Patent Grant
9547027
The field relates to semiconductor devices for use in a variety of systems.
Advances in semiconductor processing and logic design have permitted an increase in the amount of logic that may be present on integrated circuit devices. As a result, computer system configurations have evolved from a single or multiple integrated circuits in a system to multiple hardware threads, multiple cores, multiple devices, and/or complete systems on individual integrated circuits. Additionally, as the density of integrated circuits has grown, the power requirements for computing systems (from embedded systems to servers) have also escalated. Furthermore, software inefficiencies, and its requirements of hardware, have also caused an increase in computing device energy consumption. In fact, some studies indicate that computing devices consume a sizeable percentage of the entire electricity supply for a country, such as the United States of America. As a result, there is a vital need for energy efficiency and conservation associated with integrated circuits. These needs will increase as servers, desktop computers, notebooks, ultrabooks, tablets, mobile phones, processors, embedded systems, etc. become even more prevalent (from inclusion in the typical computer, automobiles, and televisions to biotechnology).
Power management for integrated circuits such as processors (used in both server and client systems) depends on accurate measurements of estimates of current processor power consumption. Various components of a processor may have their voltage and frequency modulated to stay within specified power limits. Since exceeding a power constraint is undesirable, processors are tuned to always stay under the power limit. Errors in power measurement are accounted for as a guardband, resulting in reduced power being available for processor performance.
Consider a processor that has a 100 watt (W) power limit, also referred to as a thermal design power (TDP), and a +/−5% error in power measurement. Since the processor must stay below its power limit, it caps power when a power consumption of 95 W is measured, making 5 W unavailable for use because it is reserved as a guardband. The size of this guardband is directly proportional to the amount of the error. These guardbands thus reduce available power. Further, inaccuracies of different types of power monitors can vary at low and high loads.
Embodiments may be used to provide a power measurement for a processor that is highly accurate at all load levels of the processor. To this end, embodiments can obtain information both from a digital power meter and a voltage regulator-based current sensor to determine power consumption levels from such sensor information. More specifically, embodiments may generate a combined or hybrid power measurement based on sensor information obtained from these multiple power sensors. In this way, a more accurate power measurement can be obtained. Then based on this accurate power measurement, a greater power head room is realized. Thus based on this power information, a processor may be controlled to operate at higher operating frequencies and/or voltages to realize greater performance within a given power budget.
Referring now to
As seen, processor 110 may be a single die processor including multiple cores 120a-120n. In addition, each core may be associated with an individual voltage regulator 125a-125n. Accordingly, a fully integrated voltage regulator (FIVR) implementation may be provided to allow for fine-grained control of voltage and thus power and performance of each individual core. As such, each core can operate at an independent voltage and frequency, enabling great flexibility and affording wide opportunities for balancing power consumption with performance. As further seen, each core 120 can include at least one event counter 1220-122n to count certain events occurring on the core, such as instructions that use certain core circuitry (e.g., high power circuitry). This event information can be used to provide sensor information for a digital power meter in accordance with an embodiment of the present invention. Note that in some embodiments, the cores themselves may contain a digital power meter to use this information to generate a power consumption level for the core, and communicate this information to a power control unit described below.
Also although only a single event counter per core is shown for ease of illustration, understand that the scope of the present invention is not limited in this regard. For example, in other implementations of a number of event counters can be present per core. Each counter can be configured to count a number of instructions executed of a given type in the corresponding core. For example, all instructions of an instruction set architecture (ISA) can be associated with one of these counters, where instructions of roughly the same power consumption level (e.g., due to the units of the core used for such instruction's execution) can be associated or bucketed with the same counter. In this way, a relatively accurate measure of actual power consumption based on the instructions being executed in the cores can be achieved. In one embodiment, each of these counters may be associated with a given weight value, generally corresponding to its relative power consumption level. Thus in one embodiment, a digital power meter (DPM) can operate based on counting events, assigning an energy weight to each event, and scaling for temperature. The rate at which various events occur is an indicator of the dynamic power consumption of the processor.
