US20080046713A1 - Overriding processor configuration settings - Google Patents

Overriding processor configuration settings Download PDF

Info

Publication number
US20080046713A1
US20080046713A1 US11/973,837 US97383707A US2008046713A1 US 20080046713 A1 US20080046713 A1 US 20080046713A1 US 97383707 A US97383707 A US 97383707A US 2008046713 A1 US2008046713 A1 US 2008046713A1
Authority
US
United States
Prior art keywords
processor
register
override
bits
feature
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Abandoned
Application number
US11/973,837
Inventor
Edward Barragy
Chris Newburn
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Individual
Original Assignee
Individual
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Individual filed Critical Individual
Priority to US11/973,837 priority Critical patent/US20080046713A1/en
Publication of US20080046713A1 publication Critical patent/US20080046713A1/en
Abandoned legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/30098Register arrangements
    • G06F9/30101Special purpose registers
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/30181Instruction operation extension or modification

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Executing Machine-Instructions (AREA)
  • Advance Control (AREA)
  • Stored Programmes (AREA)
  • Debugging And Monitoring (AREA)

Abstract

In one embodiment, the present invention includes a method for setting an entry in an override register corresponding to a processor feature to override a processor configuration setting for the processor feature and overriding the processor configuration setting for the processor feature using the entry. The entry may be set with a user-level application, for example.

