WO2008125509A2 - Method and system for analog frequency clocking in processor cores - Google Patents
Method and system for analog frequency clocking in processor cores Download PDFInfo
- Publication number
- WO2008125509A2 WO2008125509A2 PCT/EP2008/054011 EP2008054011W WO2008125509A2 WO 2008125509 A2 WO2008125509 A2 WO 2008125509A2 EP 2008054011 W EP2008054011 W EP 2008054011W WO 2008125509 A2 WO2008125509 A2 WO 2008125509A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- frequency
- analog
- core
- chip
- processor
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/04—Generating or distributing clock signals or signals derived directly therefrom
- G06F1/08—Clock generators with changeable or programmable clock frequency
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/04—Generating or distributing clock signals or signals derived directly therefrom
- G06F1/06—Clock generators producing several clock signals
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03L—AUTOMATIC CONTROL, STARTING, SYNCHRONISATION, OR STABILISATION OF GENERATORS OF ELECTRONIC OSCILLATIONS OR PULSES
- H03L7/00—Automatic control of frequency or phase; Synchronisation
- H03L7/06—Automatic control of frequency or phase; Synchronisation using a reference signal applied to a frequency- or phase-locked loop
- H03L7/16—Indirect frequency synthesis, i.e. generating a desired one of a number of predetermined frequencies using a frequency- or phase-locked loop
- H03L7/18—Indirect frequency synthesis, i.e. generating a desired one of a number of predetermined frequencies using a frequency- or phase-locked loop using a frequency divider or counter in the loop
- H03L7/197—Indirect frequency synthesis, i.e. generating a desired one of a number of predetermined frequencies using a frequency- or phase-locked loop using a frequency divider or counter in the loop a time difference being used for locking the loop, the counter counting between numbers which are variable in time or the frequency divider dividing by a factor variable in time, e.g. for obtaining fractional frequency division
- H03L7/1974—Indirect frequency synthesis, i.e. generating a desired one of a number of predetermined frequencies using a frequency- or phase-locked loop using a frequency divider or counter in the loop a time difference being used for locking the loop, the counter counting between numbers which are variable in time or the frequency divider dividing by a factor variable in time, e.g. for obtaining fractional frequency division for fractional frequency division
- H03L7/1976—Indirect frequency synthesis, i.e. generating a desired one of a number of predetermined frequencies using a frequency- or phase-locked loop using a frequency divider or counter in the loop a time difference being used for locking the loop, the counter counting between numbers which are variable in time or the frequency divider dividing by a factor variable in time, e.g. for obtaining fractional frequency division for fractional frequency division using a phase accumulator for controlling the counter or frequency divider
Definitions
- This invention generally relates to data processing systems, and more specifically, to frequency clocking in processor cores. Even more specifically, in the preferred embodiment, the invention relates to the analog multi- frequency clocking in multi-chip/multi-core processors.
- EMI reduction with multiple oscillators makes "synchronous spreading" very difficult or impossible.
- Prior art technology is based on distribution of clocking signals across a wiring network known as a clock-tree. With the growth in the number of cores in multi-core microprocessors, clock-trees also grow into enormous complexity, creating serious chip layout design difficulties and translating into detractors to final product yield and related increase in manufacturing cost.
- a method of and system for frequency clocking in a processor core are provided. At least one processor core is provided, and that at least one processor core has a clocking subsystem for generating an analog output clock signal at a variable frequency. Digital frequency control data and an analog signal are both transmitted to that at least one processor core; and that processor core uses the received analog signal and digital frequency control data to set the frequency of the output clock signal of the clocking subsystem.
- multiple cores are asynchronously clocked and the core frequencies are independently set.
- a plurality of processor cores are provided, and each of the processor cores has a respective clocking subsystem for generating an analog output clock signal at a variable frequency.
- an analog signal and individual digital frequency control data are transmitted to each processor core; and each processor core receives the analog signal and digital frequency control data transmitted to the core, and uses the received analog signal and digital control data to set locally (on the core) the frequency of the output clock signal of the clocking subsystem of the processor core.
- the preferred embodiment of the invention provides a computing system (Server) clocking subsystem solution with a single system reference oscillator, which may be spread (for spread-spectrum) to satisfy EMl requirements.
- the invention achieves clock distribution to each core via a classical multi-cascade analog tree distribution network and a digital data distribution network to each core.