Further understand that instead of having an internal (to the core) digital power meter, instead the weighted count information from each of the event counters (of each core) can be provided to logic of a power control unit that can perform digital power measurements based on this information. In addition, this power controller-based digital power meter can further receive information from other portions of a processor such as uncore or system agent circuitry, interface circuitry, interconnect circuitry and so forth. Based on all such information, this digital power meter can generate a relatively accurate measure of power consumption. Note that for the system agent or uncore circuitry, the information may be associated with cache accesses. For interconnect circuitry, relative bandwidth may be used as an indication of power consumption. Similarly, for interface circuitry, a measure of the amount of data packets sent and received can be a good proxy for power consumption. While such a digital power meter can be relatively accurate across a full load line of the processor, it can be very difficult to tune the meter appropriately for the various operations and events occurring within the processor. Accordingly, a hybrid power meter in accordance with an embodiment of the present invention can improve the accuracy without the need for more complex tuning of a digital power meter.
Still referring to
Also shown is a power control unit (PCU) 138, which may include hardware, software and firmware to perform power management operations with regard to processor 110. In various embodiments, PCU 138 may include logic to perform digital power measurements, as described above. In addition, PCU 138 may include logic to perform hybrid power measurements in accordance with an embodiment of the present invention. Furthermore, PCU 138 may be coupled via a dedicated interface to external voltage regulator 160. In this way, PCU 138 can instruct the voltage regulator to provide a requested regulated voltage to the processor. In addition, voltage regulator 160 can provide information regarding its current delivery to the processor. In different implementations, voltage regulator 160 can store this information in a register of the voltage regulator that the PCU accesses. Or a current sensor, located either in voltage regulator 160 or on the path between voltage regulator 160 and PCU 138 can provide this information. This current information can be used by power meter logic of PCU 138 to generate a power consumption level based on this current delivery. Thus a voltage regulator-based current sensor can directly measure the current supplied by voltage regulator 160 to the processor. When multiplied by the supply voltage, this provides a measurement of processor power consumption.
As will be described below, logic within PCU 138 can be used to both calculate power consumption levels in multiple manners, including as described above as well as possibly in other manners and then, using a hybrid power measurement logic in accordance with an embodiment of the present invention, determine a hybrid power consumption level based on a combination of these individual power consumption levels.
While not shown for ease of illustration, understand that additional components may be present within processor 110 such as uncore logic, and other components such as internal memories, e.g., one or more levels of a cache memory hierarchy and so forth. Furthermore, while shown in the implementation of
Although the following embodiments are described with reference to energy conservation and energy efficiency in specific integrated circuits, such as in computing platforms or processors, other embodiments are applicable to other types of integrated circuits and logic devices. Similar techniques and teachings of embodiments described herein may be applied to other types of circuits or semiconductor devices that may also benefit from better energy efficiency and energy conservation. For example, the disclosed embodiments are not limited to any particular type of computer systems, and may be also used in other devices, such as handheld devices, systems on chip (SoCs), and embedded applications. Some examples of handheld devices include cellular phones, Internet protocol devices, digital cameras, personal digital assistants (PDAs), and handheld PCs. Embedded applications typically include a microcontroller, a digital signal processor (DSP), network computers (NetPC), set-top boxes, network hubs, wide area network (WAN) switches, or any other system that can perform the functions and operations taught below. Moreover, the apparatus', methods, and systems described herein are not limited to physical computing devices, but may also relate to software optimizations for energy conservation and efficiency. As will become readily apparent in the description below, the embodiments of methods, apparatus', and systems described herein (whether in reference to hardware, firmware, software, or a combination thereof) are vital to a ‘green technology’ future, such as for power conservation and energy efficiency in products that encompass a large portion of the US economy.
Referring now to
As seen in
Still referring to
As further shown in
Finally, control passes to block 260 where an operating frequency and/or voltage of the processor can be controlled based on this hybrid power consumption level and a power limit for the processor. As an example, typical processors can have a thermal design power (TDP) that corresponds to a maximum power dissipation that the processor can output (that can be handled by a cooling system). Using this as a maximum value and understanding a present loading of the processor and thus a current power consumption level (namely this hybrid power consumption level), in addition to obtaining a more accurate power measurement regardless of where on the load line the processor is executing, it may be possible to increase the operating frequency and/or voltage to thus obtain greater performance while remaining within the power budget, namely the TDP. Although shown at this high level in the embodiment of
As discussed, in different embodiments different manners of combining the power consumption levels determined by the different sensors can occur. Referring now to
If instead this first power consumption level is greater than the threshold, control passes to block 320 where instead the second power consumption level can be used to control operating frequency and/or voltage, as a current-based sensor may be more accurate at higher load levels. Although the scope of the present invention is not limited in this regard in some embodiments this threshold level may correspond to a load level of the processor between approximately 40 and 60% of a processor utilization, e.g., corresponding to roughly half of a TDP value of the processor.