Description

  • This application is a continuation of U.S. patent application Ser. No. 10/959,356 filed Oct. 6, 2004 entitled “OVERRIDING PROCESSOR CONFIGURATION SETTINGS” the content of which is hereby incorporated by reference.
  • BACKGROUND
  • A processor, such as a central processing unit (CPU) of a system typically includes a number of processor features such as caches, buffers, arrays, and the like. The processor may further include various processor features that are logic units, such as a branch prediction unit, and a prefetch unit, among others. One way in which the performance of a processor is improved is through the use of a multiple-stage pipeline architecture, in which various pipeline features may be used to more efficiently execute instructions.
  • For the processor to perform instructions efficiently, certain features within the processor may be defaulted to an enabled or disabled state during the manufacture of the processor. Typically, this default state cannot be modified under user-level control. That is, certain processor features are enabled or disabled by default when the processor is manufactured, with different default configurations possible for different target markets. In other words, current practice is ‘all or nothing’, irrespective of the application executing on the processor. Thus, certain processor features are enabled or disabled according to a predetermined policy, typically determined before the system is shipped to an end user.
  • For example, a hardware prefetcher of a processor is typically configured to a default condition during manufacture. This default condition is set when a system including the processor is booted, for example, during basic input/output system (BIOS) execution. Thus control of the hardware prefetcher is not visible to an end user and cannot be reconfigured by the end user.
  • A need thus exists to improve processor performance with regard to particular processor features.
  • BRIEF DESCRIPTION OF THE DRAWINGS
  • FIG. 1 is a block diagram of a partial set of state registers of a processor in accordance with one embodiment of the present invention.
  • FIG. 2 is a block diagram of multiple override registers in accordance with another embodiment of the present invention.
  • FIG. 3 is a logic implementation used to couple override registers in accordance with one embodiment of the present invention.
  • FIG. 4 is a block diagram of override registers in accordance with another embodiment of the present invention.
  • FIG. 5 is a flow diagram of a method according to one embodiment of the present invention.
  • FIG. 6 is a block diagram of a representative computer system in accordance with one embodiment of the present invention.
  • DETAILED DESCRIPTION
  • In various embodiments, user-level software control of different aspects of a processor's micro-architectural configuration may be effected. While such user-level control may be achieved in many different ways, in certain embodiments one or more additional registers may be present within a processor to store information regarding the processor's configuration. In one embodiment, the additional register may be referred to as a micro-architectural configuration register, or MCR. As part of the processor's state, the MCR is saved and restored across context switches. In certain embodiments, each bit in the MCR may correspond to an enabled or disabled condition of a specific micro-architectural feature. As an additional register, reading and writing of the MCR is very fast, and is comparable in speed to modifying of general purpose (GP) registers within the processor.
  • Referring now to FIG. 1, shown is a block diagram of a partial set of state registers in a processor in accordance with one embodiment of the present invention. The state registers of FIG. 1 may be present in a processor in accordance with an Intel® architecture (IA-32 or x86) specification, although the scope of the present invention is not so limited. That is, while discussed primarily herein with reference to an IA-32 processor, embodiments may be implemented in other processors available from Intel Corporation such as an ITANIUM® processor, or processors of other manufacturers having a similar or different instruction set architecture (ISA). As shown in FIG. 1, the state registers may be within a core of a processor and may be used to store information regarding processor state for a given context. While shown with the specific registers present in FIG. 1, it is to be understood that in other embodiments, additional status, control and state registers may be present, in addition to other program execution registers.
  • As shown in FIG. 1, registers present may include a plurality of GP registers 10. Such registers may be used to store various information during processor operation, such as operands for logical and arithmetic operations, operands for address calculations, memory pointers, and the like. Additionally, status and control (EFLAGS) register 20 may include various status flags, control flags, and system flags. EFLAGS register 20 may be used to report on program status. Instruction pointer (EIP) register 30 may be used to store a pointer to the next instruction of a program to be executed.
  • As further shown in FIG. 1, a plurality of segment registers 40 may be present. Segment registers 40 may store segment selectors which act as a pointer to identify a memory segment. In other words, such segment registers 40 address code, data and stack segments in memory.
  • As further shown in FIG. 1, at least one micro-architectural configuration register (MCR) 50 is present. MCR 50 may be used to store information regarding different processor features. For example, each entry or bit of MCR 50 may correspond to a different processor feature. MCR 50 may be architecturally visible, such that it may be accessed by various software levels via the ISA. The outputs of each bit may be coupled to a corresponding feature or other processor hardware to control the enabling or disabling of the feature. Based on the status of the given bit, the corresponding feature may be enabled or disabled.
  • As an example, certain features may be enabled or disabled by default when a processor is manufactured. However, during operation, certain features may have their default setting changed (e.g., enabled to disabled) under software control. Such software control may be effected in user-level software, such as an application program. In such manner, non-operating system (OS) or non-BIOS control of processor features may be effected.
  • Collectively, the registers shown in FIG. 1, along with other registers not shown in FIG. 1, such as architectural control registers (e.g., control register 0, control register 1, control register 2, control register 3, and the like), may be considered to be context registers, as such registers may be programmed with different values based on a given context that is running on the processor. A context switch may occur when a multitasking OS stops running one process (e.g., an application or a software thread) and begins running another. Upon a context switch, the values stored in these registers may be saved away to save the processor state for the current context. Then these registers may be loaded with the status for the new context to provide its processor state. On another context switch, e.g., back to the original context, the saved processor state may be restored to the registers.
  • Certain processor features may be appropriate or inappropriate during a given program phase. Accordingly, in operation, bits of MCR 50 may be enabled or disabled based on information regarding the utility, efficiency and the like of a particular processor feature for a given application. For example, either via a predetermined policy, dynamic profiling, or during specific program phases it is known that enabling or disabling a specific feature improves overall program performance. Software may determine the appropriateness of overriding a processor configuration setting and cause enabling or disabling of processor features in various manners. First, when software such as an application program is written, a programmer may insert code to cause the desired feature to be turned on or off (i.e., enabled or disabled) at certain phases of a program. In other embodiments, a compiler may determine that a particular processor feature is desired for certain program phases and during compilation may insert code to cause the desired enabling or disabling of one or more processor features. Still further, in a managed runtime environment (MRTE) (e.g., a Java™ or .NET environment), profiling may be performed and the results of the profiling may indicate that certain processor features are appropriate during certain program phases, but are inappropriate during other phases.