- Each core takes both inputs to generate a precise frequency clock for the core, which may be unique to that core.
- the local core clock synthesizer frequency is determined by the digital control data which is used in conjunction with the analog core clock input to set the precise core frequency of operation using digital signal processing or other digital means.
- the frequency can be established based upon a policy set by the server manufacturer or customer. For example, the frequency can be set to the maximum capability of each core based upon a particular voltage of operation for all cores.
- the frequency control information is sent to each core as moderate speed (10- 100 Mb/s) digital data words thereby avoiding the problems with high-speed analog signal transmission.
- the frequency control information has high noise immunity and low signal distortion since it is in the form of digital data.
- the frequency control information is sent as individual control data words (v data) to each core.
- the data is latched into the core "clock synthesizer memory" from the server SEEPROM, which contains the vital chip data (VCD) for each core in the server.
- the single system reference oscillator is set at a moderate frequency (10- 100 MHz), which is distributed to each core via analog transmission line techniques; phase locked loops (PLL), and re-drive circuits.
- the analog clock signal frequencies are kept moderate prior to the individual core clock synthesizers to avoid highspeed distortion effects.
- the system reference clock, chip clock, and generic core clock signals are continuously required to maintain a stable core clock.
- the fundamental core operating frequency changes infrequently (except for certain spread spectrum techniques) such that speed v data changes are infrequent and only periodic v data updates are sufficient to generate a clock for each core.
- Each core is running asynchronous from each of the other cores and with respect to local cache. It will be appreciated that, once the different regions of a chip are asynchronous, some handshaking/buffering will be required to transfer data between regions, so there will be some added latency. Techniques are known to minimize this latency. Nevertheless, the net performance gain of operating each core at its maximum frequency will be substantial
- the present invention can be applied to any processing platform that uses multi- microprocessor core silicon chips.
- client uP platforms storage controllers, data communication switches, etc.
- Figure 1 shows an analog multi- frequency clocking of a processor subsystem.
- FIG. 2 illustrates an analog multi- frequency clocking of processor chips.
- FIG. 3 shows a local core clock synthesizer embodying the present invention.
- Figure 4 shows an alternate processor configuration in which multi-core groups share an L2 cache.
- Figure 5 illustrates a further alternate processor configuration in which multi-core groups share an L2 cache and a common local clock generator.
- FIG. 1 illustrates a typical computing Server 100 that is composed of multiple microprocessor (uP) chips (N) 102 which has internal clocking functions (e.g. digital signal processor, DSP, core clock generator, etc.) that utilize the server reference oscillator (vR) as the basic system clock.
- uP microprocessor
- N microprocessor
- internal clocking functions e.g. digital signal processor, DSP, core clock generator, etc.
- VR server reference oscillator
- a Master PLL and distribution ASIC Application Specific Integrated Circuit
- the output of the Master PLL & Distribution ASIC is a chip clock signal (vch) that is distributed throughout the processor chip .
- the reference oscillator 104 clock frequency (v R) is a relatively low frequency (typically 10-100 MHz) such that it can be easily routed throughout the PC board without significant signal degradation yet fast enough to enable feasible up-conversions rates to insure the uP high speed clock (typically 5-10 GHz) is stable and remains within the platform deviation requirement (typically 10 -100 ppm, parts per million).
- the distribution network is generally point-to-point (illustrated in Figure 1) for best reference clock integrity with signal re-drive at the up-conversion points.
- the first up-conversion and re-drive point is the Master PLL 106 which is used to generate the chip frequency (v ch) clock for each microprocessor chip in the server.
- the Master PLL not only re-drives the signal but also multiplies the reference oscillator by typically 2- 10x.
- the uP chip clock signal is, in turn, distributed within a chip by a second level distribution ASIC for use by each core clock synthesizer to generate the fundamental core clock, described below.
- Figure 1 also shows the interconnection from the uP chips to the I/O Subsystem, System
- the Clustering fabric is used to interconnect multiple MCMs together to construct a larger multi-processor Server where the MCMs are connected in a symmetric multi-processing (SMP) configuration.
- SMP symmetric multi-processing
- the memory is coherent to all the processors within the SMP. In this case, all the MCMs are synchronized to a single
- Reference Oscillator 104 (illustrated in Figure 1 outside the MCM).
