Dynamic selection of memory virtualization techniques

• US 7,752,417 B2
• Filed: 06/05/2006
• Issued: 07/06/2010
• Est. Priority Date: 06/05/2006
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Vick, Christopher A., Paleczny, Michael H., Freeman, Jay R., Gustafson, Phyllis E., Manczak, Olaf
• Primary Examiner(s): Chace; Christian P
• Assistant Examiner(s): DAVIDSON, CHAD
• Application Number:
  US11/447,189
• Publication Number:
  US 20070283125A1
• Time in Patent Office: 1,492 Days
• Field of Search: None
• US Class Current: 711/209
• CPC Class Codes: G06F 11/3419 by assessing time
• CPC Class Codes: G06F 11/3433 for load management allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring
• CPC Class Codes: G06F 12/023 Free address space management
• CPC Class Codes: G06F 12/10 Address translation
• CPC Class Codes: G06F 12/1009 using page tables, e.g. page table structures
• CPC Class Codes: G06F 12/1027 using associative or pseudo-associative address translation means, e.g. translation look-aside buffer [TLB]
• CPC Class Codes: G06F 12/1036 for multiple virtual address spaces, e.g. segmentation G06F12/1045 takes precedence
• CPC Class Codes: G06F 2201/81 Threshold
• CPC Class Codes: G06F 2201/865 Monitoring of software
• CPC Class Codes: G06F 2201/88 Monitoring involving counting

Abstract

A computer system may be configured to dynamically select a memory virtualization and corresponding virtual-to-physical address translation technique during execution of an application and to dynamically employ the selected technique in place of a current technique without re-initializing the application. The computer system may be configured to determine that a current address translation technique incurs a high overhead for the application'"'"'s current workload and may be configured to select a different technique dependent on various performance criteria and/or a user policy. Dynamically employing the selected technique may include reorganizing a memory, reorganizing a translation table, allocating a different block of memory to the application, changing a page or segment size, or moving to or from a page-based, segment-based, or function-based address translation technique. A selected translation technique may be dynamically employed for the application independent of a translation technique employed for a different application.

Claims

1. A computer-implemented method, comprising, during execution of a first application:

receiving, on behalf of said first application, a first memory request directed to a first virtual memory address;

translating said first virtual memory address to a corresponding first physical address according to a first address translation technique;

analyzing performance of said first application;

dynamically selecting a second address translation technique for translating virtual memory addresses to corresponding physical memory addresses on behalf of said first application; and

subsequent to said selecting a second address translation technique, receiving, on behalf of said first application, a second memory request directed to a second virtual memory address, and translating said second virtual memory address to a corresponding second physical address according to said second address translation technique;

wherein said dynamically selecting a second address translation technique is dependent, at least in part, on a user policy and on results of said analyzing; and

wherein said dynamically selecting and said translating according to said second address translation technique are performed transparently to said first application.

2. The method of claim 1, further comprising determining said first address translation technique dependent on a predicted distribution of virtual memory addresses to be accessed by said first application.

3. The method of claim 1, wherein said analyzing performance of said first application comprises:

measuring one or more of;

a number of translation lookaside buffer (TLB) misses, a number of trap handler calls, a memory access time, a number of table look-up operations, a distribution of memory accesses, and a size of a data structure; and

comparing results of said measuring to one or more performance criteria.

4. The method of claim 1, wherein each of said first address translation technique and said second address translation technique comprises one or more of:

page-based translation of virtual memory addresses to physical memory addresses, segment-based translation of virtual memory addresses to physical memory addresses, a table-lookup of a virtual-to-physical address mapping, and function-based translation of virtual memory addresses to physical memory addresses.

5. The method of claim 1, further comprising, during said execution, reorganizing a memory in conjunction with said selecting a second address translation technique, wherein said reorganizing said memory comprises one or more of:

reducing fragmentation of a portion of said memory, reorganizing a page table corresponding to said memory, changing one or more mappings between a virtual address space and a corresponding physical address space of said memory, and allocating a different portion of said memory to said first application.

6. The method of claim 1, wherein said first address translation technique and said second address translation technique differ in one or more of:

a page size, a segment size, and a virtual-to-physical address space mapping.

7. The method of claim 1, further comprising, during said execution:

selecting a third address translation technique for translating virtual addresses to corresponding physical addresses on behalf of a second application;

subsequent to said selecting a third address translation technique, translating a third virtual memory address to a third physical memory address on behalf of said second application according to said third address translation technique;

wherein said third address translation technique is different from one or more of said first address translation technique and said second address translation technique; and

wherein said selecting said third address technique and said translating according to said third address translation technique are performed transparently to said first application.