  • Upon entering the first program phase, the appropriate bit of MCR 50 (for example) is set to either enable or disable the specified feature. Upon exiting of this program phase, the MCR bit is reset to its original value. Because the contents of MCR 50 is part of the processor state and is preserved across context switches, it is specific to a given software process and thread.
  • In other embodiments during dynamic profiling, it may be determined that when an application reaches a certain program counter (PC) value, a given feature is to be enabled or disabled. In still other embodiments, a given feature may be enabled or disabled for particular program types. For example, a hardware prefetcher (HWP) such as a striding hardware prefetcher may be enabled for a scientific computing application, while it may be disabled for a database application, as the mechanics of these different types of applications cause the prefetcher to be less effective for database operations.
  • While embodiments of the present invention may be used with a variety of micro-architectural or other processor features, the following discussion relates to a HWP. In the case of MRTE's for server-type applications a HWP may provide very large performance boosts during garbage collection (GC) phases of the computation but may hurt overall performance. Thus to improve performance, the HWP may be disabled during general phases of computation, but enabled during garbage collection. When a virtual machine (VM) of the MRTE enters a garbage collection phase, it accesses MCR 50 and sets a bit corresponding to the HWP configuration to ‘enabled’. When the garbage collection phase is completed this bit is set to ‘disabled’.
  • In other embodiments, a plurality of registers may be used to provide user-level control of processor features. In such manner, control of processor features, such as micro-architectural features, may support both a processor configuration setting that applies to all software threads by default, and hints by each software thread to override that default. Thus, a first register may be present to store information regarding default processor configuration settings and one or more override registers may be present for use by software to override the default setting. Then based on the information in the override registers, processor configuration settings may be overwritten by forcing a feature to be enabled or to be disabled.
  • Referring now to FIG. 2, shown is a block diagram of multiple registers in accordance with another embodiment of the present invention. As shown in FIG. 2, a processor configuration register (PCR) 70 is present. PCR 70 may include N bits, where each bit corresponds to a different processor feature, such as micro-architectural features and the like. PCR 70 may store values that correspond to a processor configuration setting, applicable as a default across all processes. For example, a value of 1 may indicate that the corresponding feature is enabled while a value of 0 may indicate that the corresponding feature is disabled.
  • However, the values in PCR 70 may be overwritten based on user-level control of a given software thread. Specifically, the default settings of PCR 70 may be overridden based on information in an enable override register (EOR) 80 and a disable override register (DOR) 90. EOR 80 and DOR 90 may be programmed during execution of a given software thread to override the default settings, and to cause a processor feature to be enabled or disabled. For example, an independent software vendor (ISV) developing an application program may include code to cause a particular processor feature to be enabled to improve operation and efficiency of the program, for example, during certain program phases.
  • In different embodiments, the corresponding bits of PCR 70, EOR 80 and DOR 90 may be logically combined in order to effect the desired disabling or enabling of a processor feature. The registers may be logically joined in different manners to effect feature selection under user control.
  • Referring now to FIG. 3, shown is a logic implementation used to couple override registers in accordance with one embodiment of the present invention. Specifically, FIG. 3 shows corresponding bits of PCR 70, EOR 80 and DOR 90. As further shown in FIG. 3, logic gates may be used to couple the corresponding bits. Specifically, an OR gate 92 may receive corresponding bits from PCR 70 and EOR 80. In turn, the output of OR gate 92 may be coupled to an input of an AND gate 94, which is also coupled to receive an inverted output of a corresponding bit of DOR 90. The resulting output of AND gate 94 may be used to control the processor configuration setting for the corresponding processor feature. For example, the output of AND gate 94 may be provided to circuitry of the processor feature to enable or disable the feature accordingly.
  • Using the logic implementation shown in FIG. 3, the combination of the bit settings for the three registers may cause associated processor features to be enabled or disabled. Specifically, under normal operating conditions without regard to EOR 80 and DOR 90, a value of 1 in bit 0 (for example) of PCR 70 indicates that the corresponding feature is enabled. A value of 1 in bit 0 of EOR 80 and DOR 90 indicates that the value of the PCR 70 is to be overridden, as enable and disable, respectively. That is, an active high value for a bit of EOR 80 is used to force the corresponding feature on, while an active high bit in DOR 90 is used to force the corresponding feature off.
  • For example, in a desktop system, a PCR bit corresponding to a HWP may be set to 1 ‘at the factory’ (or by an original equipment manufacturer (OEM) via BIOS) so that the HWP is enabled by default. Because such a HWP may be desirable for an application running on the desktop system, the application (i.e., user-level) code does nothing with respect to the override registers and the corresponding bits of EOR 80 and DOR 90 are defaulted to 0. Thus the default behavior for the feature is given by the PCR bit, i.e., the HWP is enabled for the desktop system.
  • However for a server system, it may be desired to disable the HWP by default. In this case, the PCR bit corresponding to the HWP is set ‘at the factory’ to 0. For this default setting, the corresponding bits of EOR 80 and DOR 90 are 0, and thus the HWP is disabled by default. During execution of a program or certain portions of a program, it may be desired to enable the HWP. For example, a Java™ application may be running on the server system. Based on profiling, it is known that the HWP may improve performance during a GC phase, thus the HWP may be enabled during GC, but disabled elsewhere during the application. Because the HWP is off by default, during GC software sets the corresponding bit of EOR 80 to 1. The corresponding bit of DOR 90 remains at its default value of 0. Based on the logic implementation shown in FIG. 3, this gives a final output of 1 from AND gate 94 and the HWP is enabled, as desired.
  • Conversely, suppose the same Java™ application is running on a desktop system. Here the PCR bit is set at the factory to 1. Because the application runs more efficiently without the HWP enabled during non-GC phases, at the end of a GC phase, software sets the corresponding DOR bit to 1. This value of DOR 90 indicates a disable override, and the HWP is disabled. The application executes until the start of the next GC phase with the HWP disabled, at which point it resets the DOR bit to 0, which enables the HWP. Because the software sets the bits at the appropriate program locations, the save/restore cost at context switches may be reduced.
  • As discussed above, the software hint settings are software thread specific and may be saved and restored upon context switches. In designs in which a save and restore procedure is optimized to save only those settings which are not a logic 0, the override registers may only be set if they are useful. If the override bit would provide the same setting as the processor configuration setting, i.e., there is no need for an override (it is useless), then the corresponding override bit is not set. In such an embodiment, the input to the appropriate bits of EOR 80 and DOR 90 may be an output of logically ANDing the software-specified bit and the inverted value of the corresponding processor configuration setting bit. If the hardware does not support functionality associated with the software—specified bits, there is nothing to override, and they are useless. Accordingly, those bits may be read only, with values of 0.
  • In certain embodiments, hardware may override the software hint overrides. That is, based on certain information the hardware may determine that the software hint values do not improve program execution or are inappropriate for a given hardware implementation and therefore the hardware may override such software hints.