- the preferred method of this invention can also be used on a configuration of uP chips contained on multiple Single Chip Modules (SCM) mounted on a common glass epoxy printed circuit (PC) board.
- SCM Single Chip Modules
- PC common glass epoxy printed circuit
- This alternate packaging configuration may be used for smaller systems.
- the Distribution ASIC is also mounted in an SCM on the system board and interconnection to each processor chip is done via system PC board wiring.
- the MCM and/or PC board contains vital core frequency data (VCD) for each core in the server. This information is typically maintained in a Serial Electrically Erasable Programmable Read Only Memory (SEEPROM). This SEEPROM contains the vital core frequency data (v data) for each connected processor (core).
- V data is the digital representation of the optimum processor (core) frequency along with identification (Id) of the appropriate chip and core. The Id information is used to insure the correct VCD is transmitted and stored in the VCD Interface function on each chip, for all cores on the chip.
- the VCD is derived from the frequency characterization data, voltage characterization data, power characterization, etc. gathered by the Service Element (SE).
- SE Service Element
- the SE analyzes and reformats the data and loads the data into the system SEEPROM via an appropriate digital interface (e.g. I2C).
- the totality of data gathered and analyzed by the SE is used to set the optimum frequency, voltage, etc. for each core to achieve the highest performance possible or other policy established by the customer.
- a novel aspect of this invention is the use of data to generate the optimum processor frequency locally (within core) in conjunction with the up-converted reference clock versus today's approach of transmitting the same analog clock signal to all cores.
- the data for each core/chip can be obtained during the chip test/verification stage in the manufacturing process or as part of a training paradigm during power-on sequence of the server. The latter approach would be part of the initialization and set-up process of the server.
- a representative server processor chip (one of several for a typical server) configuration with multi-cores (4) and shared L2 cache is illustrated at 200 in Figure 2.
- the four core clock synthesizers 202 within the processor chip receive the generic core clock (v gc) from the second level PLL and distribution ASIC 204 by means of the second level distribution network, which is contained on the chip.
- the generic core clock signal (v gc) is transmitted to each core using a multi-drop bus (illustrated) or a point-to-point star interconnection.
- the second level distribution ASIC 204 provides the necessary frequency up-conversion to generate the generic core clock (typically 10-2Ox), re-drive circuits, and a clock (v ch ) for the VCD Interface function.
- the VCD Interface function contains the VCD interface to the SEEPROM (See Figure 1) to receive and store the appropriate data for setting the precise frequency of each of the cores within the chip along with the appropriate Ids.
- the VCD Interface function interrogates the VCD interface to the SEEPROM (See Figure 1) to receive and store the appropriate data for setting the precise frequency of each of the cores within the chip along with the appropriate Ids.
- the VCD Interface function interrogates the VCD interface to the SEEPROM (See Figure 1) to receive and store the appropriate data for setting the precise frequency of each of the cores within the chip along with the appropriate Ids.
- the VCD Interface function interrogates the VCD interface to the SEEPROM (See Figure 1) to receive and store the appropriate data for setting the precise frequency of each of the cores within the chip along with the appropriate Ids.
- the VCD Interface function interrogates the VCD interface to the SEEPROM (See Figure 1) to receive and store the appropriate data for setting the precise frequency of each of the cores within the chip along with the appropriate Id
- SEEPROM and obtains the appropriate data (typically through an I2C interface) for its' cores. It may contain some SRAM and state machines or small controller in addition to the I2C interface to perform this function.
- the VCD Interface function also performs the distribution function by transmitting the V Data to the appropriate core synthesizer only.
- a unique chip and core Id is included which is related to the chip and module serial number.
- This core Id is used by the VCD Interface function to route the V data to the appropriate port.
- V Data intended for core "0" is routed to port "DO" ( Figure 2).
- the V data is stored in the clock synthesizer and is used as the processor clock frequency data until it is updated by the VCD function on chip. If no changes are forthcoming, no data is sent from the VCD Interface function or the SEEPROM.
- the V data is not sent continuously, but only when it is updated. This is in contrast to the state-of-the-art analog technique where the signal must be sent continuously. However, the analog clock is sent continuously to ensure a stable core clock.
- Each core 206 is comprised of the microprocessor, dedicated cache 210, and the core clock synthesizer 202.