8. A computer system, comprising:

a processor configured to support a plurality of memory virtualization techniques; and

a memory configured to communicate with said processor;

wherein said processor is further configured to;

determine a first one of said plurality of address translation techniques dependent on a predicted distribution of virtual memory locations to be accessed by a first application;

translate, on behalf of said first application executing on said processor, a memory request directed to a first virtual memory address to a first physical memory address of said memory according to said first one of said plurality of memory virtualization techniques;

during said executing, dynamically select a second one of said plurality of memory virtualization techniques for translating virtual addresses, on behalf of said first application, to corresponding physical addresses of said memory;

during said executing and subsequent to said selecting a second one of said plurality of memory virtualization techniques, translate, on behalf of said first application, a second memory request directed to a virtual memory address to a corresponding physical memory address of said memory according to said second one of said plurality of memory virtualization techniques;

wherein said dynamically selecting and said translating according to said second one of said plurality of memory virtualization techniques are performed transparently to said first application.

9. The computer system of claim 8,

wherein said processor is further configured to analyze performance of said first memory virtualization technique as applied to said first application executing on said processor during said executing;

wherein to analyze performance of said first one of said plurality of memory virtualization techniques, said processor is configured to;

monitor one or more of;

a number of translation lookaside buffer (TLB) misses, a number of trap handler calls, a memory access time, a number of table look-up operations, a distribution of memory accesses, and a size of a data structure; and

compare results of said monitoring to one or more performance criteria; and

wherein said dynamically selecting a second one of said plurality of memory virtualization techniques is dependent on said analyzing.

10. The computer system of claim 8, wherein each of said first one of said plurality of memory virtualization techniques and said second one of said plurality of memory virtualization techniques comprises a different one of:

page-based translation of virtual memory addresses to physical memory addresses, segment-based translation of virtual memory addresses to physical memory addresses, a table-lookup of a virtual-to-physical address mapping, and function-based translation of virtual memory addresses to physical memory addresses.

11. The computer system of claim 8, wherein said processor is further configured to reorganize said memory, during said executing, in conjunction with said selecting a second one of said plurality of address translation techniques, wherein to reorganize said memory, said processor is configured to perform one or more of:

reducing fragmentation of a portion of said memory, reorganizing a page table corresponding to said memory, changing one or more mappings between a virtual address space and a corresponding physical address space of said memory, and allocating a different portion of said memory to said first application.

12. The computer system of claim 8, wherein said first one of said plurality of address translation techniques and said second one of said plurality of address translation techniques differ in one or more of:

a page size, a segment size, and a virtual-to-physical address space mapping.

13. The computer system of claim 8, wherein said dynamically selecting a second one of said plurality of memory virtualization techniques is dependent, at least in part, on a policy specified by a user.

14. A processor, comprising:

an address translator configured to support a plurality of memory virtualization techniques; and

a virtualization technique selector;

wherein said address translator is configured to translate, on behalf of a first application executing on said processor, a memory request directed to a first virtual memory address to a first physical memory address according to a first memory virtualization technique;

wherein said virtualization technique selector is configured to dynamically select a second memory virtualization technique for translating virtual addresses to physical addresses on behalf of said first application during said executing;

wherein said first memory virtualization technique comprises translating virtual addresses to physical addresses using one or more look-up operations on a stored mapping of virtual addresses to physical addresses;

wherein said second memory virtualization technique comprises translating virtual addresses to physical addresses by generating physical addresses from respective virtual addresses using a mathematical or logical function applied to the respective virtual addresses without performing a look-up operation on a stored mapping of virtual addresses to physical addresses;

wherein, during said executing and subsequent to said virtualization technique selector selecting a second memory virtualization technique, said address translator is configured to translate a second memory request directed to a second virtual memory address to a second physical memory address on behalf of said first application according to said second memory virtualization technique; and

wherein said dynamically selecting and said translating according to said second memory virtualization technique are performed transparently to said first application.

15. The processor of claim 14, further comprising a performance analyzer configured to analyze performance of said first memory virtualization technique as applied to said first application executing on said processor during said executing, wherein said dynamically selecting a second memory virtualization technique is dependent on said analyzing.

16. The processor of claim 15, wherein to analyze performance of said first memory virtualization technique, said performance analyzer is configured to:

monitor one or more of;

a number of translation lookaside buffer (TLB) misses, a number of trap handler calls, a memory access time, a number of table look-up operations, a distribution of memory accesses, and a size of data structures accessed; and

compare results of said monitoring to one or more performance criteria.

17. The processor of claim 14, further comprising a memory reorganizer;

wherein in conjunction with said dynamically selecting a second memory virtualization technique said memory reorganizer is configured to perform one or more of;

reducing fragmentation of a portion of said memory, reorganizing a page table corresponding to said memory, changing one or more mappings between a virtual address space and a corresponding physical address space of said memory, and allocating a different portion of said memory to said first application.

18. The processor of claim 14, wherein said first memory virtualization technique and said second memory virtualization technique further differ in a virtual-to-physical address space mapping.

19. The processor of claim 14, wherein said dynamically selecting a second memory virtualization technique is dependent, at least in part, on a policy specified by a user.