  • As a result, if the hardware can determine empirically that the setting suggested by the software hint override is not effective, then hardware can clear the override bit. For example, the hardware may first accept a software hint and execute an application accordingly. However, the hardware may determine that execution is not efficient and the enabled or disabled feature does not improve processor performance or may violate another type of execution constraint, such as a thermal limit. In such case, the hardware may choose to override the software hint. In other embodiments, the hardware may choose to override a software hint where it is already using the processor feature according to the hint value. Or it may choose to override the software hint in situations where the hardware has better control and understanding of the processor feature. A hardware override may be appropriate if the software suggests that the hardware should try a speculative optimization that has a large potential upside, but that might be costly to try, hence the responsibility of indicating the possibility of potential gain rests with software.
  • Referring now to FIG. 4, shown is a block diagram of override registers in accordance with another embodiment of the present invention. As shown in FIG. 4, a processor configuration register (PCR) 70, an enable override register (EOR) 80 and a disable override register (DOR) 90 may be present. In the embodiment of FIG. 4, the override registers may be configured to provide for hardware override of the software override registers. Specifically, as shown in FIG. 4, bits of EOR 80 and DOR 90 may receive a reset signal. This signal may be used to override a software-specified value for the corresponding bit.
  • For example, for a given processor feature, the corresponding bit of PCR 70 may be set at a default value of 0. However, an application desires to override that setting and cause the feature to be enabled. Accordingly, the software causes the corresponding bit of EOR 80 to be set at a value of 1. The corresponding bit of DOR 90 may remain set at a default of 0, which would cause the corresponding feature to be enabled using the logic implementation shown in FIG. 3. However, in an embodiment in which the hardware chooses to override the software specified value, the reset signal may be applied to the corresponding bit of EOR 80, causing the bit value to be returned to its default value of 0. In such manner, the hardware overrides the software hint. While shown as directly coupled to the bits of EOR 80 and DOR 90, in other embodiments, the reset signal may be coupled to logic, the output of which may be used to override the software hint values. Upon a context switch, the values thus present in the registers as affected by the hardware override of the software hints may be saved.
  • Embodiments of the present invention thus enable dynamic configuration of processor features. Furthermore, in various embodiments of the present invention OS privilege is not needed to enable or disable processor features, allowing fast access, and facilitating fine-grained optimization of application performance with respect to various processor features, such as HWP. Thus user-level hints and/or overrides to processor configurations may be implemented, instead of restricting such management to the OS and BIOS for all processes. Accordingly, an application program can dynamically configure one or more processor features, and do so only for that program. Other processor features can be controlled using software override hints, such as branch prediction logic and the like. For example, branch prediction may be enabled during certain program phases and disabled during other phases. Still other features may include features such as dynamic random access memory (DRAM) page open/close policies, cache allocation policies, bus protocols, other memory protocols, and the like.
  • Referring now to FIG. 5, shown is a flow diagram of a method according to one embodiment of the present invention. As shown in FIG. 5, method 200 may be performed by software to override a desired processor feature. At block 210, the processor feature to be overridden may be selected. Then, in an implementation incorporating one or more override registers, the appropriate override register may be set (block 220). For example, with reference to the implementation of FIG. 3, to enable a processor feature that is disabled by default (e.g., PCR 70 has a corresponding bit value of 0), EOR 80 may have its corresponding bit set to a value of 1.
  • Then in an embodiment in which hardware may choose to override such a software hint value, next it may be determined whether the hardware wants to override the software hint (diamond 230). If so, the appropriate override register (in this case, EOR 80) may be reset (block 240), and control may pass to block 250. If the hardware chooses not to override the software hint, control may pass directly to block 250 from diamond 230.
  • Corresponding bits of the override registers and the processor configuration register then may be applied to a logic implementation (block 250). For example, with reference to FIG. 3, corresponding bits of PCR 70 and EOR 80 may be applied to OR gate 92, the output of which may be applied to AND gate 94 along with the corresponding bit of DOR 90. Based on the output of AND gate 94, the processor configuration setting may be overridden using the output of the logic implementation (block 260).
  • The above operation to override a processor configuration setting may be performed during a first context of a program. For example, the first context may correspond to a software thread, such as a GC phase of a program. Next it may be determined whether a context switch occurs (diamond 270). For example, a context switch may occur at the end of the GC phase of the program. If a context switch does not occur, control may loop back to diamond 270.
  • Upon a context switch, the values in the override registers and the processor configuration register may be saved (block 280). For example, such values may be saved along with the values in other context registers to allow for restoring of the processor state values on a later context switch back to the original context.
  • After the context switch, the override registers and processor configuration register may be loaded with values for the new context (block 290). For example, if such context was previously active, the values saved away with the processor state may be restored.
  • Embodiments may be implemented in a computer program that may be stored on a storage medium having instructions to program a computer system to perform the embodiments. The storage medium may include, but is not limited to, any type of disk including floppy disks, optical disks, 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 and static RAMs, erasable programmable read-only memories (EPROMs), electrically erasable programmable read-only memories (EEPROMs), flash memories, magnetic or optical cards, or any type of media suitable for storing electronic instructions. Other embodiments may be implemented as software modules executed by a programmable control device.
  • Referring now to FIG. 6, shown is a block diagram of a computer system in accordance with one embodiment of the present invention. As shown in FIG. 6, the computer system includes a processor 501. Processor 501 may be coupled over a front-side bus 520 to a memory hub 530 in one embodiment, which may be coupled to a shared main memory 540 via a memory bus. Processor 501 may include various registers in accordance with an embodiment of the present invention. Specifically, as shown in FIG. 6, processor 501 may include a processor configuration register (PCR) 502, an enable override register (EOR) 504, and a disable override register (DOR) 506. The corresponding bits in each of these registers may be coupled to logic 507 which may be used to logically combine the bits to override a processor configuration setting for a given processor feature. As shown in FIG. 6, logic 507 may be coupled to various processor resources, for example, a HWP 508 and a branch prediction unit (BPU) 509. Based on the output of logic 507, these processor features may be enabled or disabled. As discussed above, such feature selection may be under user control.
  • Memory hub 530 may also be coupled (via a hub link) to an input/output (I/O) hub 535 that is coupled to an I/O expansion bus 555 and a peripheral bus 550. In various embodiments, I/O expansion bus 555 may be coupled to various I/O devices such as a keyboard and mouse, among other devices. Peripheral bus 550 may be coupled to various components such as a peripheral device 570 which may be a memory device such as a flash memory, add-in card, and the like. Although the description makes reference to specific components of the system of FIG. 6, numerous modifications of the illustrated embodiments may be possible.
  • 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.