- the core frequency is set by the core clock synthesizer and the digital V data in the VCD for each core.
- Each core is likely to have different frequency settings.
- the number of cores within the processor chip is determined by the technology and manufacturing process capability. Four are shown in Figure 2 for illustrative purposes. The technical approach described herein easily scales with the number of cores, which will likely increase in the future.
- the chip 200 also contains the appropriate interfaces 210, 212, 214 to the I/O, Memory, and Fabric controllers.
- the design of the core clock synthesizer is illustrated at 300 in Figure 3. It is comprised of a voltage controlled high speed oscillator (VCO) 302, a low pass filter (LPF) 304, a digitally controlled integer-N divider 306, and a Delta-Sigma modulator 310 in conjunction with a digital signal processor (DSP) 312.
- VCO voltage controlled high speed oscillator
- LPF low pass filter
- DSP digital signal processor
- VCO design and technology The VCO is tuned to a precise fractional frequency by changing the analog control voltage up or down in precise increments to achieve the desired frequency.
- a portion of the core clock output of the VCO is sent to the integer-N divider, which divides the incoming core clock frequency by an integer N value from the Delta-Sigma modulator.
- the Delta-Sigma modulator provides an output bit stream of time discrete integer values such that the average of the division ratio is equal to the input desired fractional division ratio.
- the desired fractional division ratio is generated by the DSP.
- the DSP 312 converts the desired V data digital frequency value to the appropriate fractional division ratio to yield the desired optimum core frequency.
- the reference frequency may be set at the factory based on the desired generic core frequency, which is the basis for determining the desired fractional division ratio.
- the divided output signal of the Integer-N divider 302 is phase compared to the generic core frequency "v gc" in the analog phase detector 314. If the two signals are matched, no frequency correction signal is generated and the clock synthesizer core output is equal to the desired core frequency, which is defined by the core V data input to the DSP. If there is a mismatch, a correction signal voltage is generated, which is passed through a low pass filter (LPF) 304 to remove high frequency noise prior to being applied to the voltage-controlled oscillator (VCO) 302. The error signal directs the VCO to alter its' output frequency in the direction to drive the correction signal to zero and achieve a frequency match at the phase detector.
- LPF low pass filter
- VCO voltage-controlled oscillator
- each core is likely to be at a different frequency, any issues associated with electromagnetic interference (EMI) are likely to be mitigated and the need for spread spectrum techniques minimized. Nevertheless, this approach offers a novel spread spectrum technique, which is not available with today's technology to reduce EMI even further.
- the DSP could systematically add and subtract a predefined amount from the V data value in the Data Control Register 316. This is done in a way such that the mean value always remains the same as the base V data value.
- Each core clock frequency (VCO output) will oscillate about the mean frequency value based upon a spread spectrum oscillating frequency, which is independently chosen for each core. This approach allows the spread spectrum approach to be asynchronous for each core, thereby lowering the total EMI.
- An alternative is to have the spread spectrum oscillating frequency the same for each core.
- Inherent to the Delta-Sigma modulator is a harmonic dither driver, thereby eliminating the need to add an external dither modulator to effect the spread-spectrum EMI mitigation.
- Figure 4 illustrates at 400 an alternate processor chip configuration (versus Figure 2) where multi-core groups 402, 404 share an L2 cache 406, 410.
- the chip 400 also contains the appropriate interfaces to the I/O, Memory, and Fabric controllers (not shown).
- the generic core clock signal (v gc) is star connected to each core clock synthesizer 412.
- v ch is shown as direct connected to the VCD Interface function 414 from the Master PLL & Distribution ASIC but may include a re-drive circuit at the junction point.
- the digital clocking attributes and functions discussed for Figure 2 also apply to this configuration.
- the configuration in Figure 4 could have common L2 cache clocking frequency or separate frequencies, depending on regional variability in cache. This arrangement is optimal for wiring resource: local processor/Ll cache clock grids, and Vdd (power supply voltage) grids.
- the output signal from the VCO to each core, or, to any grouping or subset of cores on the multicore processor chip provides a natural interconnected organization which enables a locally addressable switch or 'gate control' to selectively shut- off any pathway to said core or grouping of cores.
- the switching off of the local core clock(s) enables fine-grained power management without inducing power-fluctuations in the power-grid supply voltage, since the present invention teaches a method of clock frequency control not based on the use of varying the power supply or power grid voltages, nor, of varying Vdd.