Still other manners of combining power consumption levels determined by multiple power sensors can occur. As an example, a correlation-based combination can occur. In this way, one of the power consumption levels can be used to apply a correction factor to the other power consumption level. Assume for example given that a DPM may be accurate at low loads and a current sensor method is more accurate at higher loads, a correlation factor can be computed by executing a low power load and obtaining sensor information, and then executing a high power load and obtaining sensor information.
Referring now to
Offset=First Power Consumption Level−Second Power Consumption Level.
In one embodiment, this offset corresponding to the difference between the power consumption levels can be stored in a power management storage, e.g., present in a PCU. Although described as being stored within a storage of the PCU, understand the scope of the present invention is not limited in this regard. For example, in other implementations, this offset value (in other calibration values discussed further below) can be stored in a non-volatile storage, e.g., a non-volatile storage associated with the basic input/output system (BIOS).
Still referring to
Slope=(Second Power Consumption Level−Offset)/Second Power Consumption Level. Note that this slope can also be stored in the power management storage.
These operations as discussed above in
Referring now to
Hybrid=Second Power Consumption Level×Slope+Offset.
Accordingly, a hybrid power consumption level can be determined and can be used at block 390 to control the operating frequency and/or voltage of the processor. Although shown at this high level in the embodiment of
Embodiments can be implemented in processors for various markets including server processors, desktop processors, mobile processors and so forth. Referring now to
In various embodiments, power control unit 455 may include a hybrid power meter logic 459 in accordance with an embodiment of the present invention. As described above, this power meter can receive sensor information from different power sensors, including a digital power meter and an analog or current-based sensor. Note that different implementations are possible, such as a hybrid power meter that receives sensor information from each of multiple pairs of such sensors, e.g., where each pair is associated with a given core. Or, digital power meters may be present in the individual cores (and other processor circuitry) and instead a single current sensor-based power meter may be present in the processor. Variations on these implementations are of course also possible.
Based on the information received from these various sensors, hybrid power meter logic 459 can combine the sensor information in a selected manner to obtain a very accurate measure of power consumption in the processor. In this way, processor 400 can be configured to operate with very little guardband from a maximum power consumption level, e.g., a TDP level of the processor.
With further reference to
Referring now to
In general, each core 510 may further include low level caches in addition to various execution units and additional processing elements. In turn, the various cores may be coupled to each other and to a shared cache memory formed of a plurality of units of a last level cache (LLC) 5400-540n. In various embodiments, LLC 540 may be shared amongst the cores and the graphics engine, as well as various media processing circuitry. As seen, a ring interconnect 530 thus couples the cores together, and provides interconnection between the cores, graphics domain 520 and system agent circuitry 550. In one embodiment, interconnect 530 can be part of the core domain. However in other embodiments the ring interconnect can be of its own domain.
As further seen, system agent domain 550 may include display controller 552 which may provide control of and an interface to an associated display. As further seen, system agent domain 550 may include a power control unit 555 which can include a hybrid power meter logic 559 in accordance with an embodiment of the present invention to dynamically and accurately measure power consumption in the processor to enable greater processor performance in view of greater available processing power. In various embodiments, this logic may execute the algorithms described above in one or more of
As further seen in
Embodiments may be implemented in many different system types. Referring now to
Still referring to
Furthermore, chipset 690 includes an interface 692 to couple chipset 690 with a high performance graphics engine 638, by a P-P interconnect 639. In turn, chipset 690 may be coupled to a first bus 616 via an interface 696. As shown in
Referring now to
In one aspect, a processor includes multiple cores to independently execute instructions, a first sensor to measure a first power consumption level of the processor based at least in part on events occurring on the cores, and a hybrid logic to combine the first power consumption level with a second power consumption level determined based on a dynamic current provided to the processor. To this end, a power controller included in or coupled to the processor may control at least one of an operating frequency and a voltage of the processor based on this combined consumption level and a power limit of the processor.
In another aspect, a method includes receiving, in a first logic of a processor, sensor information from a digital power meter of the processor, and calculating a first power consumption level of the processor using this sensor information; receiving, in the logic, sensor information from a current sensor configured to measure a current delivered by a voltage regulator coupled to the processor and calculating a second power consumption level of the processor using the sensor information from the current sensor; and combining, in the first logic, the first and second power consumption levels to obtain a hybrid power consumption level of the processor. From this information and a power limit of the processor, an operating frequency and/or voltage of the processor can be controlled.
Yet another aspect includes a system with a multicore processor and a system memory. The cores may each include one or more event counters to count events occurring on the core, a digital power meter to calculate a first power consumption level based on the count information, a second power meter to calculate a second power consumption level based on a current delivered to the processor from a voltage regulator, and a power controller including logic to generate a combined power consumption level of the processor using the first and second power consumption levels.
In another aspect, a processor means includes execution means each for independently executing instructions, sensor means for measuring a first power consumption level of the processor means based at least in part on events occurring on the execution means, and means for combining the first power consumption level and a second power consumption level of the processor means determined based on a dynamic current provided to the processor means. In turn an operating frequency and/or voltage of the processor means can be controlled via a controller means based on the combined first and second power consumption levels and a power limit of the processor means.
Embodiments may be used in many different types of systems. For example, in one embodiment a communication device can be arranged to perform the various methods and techniques described herein. Of course, the scope of the present invention is not limited to a communication device, and instead other embodiments can be directed to other types of apparatus for processing instructions, or one or more machine readable media including instructions that in response to being executed on a computing device, cause the device to carry out one or more of the methods and techniques described herein.
Embodiments may be implemented in code and may be stored on a non-transitory storage medium having stored thereon instructions which can be used to program a system to perform the instructions. The storage medium may include, but is not limited to, any type of disk including floppy disks, optical disks, solid state drives (SSDs), compact disk read-only memories (CD-ROMs), compact disk rewritables (CD-RWs), and magneto-optical disks, semiconductor devices such as read-only memories (ROMs), random access memories (RAMs) such as dynamic random access memories (DRAMs), static random access memories (SRAMs), erasable programmable read-only memories (EPROMs), flash memories, electrically erasable programmable read-only memories (EEPROMs), magnetic or optical cards, or any other type of media suitable for storing electronic instructions.
While the present invention has been described with respect to a limited number of embodiments, those skilled in the art will appreciate numerous modifications and variations therefrom. It is intended that the appended claims cover all such modifications and variations as fall within the true spirit and scope of this present invention.
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/US2012/031464 | 3/30/2012 | WO | 00 | 6/20/2013 |
| Publishing Document | Publishing Date | Country | Kind |
|---|---|---|---|
| WO2013/147849 | 10/3/2013 | WO | A |
| Number | Name | Date | Kind |
|---|---|---|---|
| 5163153 | Cole et al. | Nov 1992 | A |
| 5522087 | Hsiang | May 1996 | A |
| 5590341 | Matter | Dec 1996 | A |
| 5621250 | Kim | Apr 1997 | A |
| 5931950 | Hsu | Aug 1999 | A |
| 6564328 | Grochowski | May 2003 | B1 |
| 6748546 | Mirov et al. | Jun 2004 | B1 |
| 6792392 | Knight | Sep 2004 | B1 |
| 6823516 | Cooper | Nov 2004 | B1 |
| 6829713 | Cooper et al. | Dec 2004 | B2 |
| 6996728 | Singh | Feb 2006 | B2 |
| 7010708 | Ma | Mar 2006 | B2 |
| 7043649 | Terrell | May 2006 | B2 |
| 7093147 | Farkas et al. | Aug 2006 | B2 |
| 7111179 | Girson et al. | Sep 2006 | B1 |
| 7194643 | Gonzalez et al. | Mar 2007 | B2 |
| 7272730 | Acquaviva et al. | Sep 2007 | B1 |
| 7412615 | Yokota et al. | Aug 2008 | B2 |
| 7434073 | Magklis | Oct 2008 | B2 |
| 7437270 | Song et al. | Oct 2008 | B2 |
| 7454632 | Kardach et al. | Nov 2008 | B2 |
| 7529956 | Stufflebeam | May 2009 | B2 |
| 7539885 | Ma | May 2009 | B2 |
| 7730340 | Hu et al. | Jun 2010 | B2 |
| 20010044909 | Oh et al. | Nov 2001 | A1 |
| 20020194509 | Plante et al. | Dec 2002 | A1 |
| 20030061383 | Zilka | Mar 2003 | A1 |
| 20030065960 | Rusu | Apr 2003 | A1 |
| 20040064752 | Kazachinsky et al. | Apr 2004 | A1 |
| 20040098560 | Storvik et al. | May 2004 | A1 |
| 20040123170 | Tschanz | Jun 2004 | A1 |
| 20040139356 | Ma | Jul 2004 | A1 |
| 20040268166 | Farkas et al. | Dec 2004 | A1 |
| 20050022038 | Kaushik et al. | Jan 2005 | A1 |
| 20050033881 | Yao | Feb 2005 | A1 |
| 20050132238 | Nanja | Jun 2005 | A1 |
| 20050210905 | Burns et al. | Sep 2005 | A1 |
| 20060050670 | Hillyard et al. | Mar 2006 | A1 |
| 20060053326 | Naveh | Mar 2006 | A1 |
| 20060059286 | Bertone et al. | Mar 2006 | A1 |
| 20060069936 | Lint et al. | Mar 2006 | A1 |
| 20060090086 | Rotem et al. | Apr 2006 | A1 |
| 20060117202 | Magklis et al. | Jun 2006 | A1 |
| 20060184287 | Belady et al. | Aug 2006 | A1 |
| 20070005995 | Kardach et al. | Jan 2007 | A1 |
| 20070016814 | Rusu et al. | Jan 2007 | A1 |
| 20070016817 | Albonesi et al. | Jan 2007 | A1 |
| 20070079294 | Knight | Apr 2007 | A1 |
| 20070106827 | Boatright et al. | May 2007 | A1 |
| 20070156992 | Jahagirdar | Jul 2007 | A1 |
| 20070214342 | Newburn | Sep 2007 | A1 |
| 20070239398 | Song et al. | Oct 2007 | A1 |
| 20070245163 | Lu et al. | Oct 2007 | A1 |
| 20080028240 | Arai et al. | Jan 2008 | A1 |
| 20080082844 | Ghiasi | Apr 2008 | A1 |
| 20080168287 | Berry | Jul 2008 | A1 |
| 20080244294 | Allarey | Oct 2008 | A1 |
| 20080250260 | Tomita | Oct 2008 | A1 |
| 20090006871 | Liu et al. | Jan 2009 | A1 |
| 20090150695 | Song et al. | Jun 2009 | A1 |
| 20090150696 | Song et al. | Jun 2009 | A1 |
| 20090158061 | Schmitz et al. | Jun 2009 | A1 |
| 20090158067 | Bodas et al. | Jun 2009 | A1 |
| 20090172375 | Rotem et al. | Jul 2009 | A1 |
| 20090172428 | Lee | Jul 2009 | A1 |
| 20090235105 | Branover et al. | Sep 2009 | A1 |
| 20100083009 | Rotem | Apr 2010 | A1 |
| 20100115293 | Rotem et al. | May 2010 | A1 |
| 20100115309 | Carvalho et al. | May 2010 | A1 |
| 20100146513 | Song | Jun 2010 | A1 |
| 20100191997 | Dodeja et al. | Jul 2010 | A1 |
| 20100332877 | Yarch et al. | Dec 2010 | A1 |
| 20110154090 | Dixon et al. | Jun 2011 | A1 |
| 20120023345 | Naffziger | Jan 2012 | A1 |
| 20120079290 | Kumar | Mar 2012 | A1 |
| 20120144217 | Sistla | Jun 2012 | A1 |
| 20120204042 | Sistla | Aug 2012 | A1 |
| 20120246506 | Knight | Sep 2012 | A1 |
| 20130061064 | Ananthakrishnan et al. | Mar 2013 | A1 |
| 20130080803 | Ananthakrishnan et al. | Mar 2013 | A1 |
| 20130080804 | Ananthakrishnan et al. | Mar 2013 | A1 |
| 20130111120 | Ananthakrishnan et al. | May 2013 | A1 |
| 20130111121 | Ananthakrishnan et al. | May 2013 | A1 |
| 20130111226 | Ananthakrishnan et al. | May 2013 | A1 |
| 20130111236 | Ananthakrishnan et al. | May 2013 | A1 |
| Number | Date | Country |
|---|---|---|
| 1739080 | Feb 2006 | CN |
| 101901033 | Dec 2010 | CN |
| 101937265 | Jan 2011 | CN |
| 1 282 030 | May 2003 | EP |
| 2008083906 | Jul 2008 | WO |
| Entry |
|---|
| Intel Developer Forum, IDF2010, Opher Kahn, et al., “Intel Next Generation Microarchitecture Codename Sandy Bridge: New Processor Innovations,” Sep. 13, 2010, 58 pages. |
| Spec-Power and Performance, Design Overview V1.10, Standard Performance Information Corp., Oct. 21, 2008, 6 pages. |
| Intel Technology Journal, “Power and Thermal Management in the Intel Core Duo Processor,” May 15, 2006, pp. 109-122. |
| Anoop Iyer, et al., “Power and Performance Evaluation of Globally Asynchronous Locally Synchronous Processors,” 2002, pp. 1-11. |
| Greg Semeraro, et al., “Hiding Synchronization Delays in a GALS Processor Microarchitecture,” 2004, pp. 1-13. |
| Joan-Manuel Parcerisa, et al., “Efficient Interconnects for Clustered Microarchitectures,” 2002, pp. 1-10. |
| Grigorios Magklis, et al., “Profile-Based Dynamic Voltage and Frequency Scalling for a Multiple Clock Domain Microprocessor,” 2003, pp. 1-12. |
| Greg Semeraro, et al., “Dynamic Frequency and Voltage Control for a Multiple Clock Domain Architecture,” 2002, pp. 1-12. |
| Greg Semeraro, “Energy-Efficient Processor Design Using Multiple Clock Domains with Dynamic Voltage and Frequency Scaling,” 2002, pp. 29-40. |
| Diana Marculescu, “Application Adaptive Energy Efficient Clustered Architectures,” 2004, pp. 344-349. |
| L. Benin!, et al., “System-Level Dynamic Power Management,” 1999, pp. 23-31. |
| Ravindra Jejurikar, et al., “Leakage Aware Dynamic Voltage Scaling for Real-Time Embedded Systems,” 2004, pp. 275-280. |
| Ravindra Jejurikar, et al., “Dynamic Slack Reclamation With Procrastination Scheduling in Real-Time Embedded Systems,” 2005, pp. 13-17. |
| R. Todling, et al., “Some Strategies for Kalman Filtering and Smoothing,” 1996, pp. 1-21. |
| R.E. Kalman, “A New Approach to Linear Filtering and Prediction Problems,” 1960, pp. 1-12. |
| International Application No. PCT/US2012/028865, filed Mar. 13, 2012, entitled “Providing Efficient Turbo Operation of a Processor,” by Intel Corporation. |
| International Application No. PCT/US2012/028902, filed Mar. 13, 2012, entitled “Dynamically Controlling Interconnect Frequency in a Processor,” by Intel Corporation. |
| International Application No. PCT/US2012/028876, filed Mar. 13, 2012, entitled “Dynamically Computing an Electrical Design Point (EDP) for a Multicore Processor,” by Intel Corporation. |
| U.S. Appl. No. 13/600,568, filed Aug. 31, 2012, entitled, “Configuring Power Management Functionality in a Processor,” by Malini K. Bhandaru, et al. |
| International Searching Authority, “Notification of Transmittal of the International Search Report and the Written Opinion of the International Searching Authority,” mailed Nov. 16, 2012, in International application No. PCT/US2012/031464. |
| State Intellectual Property Office, P.R. China, Office Action mailed Jan. 14, 2016, in Chinese Patent Application No. 201280072149.X. |
| State Intellectual Property Office of the People's Republic of China, Second Office Action mailed Jun. 21, 2016 in Chinese Patent Application No. 201280072149.X. |
| Number | Date | Country | |
|---|---|---|---|
| 20140195828 A1 | Jul 2014 | US |