Claims (20)

1. An apparatus comprising:
a processor configuration register to store a default setting for at least one processor feature;
a first register to store first override information to enable the at least one processor feature;
a second register to store second override information to disable the at least one processor feature; and
control logic to couple corresponding entries of the processor configuration register, the first register, and the second register.
2. The apparatus of claim 1, wherein the control logic is to output a control signal to enable or disable the at least one processor feature based on a state of the corresponding entries of the processor configuration register, the first register and the second register.
3. The apparatus of claim 2, wherein the first register and the second register are to store the first override information and the second override information, respectively, under user level software.
4. The apparatus of claim 1, wherein the first register includes a plurality of entries each to store first override information for a processor feature.
5. The apparatus of claim 4, wherein the second register includes a plurality of entries each to store second override information for a processor feature.
6. The apparatus of claim 5, further comprising a hardware unit to generate a reset signal to reset an entry of the first register, wherein the first entry is set by software.
7. The apparatus of claim 1, wherein the first register is to override a disabled default setting of the at least one processor feature and the second register is to override an enabled default setting of the at least one processor feature.
8. A system comprising:
a processor including a first register to store a plurality of first bits, each of the first bits corresponding to a default setting for a processor feature, a second register to store a plurality of second bits, at least one of the second bits to override the corresponding default setting to an enabled state, and a third register to store a plurality of third bits, at least one of the third bits to override the corresponding default setting to a disabled state; and
a memory coupled to the processor.
9. The system of claim 8, wherein the plurality of second bits are to be written by user level software.
10. The system of claim 9, further comprising a hardware prefetcher, wherein the hardware prefetcher is to reset a value of one of the plurality of second bits to override the value written by the user level software.
11. The system of claim 8, wherein the processor includes logic to receive corresponding ones of the first, second and third bits and to generate an output therefrom, wherein the output is to override the default setting for the corresponding processor feature.
12. The system of claim 11, wherein the processor includes a state storage to store the plurality of first bits, the plurality of second bits and the plurality of third bits on a context switch.
13. The system of claim 8, wherein corresponding ones of the first, second and third bits are associated with a processor feature corresponding to a hardware prefetcher.
14. A method comprising:
setting an entry corresponding to a first processor feature in a first override register, the first override register to store entries to override a processor configuration setting for corresponding processor features to an enabled state, and setting an entry corresponding to the first processor feature in a second override register, the second override register to store entries to override a processor configuration setting for corresponding processor features to a disabled state, the processor configuration setting stored in a configuration register; and
overriding the processor configuration setting for the first processor feature using the entries in the first override register and the second override register.
15. The method of claim 14, further comprising setting the entries during a first portion of a program and resetting the entries during a second portion of the program under user control.
16. The method of claim 15, further comprising storing the entries in the first and second override registers upon a context switch from a first context to a second context and loading values associated with a processor state of the second context into the first and second override registers.
17. The method of claim 16, further comprising setting the entries based on dynamic profiling of a program.
18. The method of claim 14, further comprising logically combining the entry of the first override register and the processor configuration setting of the configuration register to obtain a first value and logically combining the first value and the entry of the second override register to override the processor configuration setting.
19. The method of claim 18, further comprising overriding the entry in the first override register under hardware control after setting the entry with user level software.
20. The method of claim 14, wherein overriding the processor configuration setting comprises performing a first logical operation between the entry in the first override register and a corresponding entry in the processor configuration register and performing a second logical operation between the entry in the second override register and a result of the first logical operation.
US11/973,837 2004-10-06 2007-10-10 Overriding processor configuration settings Abandoned US20080046713A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US11/973,837 US20080046713A1 (en) 2004-10-06 2007-10-10 Overriding processor configuration settings