- workload monitors via autonomic sensor circuits can turn off idle cores, or, redistribute workloads to optimize performance at a minimum physically possible power point.
- the present invention recognizes and specifically points out the significant distinguishable advantages of eliminating noise effects associated with voltage (or power) grid variations or voltage-island designs used in prior art approaches for clock frequency variation.
- Figure 5 illustrates at 500 another alternate processor chip configuration where multi-core groups 502, 504 share an L2 cache 506, 510 and a common local clock generator 512, 514.
- each core group of four contains one clock generator.
- Figure 5 shows the core clock is multi-dropped to two cores but other interconnection topologies (e.g. star) can be used.
- the chip also contains the appropriate interfaces to the I/O, Memory, and Fabric controllers (not shown). The digital clocking attributes and functions discussed for
- FIG. 2 also apply to this configuration.
- This configuration has a common local frequency for a region of cores and the local shared cache.
- the granularity of clocking by core or core groups depends on the nature of technology variability, size of cores, etc.
- the present invention enables a level of scalability and flexibility that is not readily available with today's state-of-the art.
- the optimum core operation frequency can be determined by varying the local frequency and Vdd (power supply voltage), and the invention enables in- field calibration of optimal operating conditions (if processor circuits degrade with time or environmental operating conditions).
- the instant invention also enables redundant clocks - that is, each local clock generator could have a "Bypass" mode to allow a generic system clock or another core's clock to be used in the event that the local clock generator circuit fails (or shows low yield in early mfg.).
- clock information is in digital format (data) at relatively low speed.
- the invention may be used with a core cache (Ll) synchronous with the core, but with a separate Vdd from the core.
- Ll core cache
- the invention may also be used with a cache that is asynchronously shared among a set of processors; shown herein as running at a system frequency (ns), but the cache could also have a local, independent clock generator.
- different cores/regions/cache can have different Vdd and different frequencies
- local clock grid(s) can be driven by, for example, a local clock source or a global chip clock grid driven by a global chip clock.
- the present invention allows global spread-spectrum from the system reference oscillator; each local clock generator may track the system reference oscillator spreading to avoid the "out-of-phase spreading" problem.
- digital spread spectrum techniques via the DSP may also be used.
- Computer program, software program, program, or software in the present context mean any expression, in any language, code or notation, of a set of instructions intended to cause a system having an information processing capability to perform a particular function either directly or after either or both of the following: (a) conversion to another language, code or notation; and/or (b) reproduction in a different material form.
Abstract
Description
Claims
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2010502492A JP5306319B2 (en) | 2007-04-12 | 2008-04-03 | Method and system for analog frequency clocking in a processor core |
CN2008800115703A CN101652737B (en) | 2007-04-12 | 2008-04-03 | Method and system for analog frequency clocking in processor cores |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US11/734,334 | 2007-04-12 | ||
US11/734,334 US8161314B2 (en) | 2007-04-12 | 2007-04-12 | Method and system for analog frequency clocking in processor cores |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2008125509A2 true WO2008125509A2 (en) | 2008-10-23 |
WO2008125509A3 WO2008125509A3 (en) | 2009-01-22 |
Family
ID=39854853
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/EP2008/054011 WO2008125509A2 (en) | 2007-04-12 | 2008-04-03 | Method and system for analog frequency clocking in processor cores |
Country Status (6)
Country | Link |
