factor of 360 degrees). Since the delays due to PVT are highly variable, the calibration device 404(0) is configured to continuously calculate and update the location of the phase at 360 degrees, otherwise serious signal skew or jitter can occur. Jitter is an undesirable deviation from the true periodicity of the input clock signal 411.When the 360 degree point changes, the calibration device 404(0) ensures smooth operation by making the change effective when the current tap point is not close to the 360 degree point.
[0061] Each of the low-pass filters is configured to switch to an active high state when the input clock signal 411 and the specific delayed clock signal have been substantially equal for a sufficiently long time. At the beginning of each measurement, the reset switch 832(0) is closed to discharge the low-pass filter capacitor. The measurement time should be determined so that a threshold is only exceeded for a maximum of two sampling points. It may be necessary to vary this measurement time dynamically (for example, when the modulated clock device 402(0) is in operation).
[0062] A method for the calibration device 404(0) is described below with reference to Fig. 11 described. Overview of the procedures
[0063] Fig. Figure 9 is a flowchart of method steps for generating an output clock signal according to one embodiment of the present invention. Although the method steps are described in connection with the output clock generator 406(0), those skilled in the art will recognize that any system suitable for performing the method steps in any order is within the scope of the present invention.
[0064] As shown, a method 900 begins with an action 905 in which the output clock generator 406(0) receives an input clock signal from a clock source. In an action 910, the output clock generator 406(0) applies a series of delays to the input clock signal to generate one or more delayed phases of the input clock signal. In an action 915, the output clock generator 406(0) sends the input clock signal and the one or more phases of the input clock signal to an output phase mux, and in an action 920, the output clock generator 406(0) selects an appropriate phase of the input clock signal from the input clock signal and the one or more delayed phases of the input clock signal. In an action 925, the output clock generator 406(0) generates an output clock signal based on the appropriate phase of the input clock signal. The method 1000 may then loop back to step 1005 and continue there.
[0065] Fig. 10 is a flowchart of method steps for updating a phase of an output clock signal according to an embodiment of the present invention. Although the method steps are described in connection with update means 408(0), those skilled in the art will recognize that any system capable of performing the method steps in any order is within the scope of the present invention.
[0066] As shown, a method 1000 begins with an action 1005 in which the update device 408(0) receives an input clock signal and one or more delayed phases of the input clock signal at an update phase mux. In an action 1007, the update phase mux sends an update clock signal to an output phase subcounter. In an action 1010, the update device 408(0) applies the output phase subcounter to generate an output phase. In an action 1015, the update device 408(0) sends the output phase to an output phase mux of an output clock and to an update phase device. In an action 1020, the update device 408(0) generates an update phase via the update phase device. In an action 1025, the update device 408(0) sends the update phase to the update phase mux.The method 1000 may then return to step 1005 and continue.
[0067] Fig. Figure 11 is a flowchart of method steps for automatically calibrating a modulated clock device according to one embodiment of the present invention. Although the method steps are described in connection with calibration device 404(0), those skilled in the art will recognize that any system capable of performing the method steps in any order is within the scope of the present invention.
[0068] As shown, a method 1100 begins with an action 1105 in which the calibrator 404(0) closes the reset switch of each low-pass filter to reset (e.g., short-circuit) the respective low-pass filter capacitor. In an action 1110, the calibrator 404(0) opens the respective reset switch. In an action 1115, for each low-pass filter, the calibrator 404(0) waits a certain time T, which is necessary for at least one tapping point (e.g., XNOR gate) to cause a capacitor to exceed the threshold when the signals are (almost) completely in phase. In an action 1120, the calibrator 404(0) reads values received from all threshold comparators (e.g., XNOR gates). In a decision operation 1125, the calibration device 404(0) determines whether at least one comparator has exceeded the capacitor threshold.If not, the calibrator 404(0) increases the time period T in an action 1130 and then returns to action 1105. If at least one comparator has exceeded the capacitor threshold, the calibrator 404(0) proceeds to a decision operation 1135, in which the calibrator 404(0) determines whether more than two comparators have exceeded the capacitor threshold. If so, the calibrator 404(0) decreases the time period T in an action 1140 and returns to action 1105. However, if more than 2 comparators have not exceeded the capacitor threshold (e.g., only 1 or 2 comparators have exceeded the capacitor threshold), then the calibrator 404(0) proceeds to an action 1145 in which the calibrator 404(0) detects and / or determines that an appropriate number N of tapping points is present.Therefore, the calibration device 404(0) may inform the output phase fraction counter 726(0) or another device of the modulated clock device 402(0) that the appropriate number N is either present or not present.
[0069] These procedures of Fig. 9-11 may include other actions and / or details not explained in this overview of the method. Other actions and / or details are explained with reference to other figures and may be part of these methods, depending on the implementation. For example, if Fig.8, and if the calibration device 404(0) determines that the appropriate number N is different from a previous appropriate number N, then the calibration device 404(0) sends updated calibration parameters 832(0) to the output phase subcounter that reflect this new reality (e.g., the new appropriate number N). Further, the calibration parameters 832(0) may also include a desired frequency for the output clock signal 403(0); the modulated clock device 402(0) is configured to attempt to generate the desired frequency by coordinating operations of the output clock 406(0), the update device 408(0), and the calibration device 404(0).
[0070] Note that the calibrator 404(0) can operate in parallel with the output clock 406(0) because the comparators (e.g., XNOR gates) do not directly affect the output clock signal 403(0). The calibrator 404(0) can operate at finite intervals or continuously, depending on how rapidly PVT conditions are expected to change. For example, if PVT conditions are known to change very slowly, the calibrator 404(0) can be equipped with only a phase and threshold detector (e.g., an XNOR gate, a single low-pass filter, and a mux for phase selection); the calibrator 404(0) can then sequentially measure the optimal phase (e.g., for the first tap, then for the second tap, then for the third tap, etc.).Such a restriction saves space (e.g., silicon area) because there is only a single low-pass filter and a single XNOR gate. Alternatively, if the specifications of the delay elements are known and it is ensured that they are substantially constant (e.g., if the delay elements are substantially constant with respect to the PVT conditions), then the calibration device 404(0) may not be required. Such a reliable condition may occur, for example, if the clock source (e.g., PLL device) is configured to automatically generate different phases of a single clock signal.
[0071] Advantageously, the modulated clock system of the present technique can generate an output clock signal that is slightly faster or slightly slower than an existing source clock signal, and this can be accomplished without the need for an additional PLL for each device requiring a slightly different clock frequency.
[0072] The invention has been described with reference to specific embodiments. However, those skilled in the art will recognize 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 foregoing description and drawings are therefore to be considered as illustrative and not restrictive.
Claims
[1] An output clock generator (406) for modulating an input clock signal (411) to satisfy operating parameters of an electronic component configured to receive an output clock signal (403) of the output clock generator (406), the output clock generator (406) being configured to be included in a modulated clock device, the output clock generator (406) comprising: an output phase multiplexer (516); and a sequence of delay devices connected to the output phase multiplexer (516) and configured to receive the input clock signal from a clock source having a first frequency and to generate one or more delayed phases of the input clock signal and to send the one or more delayed phases of the input clock signal to the output phase multiplexer (516); wherein the output phase multiplexer (516) is further configured to select a suitable phase of the input clock signal (411) from the input clock signal and the one or more delayed phases of the input clock signal and to generate the output clock signal (403) based on the suitable phase of the input clock signal, wherein different phases are selected from the one or more delayed phases of the input clock signal at different times to change the first frequency, wherein a tapping point is arranged after each delay device in the sequence of delay devices, wherein each tapping point comprises a node that is common to another delay device of the sequence of delay devices and to the output phase multiplexer (516),wherein by continuously jumping from one tapping point to a later tapping point in the sequence of delay devices, an output clock signal is provided whose frequency is slower than the first frequency of the input clock signal, wherein by continuously jumping from one tapping point to an earlier tapping point in the sequence of delay devices, an output clock signal is provided whose frequency is faster than the first frequency of the input clock signal., [2] The output clock (406) of claim 1, wherein the clock source comprises an electronic circuit including a phase-locked loop. [3] The output clock (406) of claim 1, wherein each delay device of the sequence of delay devices is configured to receive a clock signal and provide a delayed clock signal delayed by a phase relative to the received clock signal. [4] The output clock (406) of claim 1, wherein the sequence of delay devices is formed as a series. [5] The output clock (406) of claim 1, wherein the sequence of delay devices is arranged at least partially in parallel. [6] The output clock (406) of claim 1, wherein the output phase multiplexer (516) is further configured to receive a selection of an output phase on which the phase at which the output phase multiplexer is to provide the output clock signal is based. [7] The output clock generator (406) of claim 1, wherein the output phase multiplexer (516) is further configured to supply the output clock signal to a component requiring a clock frequency different from a frequency of the input clock signal. [8] The output clock (406) of claim 1, wherein the output clock includes a phase-locked loop and wherein the output clock signal has a frequency that is different from a frequency of the input clock signal. [9] A computing device with: one or more electronic components; and a modulated clock system configured to supply one or more output clock signals to the one or more electronic components using one or more modulated clock devices, wherein at least one modulated clock device comprises an output clock according to claim 1.
Citation Information
Patent Citations
Method and apparatus for minimizing skew between signals
US20110221497A1