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US10/959,356 US7308571B2 (en) 2004-10-06 2004-10-06 Overriding processor configuration settings
US11/973,837 US20080046713A1 (en) 2004-10-06 2007-10-10 Overriding processor configuration settings

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
US10/959,356 Continuation US7308571B2 (en) 2004-10-06 2004-10-06 Overriding processor configuration settings

Publications (1)

Publication Number Publication Date
US20080046713A1 true US20080046713A1 (en) 2008-02-21

Family

ID=35457593

Family Applications (2)

Application Number Title Priority Date Filing Date
US10/959,356 Expired - Fee Related US7308571B2 (en) 2004-10-06 2004-10-06 Overriding processor configuration settings
US11/973,837 Abandoned US20080046713A1 (en) 2004-10-06 2007-10-10 Overriding processor configuration settings

Family Applications Before (1)

Application Number Title Priority Date Filing Date
US10/959,356 Expired - Fee Related US7308571B2 (en) 2004-10-06 2004-10-06 Overriding processor configuration settings

Country Status (4)

Country Link
US (2) US7308571B2 (en)
JP (3) JP4842956B2 (en)
CN (1) CN100524215C (en)
WO (1) WO2006041758A1 (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7584508B1 (en) * 2008-12-31 2009-09-01 Kaspersky Lab Zao Adaptive security for information devices
US7607174B1 (en) * 2008-12-31 2009-10-20 Kaspersky Lab Zao Adaptive security for portable information devices
US20120324261A1 (en) * 2011-06-16 2012-12-20 Chih-Hung Huang Usb 3.0 host with low power consumption and method for reducing power consumption of a usb 3.0 host
US20130205148A1 (en) * 2012-02-07 2013-08-08 Etron Technology, Inc. Usb 3.0 host with low power consumption and method for reducing power consumption of a usb 3.0 host
US9552478B2 (en) 2010-05-18 2017-01-24 AO Kaspersky Lab Team security for portable information devices
EP3792752A1 (en) * 2019-09-11 2021-03-17 Nokia Solutions and Networks Oy Arithmetic unit

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7308571B2 (en) * 2004-10-06 2007-12-11 Intel Corporation Overriding processor configuration settings
US7653893B2 (en) * 2005-03-04 2010-01-26 Microsoft Corporation Methods and apparatus for implementing checkin policies in source code control systems
WO2006109835A1 (en) * 2005-04-12 2006-10-19 Matsushita Electric Industrial Co., Ltd. Processor
JP4769608B2 (en) * 2006-03-22 2011-09-07 富士通株式会社 Information processing apparatus having start verification function
US7802252B2 (en) * 2007-01-09 2010-09-21 International Business Machines Corporation Method and apparatus for selecting the architecture level to which a processor appears to conform
US8125986B2 (en) * 2007-01-19 2012-02-28 International Business Machines Corporation Method for enabling secure usage of computers using a mechanism lockdown
US8327120B2 (en) 2007-12-29 2012-12-04 Intel Corporation Instructions with floating point control override
US8006082B2 (en) * 2008-09-29 2011-08-23 Intel Corporation Dynamically reconfiguring platform settings
US8935694B2 (en) 2012-01-23 2015-01-13 International Business Machines Corporation System and method for selectively saving and restoring state of branch prediction logic through separate hypervisor-mode and guest-mode and/or user-mode instructions
US9032191B2 (en) 2012-01-23 2015-05-12 International Business Machines Corporation Virtualization support for branch prediction logic enable / disable at hypervisor and guest operating system levels
US9507540B1 (en) 2013-03-14 2016-11-29 Amazon Technologies, Inc. Secure virtual machine memory allocation management via memory usage trust groups
US11669441B1 (en) * 2013-03-14 2023-06-06 Amazon Technologies, Inc. Secure virtual machine reboot via memory allocation recycling
US9916185B2 (en) 2014-03-18 2018-03-13 International Business Machines Corporation Managing processing associated with selected architectural facilities
US9588774B2 (en) 2014-03-18 2017-03-07 International Business Machines Corporation Common boot sequence for control utility able to be initialized in multiple architectures
US9582295B2 (en) 2014-03-18 2017-02-28 International Business Machines Corporation Architectural mode configuration