---|---|
US (1) | US8161314B2 (en) |
JP (1) | JP5306319B2 (en) |
KR (1) | KR20100003727A (en) |
CN (1) | CN101652737B (en) |
TW (1) | TWI417700B (en) |
WO (1) | WO2008125509A2 (en) |
Families Citing this family (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7917799B2 (en) * | 2007-04-12 | 2011-03-29 | International Business Machines Corporation | Method and system for digital frequency clocking in processor cores |
US7945804B2 (en) * | 2007-10-17 | 2011-05-17 | International Business Machines Corporation | Methods and systems for digitally controlled multi-frequency clocking of multi-core processors |
US7996743B1 (en) * | 2008-04-01 | 2011-08-09 | Altera Corporation | Logic circuit testing with reduced overhead |
JP2011160369A (en) * | 2010-02-04 | 2011-08-18 | Sony Corp | Electronic circuit, electronic apparatus, and digital signal processing method |
US8484495B2 (en) * | 2010-03-25 | 2013-07-09 | International Business Machines Corporation | Power management in a multi-processor computer system |
US8943334B2 (en) | 2010-09-23 | 2015-01-27 | Intel Corporation | Providing per core voltage and frequency control |
TW201250520A (en) * | 2011-06-13 | 2012-12-16 | Waltop Int Corp | Digitizer integration chip |
CN102445916B (en) * | 2011-09-15 | 2014-04-02 | 福建星网锐捷网络有限公司 | Method and system of programmable controller and clock frequency control |
US9471088B2 (en) | 2013-06-25 | 2016-10-18 | Intel Corporation | Restricting clock signal delivery in a processor |
US9377836B2 (en) * | 2013-07-26 | 2016-06-28 | Intel Corporation | Restricting clock signal delivery based on activity in a processor |
US9552034B2 (en) * | 2014-04-29 | 2017-01-24 | Qualcomm Incorporated | Systems and methods for providing local hardware limit management and enforcement |
KR102032330B1 (en) * | 2014-06-20 | 2019-10-16 | 에스케이하이닉스 주식회사 | Semiconductor device and its global synchronous type dynamic voltage frequency scaling method |
US20160283333A1 (en) * | 2015-03-25 | 2016-09-29 | International Business Machines Corporation | Utilizing a processor with a time of day clock error |
US20160283334A1 (en) * | 2015-03-25 | 2016-09-29 | International Business Machines Corporation | Utilizing a processor with a time of day clock error |
CN105049002B (en) * | 2015-07-02 | 2018-07-31 | 深圳市韬略科技有限公司 | A kind of method of the spreading device of electromagnetic compatibility and generation spread spectrum clock signal |
US10430354B2 (en) * | 2017-04-21 | 2019-10-01 | Intel Corporation | Source synchronized signaling mechanism |
CN108984469A (en) * | 2018-06-06 | 2018-12-11 | 北京嘉楠捷思信息技术有限公司 | Chip frequency modulation method and device of computing equipment, computing force board, computing equipment and storage medium |
CN113132272B (en) * | 2021-03-31 | 2023-02-14 | 中国人民解放军战略支援部队信息工程大学 | Network switching frequency dynamic adjustment method and system based on flow perception and network switching chip structure |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5594895A (en) * | 1992-12-15 | 1997-01-14 | International Business Machines Corporation | Method and apparatus for switching between clock generators only when activity on a bus can be stopped |
US20040251970A1 (en) * | 2003-05-29 | 2004-12-16 | Intel Corporation | Method for clock generator lock-time reduction during speedstep transition |
WO2006000929A2 (en) * | 2004-06-21 | 2006-01-05 | Koninklijke Philips Electronics N.V. | Power management |
US7187742B1 (en) * | 2000-10-06 | 2007-03-06 | Xilinx, Inc. | Synchronized multi-output digital clock manager |
Family Cites Families (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5481573A (en) * | 1992-06-26 | 1996-01-02 | International Business Machines Corporation | Synchronous clock distribution system |
US6256745B1 (en) * | 1998-06-05 | 2001-07-03 | Intel Corporation | Processor having execution core sections operating at different clock rates |
JPH1185303A (en) * | 1997-09-02 | 1999-03-30 | Nippon Steel Corp | Clock generation circuit |
JP3702148B2 (en) * | 1999-05-28 | 2005-10-05 | 三洋電機株式会社 | PLL device |
JP2001117903A (en) * | 1999-10-22 | 2001-04-27 | Seiko Epson Corp | Semiconductor integrated circuit device, microprocessor, microcomputer and electronic equipment |
US6990598B2 (en) | 2001-03-21 | 2006-01-24 | Gallitzin Allegheny Llc | Low power reconfigurable systems and methods |
US6993669B2 (en) | 2001-04-18 | 2006-01-31 | Gallitzin Allegheny Llc | Low power clocking systems and methods |
US7188261B1 (en) * | 2001-05-01 | 2007-03-06 | Advanced Micro Devices, Inc. | Processor operational range indicator |
US6898721B2 (en) | 2001-06-22 | 2005-05-24 | Gallitzin Allegheny Llc | Clock generation systems and methods |
JP4170218B2 (en) * | 2001-08-29 | 2008-10-22 | メディアテック インコーポレーテッド | Method and apparatus for improving the throughput of a cache-based embedded processor by switching tasks in response to a cache miss |
US6978389B2 (en) | 2001-12-20 | 2005-12-20 | Texas Instruments Incorporated | Variable clocking in an embedded symmetric multiprocessor system |
JP3638271B2 (en) * | 2002-07-23 | 2005-04-13 | 沖電気工業株式会社 | Information processing device |
US7124315B2 (en) * | 2002-08-12 | 2006-10-17 | Hewlett-Packard Development Company, L.P. | Blade system for using multiple frequency synthesizers to control multiple processor clocks operating at different frequencies based upon user input |
US7945803B2 (en) * | 2003-06-18 | 2011-05-17 | Nethra Imaging, Inc. | Clock generation for multiple clock domains |
JP2005085164A (en) * | 2003-09-10 | 2005-03-31 | Sharp Corp | Control method for multiprocessor system, and multiprocessor system |
JP2004152290A (en) * | 2003-10-24 | 2004-05-27 | Renesas Technology Corp | Semiconductor device |
US7321979B2 (en) | 2004-01-22 | 2008-01-22 | International Business Machines Corporation | Method and apparatus to change the operating frequency of system core logic to maximize system memory bandwidth |
US7350096B2 (en) * | 2004-09-30 | 2008-03-25 | International Business Machines Corporation | Circuit to reduce power supply fluctuations in high frequency/ high power circuits |
JP2006195602A (en) * | 2005-01-12 | 2006-07-27 | Fujitsu Ltd | System clock distribution device and system clock distribution method |
JP2007026075A (en) * | 2005-07-15 | 2007-02-01 | Canon Inc | Data processor and method for controlling data processor |
US7639764B2 (en) * | 2005-08-17 | 2009-12-29 | Atmel Corporation | Method and apparatus for synchronizing data between different clock domains in a memory controller |
US7478259B2 (en) * | 2005-10-31 | 2009-01-13 | International Business Machines Corporation | System, method and storage medium for deriving clocks in a memory system |
-
2007
- 2007-04-12 US US11/734,334 patent/US8161314B2/en active Active
-
2008
- 2008-04-03 JP JP2010502492A patent/JP5306319B2/en not_active Expired - Fee Related
- 2008-04-03 KR KR1020097014447A patent/KR20100003727A/en active Search and Examination
- 2008-04-03 WO PCT/EP2008/054011 patent/WO2008125509A2/en active Application Filing
- 2008-04-03 CN CN2008800115703A patent/CN101652737B/en active Active
- 2008-04-09 TW TW097112882A patent/TWI417700B/en active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5594895A (en) * | 1992-12-15 | 1997-01-14 | International Business Machines Corporation | Method and apparatus for switching between clock generators only when activity on a bus can be stopped |
US7187742B1 (en) * | 2000-10-06 | 2007-03-06 | Xilinx, Inc. | Synchronized multi-output digital clock manager |
US20040251970A1 (en) * | 2003-05-29 | 2004-12-16 | Intel Corporation | Method for clock generator lock-time reduction during speedstep transition |
WO2006000929A2 (en) * | 2004-06-21 | 2006-01-05 | Koninklijke Philips Electronics N.V. | Power management |
Non-Patent Citations (5)
Title |
---|
ANONYMOUS: "BIOS AND KERNEL DEVELOPER'S GUIDE FOR AMD ATHLONTM 64 AND AMD OPTERONTM PROCESSORS"[Online] February 2006 (2006-02), XP002505671 AMD Athlon 64 Processor Tech Docs page on Internet Retrieved from the Internet: URL:http://web.archive.org/web/20070315084839/http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/26094.PDF> [retrieved on 2008-11-26] * |
ANONYMOUS: "Enhanced Intel® SpeedStep® Technology for the Intel® Pentium M Processor - White Paper"[Online] March 2004 (2004-03), XP002505672 Retrieved from the Internet: URL:http://developer.intel.com/design/intarch/pentiumm/docs_pentiumm_proc.htm> [retrieved on 2008-11-26] * |
JIAN LI ET AL: "Dynamic Power-Performance Adaptation of Parallel Computation on Chip Multiprocessors" HIGH-PERFORMANCE COMPUTER ARCHITECTURE, 2006. THE TWELFTH INTERNATIONA L SYMPOSIUM ON AUSTIN, TEXAS FEBRUARY 11-15, 2006, PISCATAWAY, NJ, USA,IEEE, 11 February 2006 (2006-02-11), pages 77-87, XP010896312 ISBN: 978-0-7803-9368-4 * |
MURALIMANOHAR N ET AL: "Power efficient resource scaling in partitioned architectures through dynamic heterogeneity" PERFORMANCE ANALYSIS OF SYSTEMS AND SOFTWARE, 2006 IEEE INTERNATIONAL SYMPOSIUM ON AUSTIN, TX, USA MARCH 19-21, 2006, PISCATAWAY, NJ, USA,IEEE, 19 March 2006 (2006-03-19), pages 100-111, XP010910198 ISBN: 978-1-4244-0186-4 * |
WU-CHUN FENG AND CHUNG-HSING HSU: "Green Destiny and its Evolving Parts" 19TH INTERNATIONAL SUPERCOMPUTER CONFERENCE, [Online] June 2004 (2004-06), XP002505673 Heidelberg, Germany Retrieved from the Internet: URL:http://public.lanl.gov/radiant/pubs.html> [retrieved on 2008-11-26] * |
Also Published As
Publication number | Publication date |
---|---|
US8161314B2 (en) | 2012-04-17 |
JP2010524103A (en) | 2010-07-15 |
CN101652737A (en) | 2010-02-17 |
TW200907631A (en) | 2009-02-16 |
TWI417700B (en) | 2013-12-01 |
WO2008125509A3 (en) | 2009-01-22 |
KR20100003727A (en) | 2010-01-11 |
JP5306319B2 (en) | 2013-10-02 |
CN101652737B (en) | 2013-04-03 |
US20080256381A1 (en) | 2008-10-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8161314B2 (en) | Method and system for analog frequency clocking in processor cores | |
US7917799B2 (en) | Method and system for digital frequency clocking in processor cores | |
US7945804B2 (en) | Methods and systems for digitally controlled multi-frequency clocking of multi-core processors | |
US5481573A (en) | Synchronous clock distribution system | |
US10007293B2 (en) | Clock distribution network for multi-frequency multi-processor systems | |
US5572557A (en) | Semiconductor integrated circuit device including PLL circuit | |
US8604852B1 (en) | Noise suppression using an asymmetric frequency-locked loop | |
US8269544B2 (en) | Power-supply noise suppression using a frequency-locked loop | |
US20080174347A1 (en) | Clock synchronization system and semiconductor integrated circuit | |
CN117075683A (en) | Clock gating component, multiplexer component and frequency dividing component | |
KR100681287B1 (en) | System clock distributing apparatus and system clock distributing method | |
US7501865B1 (en) | Methods and systems for a digital frequency locked loop for multi-frequency clocking of a multi-core processor | |
KR20020045691A (en) | Apparatus for controling a frequency of bus clock in portable computer | |
US11558059B2 (en) | Concept for a digital controlled loop and a digital loop filter | |
CN112671403A (en) | Clock frequency division system, method and equipment | |
JP4156529B2 (en) | Selectable clocking architecture | |
WO1996035177A1 (en) | A modular system utilizing interchangeable processor cards | |
KR100278284B1 (en) | Method and apparatus for minimizing clock skew using synchronous bus clock and programmable interface | |
Włostowski et al. | White Rabbit and MTCA. 4 use in the LLRF upgrade for CERN’s SPS | |
CN214586628U (en) | Multi-path clock output circuit, circuit board and CT scanner | |
CN116015279A (en) | Clock configuration method, device, equipment and medium of programmable logic device | |
CN210129122U (en) | FPGA accelerator card online clock configuration device | |
JP4553428B2 (en) | Phase-locked loop (PLL) clock generator with programmable offset and frequency | |
Xiu | Spectrally Pure Clock versus Flexible Clock: Which One Is More Efficient in Driving Future Electronic Systems? | |
CN117273159A (en) | Quantum state driving signal generator, device and quantum computer system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200880011570.3 Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1020097014447 Country of ref document: KR |
|
ENP | Entry into the national phase |
Ref document number: 2010502492 Country of ref document: JP Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 08749515 Country of ref document: EP Kind code of ref document: A2 |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 08749515 Country of ref document: EP Kind code of ref document: A2 |