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5551033A (en) * 1991-05-17 1996-08-27 Zenith Data Systems Corporation Apparatus for maintaining one interrupt mask register in conformity with another in a manner invisible to an executing program
US6112288A (en) * 1998-05-19 2000-08-29 Paracel, Inc. Dynamic configurable system of parallel modules comprising chain of chips comprising parallel pipeline chain of processors with master controller feeding command and data
US6427202B1 (en) * 1999-05-04 2002-07-30 Microchip Technology Incorporated Microcontroller with configurable instruction set
US6466540B1 (en) * 1999-05-05 2002-10-15 International Business Machines Corporation Self-healing coupler for a serial raid device
US6574712B1 (en) * 1999-11-08 2003-06-03 International Business Machines Corporation Software prefetch system and method for predetermining amount of streamed data
US6662274B2 (en) * 2001-06-20 2003-12-09 Intel Corporation Method for using cache prefetch feature to improve garbage collection algorithm
US6675263B2 (en) * 2000-12-29 2004-01-06 Intel Corporation Method and apparatus for filtering prefetches to provide high prefetch accuracy using less hardware
US6675280B2 (en) * 2001-11-30 2004-01-06 Intel Corporation Method and apparatus for identifying candidate virtual addresses in a content-aware prefetcher

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS59127157A (en) * 1982-12-17 1984-07-21 Fujitsu Ltd Test data generator
EP0218335A3 (en) 1985-08-30 1989-03-08 Advanced Micro Devices, Inc. Control store for electronic processor
JPS62259139A (en) * 1986-05-06 1987-11-11 Hitachi Ltd Microprogram controller
US5790834A (en) * 1992-08-31 1998-08-04 Intel Corporation Apparatus and method using an ID instruction to identify a computer microprocessor
US5732207A (en) * 1995-02-28 1998-03-24 Intel Corporation Microprocessor having single poly-silicon EPROM memory for programmably controlling optional features
JP3451595B2 (en) * 1995-06-07 2003-09-29 インターナショナル・ビジネス・マシーンズ・コーポレーション Microprocessor with architectural mode control capable of supporting extension to two distinct instruction set architectures
US5606715A (en) 1995-06-26 1997-02-25 Motorola Inc. Flexible reset configuration of a data processing system and method therefor
US5996084A (en) * 1996-01-17 1999-11-30 Texas Instruments Incorporated Method and apparatus for real-time CPU thermal management and power conservation by adjusting CPU clock frequency in accordance with CPU activity
US5889679A (en) * 1997-07-15 1999-03-30 Integrated Device Technology, Inc. Fuse array control for smart function enable
US5978858A (en) * 1997-09-30 1999-11-02 Compaq Computer Corporation Packet protocol and distributed burst engine
GB9721659D0 (en) * 1997-10-14 1997-12-10 Philips Electronics Nv Space-limited marking structure for tracing garbage collectors
US6615344B1 (en) * 1999-09-03 2003-09-02 Infineon Technologies North America Corp. System and method for tracking selectively enabling modules used in an integrated processor using a tracking register providing configuration information to an external pin
EP1160662A1 (en) * 2000-03-10 2001-12-05 Texas Instruments France Processor with pointer tracking to eliminate redundant memory fetches
US7308571B2 (en) * 2004-10-06 2007-12-11 Intel Corporation Overriding processor configuration settings

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5551033A (en) * 1991-05-17 1996-08-27 Zenith Data Systems Corporation Apparatus for maintaining one interrupt mask register in conformity with another in a manner invisible to an executing program
US6112288A (en) * 1998-05-19 2000-08-29 Paracel, Inc. Dynamic configurable system of parallel modules comprising chain of chips comprising parallel pipeline chain of processors with master controller feeding command and data
US6427202B1 (en) * 1999-05-04 2002-07-30 Microchip Technology Incorporated Microcontroller with configurable instruction set
US6466540B1 (en) * 1999-05-05 2002-10-15 International Business Machines Corporation Self-healing coupler for a serial raid device
US6574712B1 (en) * 1999-11-08 2003-06-03 International Business Machines Corporation Software prefetch system and method for predetermining amount of streamed data
US6675263B2 (en) * 2000-12-29 2004-01-06 Intel Corporation Method and apparatus for filtering prefetches to provide high prefetch accuracy using less hardware
US6662274B2 (en) * 2001-06-20 2003-12-09 Intel Corporation Method for using cache prefetch feature to improve garbage collection algorithm
US6675280B2 (en) * 2001-11-30 2004-01-06 Intel Corporation Method and apparatus for identifying candidate virtual addresses in a content-aware prefetcher

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100138926A1 (en) * 2008-12-02 2010-06-03 Kashchenko Nadezhda V Self-delegating security arrangement for portable information devices
US8370946B2 (en) * 2008-12-02 2013-02-05 Kaspersky Lab Zao Self-delegating security arrangement for portable information devices
US7584508B1 (en) * 2008-12-31 2009-09-01 Kaspersky Lab Zao Adaptive security for information devices
US7607174B1 (en) * 2008-12-31 2009-10-20 Kaspersky Lab Zao Adaptive security for portable information devices
US9552478B2 (en) 2010-05-18 2017-01-24 AO Kaspersky Lab Team security for portable information devices
US20120324261A1 (en) * 2011-06-16 2012-12-20 Chih-Hung Huang Usb 3.0 host with low power consumption and method for reducing power consumption of a usb 3.0 host
US9372528B2 (en) * 2011-06-16 2016-06-21 Eever Technology, Inc. Universal serial bus (USB) 3.0 compatible host with lower operation power consumption and method for reducing operation power consumption of a USB 3.0 compatible host
US20130205148A1 (en) * 2012-02-07 2013-08-08 Etron Technology, Inc. Usb 3.0 host with low power consumption and method for reducing power consumption of a usb 3.0 host
US9367121B2 (en) * 2012-02-07 2016-06-14 Eever Technology, Inc. Universal serial bus USB 3.0 compatible host with lower operation power consumption and method for reducing operation power consumption of a USB compatible 3.0 host
EP3792752A1 (en) * 2019-09-11 2021-03-17 Nokia Solutions and Networks Oy Arithmetic unit

Also Published As

Publication number Publication date
CN100524215C (en) 2009-08-05
CN101031881A (en) 2007-09-05
JP5868436B2 (en) 2016-02-24
JP5634353B2 (en) 2014-12-03
US20060075218A1 (en) 2006-04-06
JP2012014716A (en) 2012-01-19
US7308571B2 (en) 2007-12-11
JP4842956B2 (en) 2011-12-21
WO2006041758A1 (en) 2006-04-20
JP2014112402A (en) 2014-06-19
JP2008516337A (en) 2008-05-15

Similar Documents

Publication Publication Date Title
US20080046713A1 (en) Overriding processor configuration settings
JP6507435B2 (en) Instruction emulation processor, method, and system
US7426648B2 (en) Global and pseudo power state management for multiple processing elements
US9703562B2 (en) Instruction emulation processors, methods, and systems
US5864692A (en) Method and apparatus for protecting memory-mapped devices from side effects of speculative instructions
US8812792B2 (en) Technique for using memory attributes
US20130159679A1 (en) Providing Hint Register Storage For A Processor
US8789169B2 (en) Microcomputer having a protection function in a register
US7318125B2 (en) Runtime selective control of hardware prefetch mechanism
US10482017B2 (en) Processor, method, and system for cache partitioning and control for accurate performance monitoring and optimization
KR100498486B1 (en) Computer system providing for recompiling a program and cxtracting threads dynamically by a thread binary compiler and Simultaneous Multithreading method thereof
CN114692166A (en) Dynamic detection of speculative holes
CN114691316A (en) Autonomous and extensible resource control based on software priority hints
CN114692142A (en) Strengthening branch hardware for speculative vulnerabilities
US6233675B1 (en) Facility to allow fast execution of and, or, and test instructions
JP2021515929A (en) Data processing
KR102560087B1 (en) Method and apparatus for translating memory addresses in manycore system
KR100802686B1 (en) System and method for low overhead boundary checking of java arrays
US6289439B1 (en) Method, device and microprocessor for performing an XOR clear without executing an XOR instruction
US20030101333A1 (en) Data processor
CN114692165A (en) Hardening execution hardware against speculation bugs
CN114692164A (en) Hardening registers to guard against speculation bugs
CN114692140A (en) Hardening storage hardware for speculative vulnerabilities
CN114692139A (en) Reinforcing loading hardware for speculative vulnerabilities
CN111512309A (en) System, method and apparatus for resisting cross-privilege linear probing

Legal Events

Date Code Title Description
STCB Information on status: application discontinuation

Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION