US20140013058A1 - Prefetching of next physically sequential cache line after cache line that includes loaded page table entry - Google Patents
Prefetching of next physically sequential cache line after cache line that includes loaded page table entry Download PDFInfo
- Publication number
- US20140013058A1 US20140013058A1 US13/872,527 US201313872527A US2014013058A1 US 20140013058 A1 US20140013058 A1 US 20140013058A1 US 201313872527 A US201313872527 A US 201313872527A US 2014013058 A1 US2014013058 A1 US 2014013058A1
- Authority
- US
- United States
- Prior art keywords
- cache line
- microprocessor
- page table
- translation lookaside
- lookaside buffer
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
- G06F12/0802—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches
- G06F12/0862—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches with prefetch
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
- G06F12/10—Address translation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2212/00—Indexing scheme relating to accessing, addressing or allocation within memory systems or architectures
- G06F2212/60—Details of cache memory
- G06F2212/6028—Prefetching based on hints or prefetch instructions
Definitions
- the present invention relates in general to microprocessors, and more particularly to the prefetching of data thereby.
- page tables that it stores in system memory that are used to translate virtual addresses into physical addresses.
- the page tables may be arranged in a hierarchical fashion, such as according to the well-known scheme employed by x86 architecture processors as described in Chapter 3 of the IA-32 Intel Architecture Software Developer's Manual, Volume 3A: System Programming Guide, Part 1, June 2006, which is hereby incorporated by reference in its entirety for all purposes.
- page tables include page table entries (PTE), each of which stores a physical page address of a physical memory page and attributes of the physical memory page. The process of taking a virtual memory page address and using it to traverse the page table hierarchy to finally obtain the PTE associated with the virtual address in order to translate the virtual address to a physical address is commonly referred to as a tablewalk.
- PTE page table entries
- processors commonly include a translation lookaside buffer (TLB) that caches the virtual to physical address translations.
- TLB translation lookaside buffer
- the size of the TLB is finite, and when a TLB miss occurs, the tablewalk must be incurred. Therefore, what is needed is a way to reduce the time required to perform a page table walk.
- the present invention provides a microprocessor.
- the microprocessor includes a translation lookaside buffer, a request to load a page table entry into the microprocessor generated in response to a miss of a virtual address in the translation lookaside buffer, and a prefetch unit.
- the prefetch unit receives a physical address of a first cache line that includes the requested page table entry and responsively generates a request to prefetch into the microprocessor a second cache line that is the next physically sequential cache line to the first cache line.
- the present invention provides a method for reducing tablewalk time by a microprocessor having a translation lookaside buffer.
- the method includes requesting to load a page table entry into the microprocessor in response to a miss of a virtual address in the translation lookaside buffer.
- the method also includes receiving a physical address of a first cache line that includes the requested page table entry.
- the method also includes generating a request to prefetch into the microprocessor a second cache line that is the next physically sequential cache line to the first cache line in response to receiving the physical address of the first cache line that includes the requested page table entry.
- the present invention provides a microprocessor.
- the microprocessor includes a translation lookaside buffer, a request to load a page table entry into the microprocessor generated in response to a miss of a virtual address in the translation lookaside buffer, and a prefetch unit.
- the prefetch unit receives a physical address of a first cache line that includes the requested page table entry and responsively generates a request to prefetch into the microprocessor a second cache line that is the previous physically sequential cache line to the first cache line.
- the present invention provides a method for reducing tablewalk time by a microprocessor having a translation lookaside buffer.
- the method includes requesting to load a page table entry into the microprocessor in response to a miss of a virtual address in the translation lookaside buffer.
- the method also includes receiving a physical address of a first cache line that includes the requested page table entry.
- the method also includes generating a request to prefetch into the microprocessor a second cache line that is the previous physically sequential cache line to the first cache line in response to receiving the physical address of the first cache line that includes the requested page table entry.
- FIG. 1 is a block diagram illustrating a microprocessor according to the present invention.
- FIG. 2 is a flowchart illustrating steps performed by the microprocessor of FIG. 1 according to the present invention.
- the microprocessor 100 includes The microprocessor 100 is a pipelined microprocessor that includes an instruction cache 102 that provides instructions to an instruction translator 104 , which translates instructions and provides them to an instruction dispatcher 106 .
- the instructions may include memory access instructions (such as load or store instructions) that are provided to a load unit 108 .
- the load unit 108 provides the virtual address 132 specified by a memory access instruction to a translation lookaside buffer (TLB) 116 that does a lookup of the virtual address 132 and provides the translated physical address 144 , if present, back to the load unit 108 . If the virtual address 132 is not present, the TLB 116 generates a miss signal 134 to a tablewalk engine 118 , which is coupled to the load unit 108 and to the TLB 116 .
- TLB translation lookaside buffer
- a prefetch unit 122 and data cache 112 are also coupled to the load unit 108 .
- a bus interface unit 114 is coupled to the data cache 112 .
- the bus interface unit 114 interfaces the microprocessor 100 to a processor bus that is coupled to physical memory 128 of the computer system in which the microprocessor 100 resides.
- the physical memory 128 stores page tables, one of which includes a first cache line of eight PTE 124 at physical address P and a second cache line of eight PTE 126 at physical address P+64 (the size of a cache line being 64 bytes in the embodiment of FIG. 1 ). In the embodiments described, it is assumed that the size of a PTE is eight bytes; thus, each cache line may hold eight PTE.
- FIG. 2 a flowchart illustrating operation of the microprocessor 100 of FIG. 1 to prefetch a next sequential cache line to a cache line implicated by a PTE load according to the present invention is shown.
- Flow begins at block 202 .
- the TLB 116 when the virtual address 132 misses in the TLB 116 , the TLB 116 generates a miss signal 134 to a tablewalk engine 118 .
- the tablewalk engine 118 responsively performs a tablewalk in order to obtain the physical address translation of the virtual address 132 missing in the TLB 116 .
- the tablewalk includes the tablewalk engine 118 generating a PTE load request 136 to the load unit 108 to load a PTE needed to perform the address translation.
- Flow proceeds to block 204 .
- the load unit 108 responsively detects the PTE load request 136 and loads the PTE from its location in the physical memory 128 . Additionally, the load unit 108 signals the prefetch unit 122 via signal 138 that it has seen a PTE load request and provides to the prefetch unit 122 the physical address 138 of the cache line 124 that includes the loaded PTE, which is at physical address P in the example of FIG. 1 . Flow proceeds to block 206 .
- the prefetch unit 122 responsively generates to the load unit 108 a prefetch request 142 .
- the prefetch request 142 requests the load unit 108 to prefetch into the data cache 112 the cache line 126 at physical address P+64, i.e., the next physically sequential cache line 126 after the cache line that includes the PTE loaded at block 204 .
- Flow proceeds to block 208 .
- the load unit 108 uses the prefetch request 142 as a hint to prefetch into the data cache 112 the next physically sequential cache line 126 .
- the load unit 108 attempts to load the cache line 126 into the data cache 112 .
- the load unit 108 may decide to forego the loading of the cache line 126 .
- One circumstance may be a functional requirement, such as that the cache line falls in an architecturally non-cacheable memory region.
- Another circumstance may be a heuristic decision, such as the utilization of resources for performing cache line allocations has risen above some watermark suggesting that the microprocessor 100 should minimize speculative allocations in favor of pursuing non-speculative allocations.
- the load unit 108 decides to load the next sequential cache line 126 from physical memory 128 , the load unit 108 commands the bus interface unit 114 to do so. Flow ends at block 208 .
- the prefetch unit generates a request to prefetch the previous sequential physical cache line, either instead of or in addition to the next sequential physical cache line.
- the prefetch unit generates a request to prefetch the next sequential physical cache line of other levels of a paging information hierarchy, such as a cache line of page descriptor entries (PDE), although one would expect a smaller benefit.
- PDE page descriptor entries
- the sheer volume of the physical memory underlying a single PDE suggests that the reward would be low due to infrequency and the risk would be high due to the relatively lethargic rate at which a program tends to traverse memory, although the access patterns of some programs might see a benefit.
- the prefetch unit generates a request to prefetch the next sequential physical cache line of information within a different page table hierarchy structure than the well-known PDE/PTE hierarchy discussed herein.
- the prefetch unit generates a request to prefetch the next physically sequential cache line to the cache line just read that includes the PTE needed to complete the tablewalk.
- each page table is 4 KB
- each PTE is eight bytes
- each cache line is 64 bytes
- the probability is particularly high if the operating system lays out the page tables physically sequential to one another.
- the prefetch unit generates a request to prefetch eight PTE, which may potentially greatly reduce the number of clock cycles required to perform a page table walk to populate the TLB for the eight memory pages whose physical address is stored in the eight PTE.
- the tablewalk engine 118 when the tablewalk engine 118 has to do a tablewalk that involves a load of any of the eight PTE in the next physically sequential cache line 126 , they will be present in the data cache 112 (unless they are subsequently evicted from the data cache 112 ), which will eliminate the need for the long latency read of physical memory 128 to get the PTE.
- Prefetch schemes which attempt to detect memory access patterns of program memory accesses, i.e., program instruction loads and stores. If the prefetcher detects the program accessing memory in a pattern, the prefetcher will attempt to anticipate the addresses of future program loads/stores and prefetch from them. If the program is accessing memory sequentially, the prefetcher may prefetch the next sequential cache line, typically based on the virtual address of the load/store. In a processor architecture for which the operating system performs tablewalks, the program load/store-based prefetcher may prefetch the next sequential cache line after a PTE load.
- the program load/store-based prefetcher will not trigger off a PTE load, since it is not a program load, and will therefore not prefetch the next sequential cache line after a PTE load.
- our prefetch unit triggers off of a non-program PTE load, which is a physical memory access generated by our tablewalk engine.
- the prefetch unit generates requests to advantageously prefetch the next physically sequential cache line, which likely contains the next few PTE in the page table.
- software can enable, for example, the function, fabrication, modeling, simulation, description and/or testing of the apparatus and methods described herein. This can be accomplished through the use of general programming languages (e.g., C, C++), hardware description languages (HDL) including Verilog HDL, VHDL, and so on, or other available programs.
- general programming languages e.g., C, C++
- HDL hardware description languages
- Verilog HDL Verilog HDL
- VHDL Verilog HDL
- VHDL Verilog HDL
- Such software can be disposed in any known computer usable medium such as semiconductor, magnetic disk, or optical disc (e.g., CD-ROM, DVD-ROM, etc.).
- Embodiments of the apparatus and method described herein may be included in a semiconductor intellectual property core, such as a microprocessor core (e.g., embodied in HDL) and transformed to hardware in the production of integrated circuits. Additionally, the apparatus and methods described herein may be embodied as a combination of hardware and software. Thus, the present invention should not be limited by any of the exemplary embodiments described herein, but should be defined only in accordance with the following claims and their equivalents. Specifically, the present invention may be implemented within a microprocessor device which may be used in a general purpose computer. Finally, those skilled in the art should appreciate that they can readily use the disclosed conception and specific embodiments as a basis for designing or modifying other structures for carrying out the same purposes of the present invention without departing from the scope of the invention as defined by the appended claims.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Memory System Of A Hierarchy Structure (AREA)
Abstract
A microprocessor includes a translation lookaside buffer, a request to load a page table entry into the microprocessor generated in response to a miss of a virtual address in the translation lookaside buffer, and a prefetch unit. The prefetch unit receives a physical address of a first cache line that includes the requested page table entry and responsively generates a request to prefetch into the microprocessor a second cache line that is the next physically sequential cache line to the first cache line.
Description
- This application is a continuation of co-pending U.S. patent application Ser. No. 13/412,934, filed Mar. 06, 2012, which is hereby incorporated by reference in its entirety. U.S. patent application Ser. No. 13/412,934 is a continuation of co-pending U.S. patent application Ser. No. 12/604,998, filed Oct. 23, 2009, which is hereby incorporated by reference in its entirety and which claims priority based on U.S. Provisional Application Ser. No. 61/164,588, filed Mar. 30, 2009, which is hereby incorporated by reference in its entirety. U.S. patent application Ser. No. 12/604,998 has issued as U.S. Pat. No. 8,161,246.
- The present invention relates in general to microprocessors, and more particularly to the prefetching of data thereby.
- Many modern microprocessors include virtual memory capability, and in particular, a memory paging mechanism. As is well known in the art, the operating system creates page tables that it stores in system memory that are used to translate virtual addresses into physical addresses. The page tables may be arranged in a hierarchical fashion, such as according to the well-known scheme employed by x86 architecture processors as described in Chapter 3 of the IA-32 Intel Architecture Software Developer's Manual, Volume 3A: System Programming Guide, Part 1, June 2006, which is hereby incorporated by reference in its entirety for all purposes. In particular, page tables include page table entries (PTE), each of which stores a physical page address of a physical memory page and attributes of the physical memory page. The process of taking a virtual memory page address and using it to traverse the page table hierarchy to finally obtain the PTE associated with the virtual address in order to translate the virtual address to a physical address is commonly referred to as a tablewalk.
- Because the latency of a physical memory access is relatively slow, the tablewalk is a relatively costly operation since it involves potentially multiple accesses to physical memory. To avoid incurring the time associated with a tablewalk, processors commonly include a translation lookaside buffer (TLB) that caches the virtual to physical address translations. However, the size of the TLB is finite, and when a TLB miss occurs, the tablewalk must be incurred. Therefore, what is needed is a way to reduce the time required to perform a page table walk.
- In one aspect, the present invention provides a microprocessor. The microprocessor includes a translation lookaside buffer, a request to load a page table entry into the microprocessor generated in response to a miss of a virtual address in the translation lookaside buffer, and a prefetch unit. The prefetch unit receives a physical address of a first cache line that includes the requested page table entry and responsively generates a request to prefetch into the microprocessor a second cache line that is the next physically sequential cache line to the first cache line.
- In another aspect, the present invention provides a method for reducing tablewalk time by a microprocessor having a translation lookaside buffer. The method includes requesting to load a page table entry into the microprocessor in response to a miss of a virtual address in the translation lookaside buffer. The method also includes receiving a physical address of a first cache line that includes the requested page table entry. The method also includes generating a request to prefetch into the microprocessor a second cache line that is the next physically sequential cache line to the first cache line in response to receiving the physical address of the first cache line that includes the requested page table entry.
- In yet another aspect, the present invention provides a microprocessor. The microprocessor includes a translation lookaside buffer, a request to load a page table entry into the microprocessor generated in response to a miss of a virtual address in the translation lookaside buffer, and a prefetch unit. The prefetch unit receives a physical address of a first cache line that includes the requested page table entry and responsively generates a request to prefetch into the microprocessor a second cache line that is the previous physically sequential cache line to the first cache line.
- In yet another aspect, the present invention provides a method for reducing tablewalk time by a microprocessor having a translation lookaside buffer. The method includes requesting to load a page table entry into the microprocessor in response to a miss of a virtual address in the translation lookaside buffer. The method also includes receiving a physical address of a first cache line that includes the requested page table entry. The method also includes generating a request to prefetch into the microprocessor a second cache line that is the previous physically sequential cache line to the first cache line in response to receiving the physical address of the first cache line that includes the requested page table entry.
-
FIG. 1 is a block diagram illustrating a microprocessor according to the present invention. -
FIG. 2 is a flowchart illustrating steps performed by the microprocessor ofFIG. 1 according to the present invention. - Referring now to
FIG. 1 , a block diagram illustrating amicroprocessor 100 according to the present invention is shown. Themicroprocessor 100 includes Themicroprocessor 100 is a pipelined microprocessor that includes aninstruction cache 102 that provides instructions to aninstruction translator 104, which translates instructions and provides them to aninstruction dispatcher 106. The instructions may include memory access instructions (such as load or store instructions) that are provided to aload unit 108. Theload unit 108 provides thevirtual address 132 specified by a memory access instruction to a translation lookaside buffer (TLB) 116 that does a lookup of thevirtual address 132 and provides the translatedphysical address 144, if present, back to theload unit 108. If thevirtual address 132 is not present, theTLB 116 generates amiss signal 134 to atablewalk engine 118, which is coupled to theload unit 108 and to theTLB 116. - A
prefetch unit 122 anddata cache 112 are also coupled to theload unit 108. Finally, a bus interface unit 114 is coupled to thedata cache 112. The bus interface unit 114 interfaces themicroprocessor 100 to a processor bus that is coupled tophysical memory 128 of the computer system in which themicroprocessor 100 resides. Thephysical memory 128, among other things, stores page tables, one of which includes a first cache line of eightPTE 124 at physical address P and a second cache line of eightPTE 126 at physical address P+64 (the size of a cache line being 64 bytes in the embodiment ofFIG. 1 ). In the embodiments described, it is assumed that the size of a PTE is eight bytes; thus, each cache line may hold eight PTE. - Referring now to
FIG. 2 , a flowchart illustrating operation of themicroprocessor 100 ofFIG. 1 to prefetch a next sequential cache line to a cache line implicated by a PTE load according to the present invention is shown. Flow begins atblock 202. - At
block 202, when thevirtual address 132 misses in the TLB 116, the TLB 116 generates amiss signal 134 to atablewalk engine 118. Thetablewalk engine 118 responsively performs a tablewalk in order to obtain the physical address translation of thevirtual address 132 missing in theTLB 116. The tablewalk includes thetablewalk engine 118 generating aPTE load request 136 to theload unit 108 to load a PTE needed to perform the address translation. Flow proceeds to block 204. - At
block 204, theload unit 108 responsively detects thePTE load request 136 and loads the PTE from its location in thephysical memory 128. Additionally, theload unit 108 signals theprefetch unit 122 viasignal 138 that it has seen a PTE load request and provides to theprefetch unit 122 thephysical address 138 of thecache line 124 that includes the loaded PTE, which is at physical address P in the example ofFIG. 1 . Flow proceeds to block 206. - At
block 206, theprefetch unit 122 responsively generates to the load unit 108 aprefetch request 142. The prefetch request 142 requests theload unit 108 to prefetch into thedata cache 112 thecache line 126 at physical address P+64, i.e., the next physicallysequential cache line 126 after the cache line that includes the PTE loaded atblock 204. Flow proceeds to block 208. - At block 208, the
load unit 108 uses theprefetch request 142 as a hint to prefetch into thedata cache 112 the next physicallysequential cache line 126. In particular, theload unit 108 attempts to load thecache line 126 into thedata cache 112. However, under some circumstances present within themicroprocessor 100, theload unit 108 may decide to forego the loading of thecache line 126. One circumstance may be a functional requirement, such as that the cache line falls in an architecturally non-cacheable memory region. Another circumstance may be a heuristic decision, such as the utilization of resources for performing cache line allocations has risen above some watermark suggesting that themicroprocessor 100 should minimize speculative allocations in favor of pursuing non-speculative allocations. If theload unit 108 decides to load the nextsequential cache line 126 fromphysical memory 128, theload unit 108 commands the bus interface unit 114 to do so. Flow ends at block 208. - Although the embodiments described have prefetched the next sequential physical cache line, in alternate embodiments, the prefetch unit generates a request to prefetch the previous sequential physical cache line, either instead of or in addition to the next sequential physical cache line. These embodiments are advantageous if the program is proceeding through memory pages in the other direction.
- Additionally, although the embodiments described have prefetched the next sequential physical cache line of PTE, embodiments are contemplated in which the prefetch unit generates a request to prefetch the next sequential physical cache line of other levels of a paging information hierarchy, such as a cache line of page descriptor entries (PDE), although one would expect a smaller benefit. The sheer volume of the physical memory underlying a single PDE suggests that the reward would be low due to infrequency and the risk would be high due to the relatively lethargic rate at which a program tends to traverse memory, although the access patterns of some programs might see a benefit. Still further, embodiments are contemplated in which the prefetch unit generates a request to prefetch the next sequential physical cache line of information within a different page table hierarchy structure than the well-known PDE/PTE hierarchy discussed herein.
- As described above, the prefetch unit generates a request to prefetch the next physically sequential cache line to the cache line just read that includes the PTE needed to complete the tablewalk. Assuming each page table is 4 KB, each PTE is eight bytes, and each cache line is 64 bytes, then there are 64 cache lines each having eight PTE in a page table; thus, it is highly probable that the next physically sequential cache line prefetched at block 208 contains the next eight PTE in the page table. The probability is particularly high if the operating system lays out the page tables physically sequential to one another.
- This is advantageous because there is a relatively high probability, particularly with small pages (typically 4 KB), that the program will eventually access at least some of the next eight pages of memory beyond the current page that was virtually accessed to cause the TLB miss at
block 202. For a relatively modest amount of additional logic added to the prefetch unit and load unit, in one embodiment the prefetch unit generates a request to prefetch eight PTE, which may potentially greatly reduce the number of clock cycles required to perform a page table walk to populate the TLB for the eight memory pages whose physical address is stored in the eight PTE. Specifically, when thetablewalk engine 118 has to do a tablewalk that involves a load of any of the eight PTE in the next physicallysequential cache line 126, they will be present in the data cache 112 (unless they are subsequently evicted from the data cache 112), which will eliminate the need for the long latency read ofphysical memory 128 to get the PTE. - Prefetch schemes exist which attempt to detect memory access patterns of program memory accesses, i.e., program instruction loads and stores. If the prefetcher detects the program accessing memory in a pattern, the prefetcher will attempt to anticipate the addresses of future program loads/stores and prefetch from them. If the program is accessing memory sequentially, the prefetcher may prefetch the next sequential cache line, typically based on the virtual address of the load/store. In a processor architecture for which the operating system performs tablewalks, the program load/store-based prefetcher may prefetch the next sequential cache line after a PTE load. However, in processors that perform the tablewalks in hardware, rather than using program load/stores, the program load/store-based prefetcher will not trigger off a PTE load, since it is not a program load, and will therefore not prefetch the next sequential cache line after a PTE load. Advantageously, in our hardware-based tablewalk microprocessor, our prefetch unit triggers off of a non-program PTE load, which is a physical memory access generated by our tablewalk engine. Thus, unlike the program load/store-based schemes, the prefetch unit generates requests to advantageously prefetch the next physically sequential cache line, which likely contains the next few PTE in the page table.
- While various embodiments of the present invention have been described herein, it should be understood that they have been presented by way of example, and not limitation. It will be apparent to persons skilled in the relevant computer arts that various changes in form and detail can be made therein without departing from the scope of the invention. For example, software can enable, for example, the function, fabrication, modeling, simulation, description and/or testing of the apparatus and methods described herein. This can be accomplished through the use of general programming languages (e.g., C, C++), hardware description languages (HDL) including Verilog HDL, VHDL, and so on, or other available programs. Such software can be disposed in any known computer usable medium such as semiconductor, magnetic disk, or optical disc (e.g., CD-ROM, DVD-ROM, etc.). Embodiments of the apparatus and method described herein may be included in a semiconductor intellectual property core, such as a microprocessor core (e.g., embodied in HDL) and transformed to hardware in the production of integrated circuits. Additionally, the apparatus and methods described herein may be embodied as a combination of hardware and software. Thus, the present invention should not be limited by any of the exemplary embodiments described herein, but should be defined only in accordance with the following claims and their equivalents. Specifically, the present invention may be implemented within a microprocessor device which may be used in a general purpose computer. Finally, those skilled in the art should appreciate that they can readily use the disclosed conception and specific embodiments as a basis for designing or modifying other structures for carrying out the same purposes of the present invention without departing from the scope of the invention as defined by the appended claims.
Claims (20)
1: A microprocessor, comprising:
a translation lookaside buffer;
a request to load a page table entry into the microprocessor generated in response to a miss of a virtual address in the translation lookaside buffer; and
a prefetch unit, configured to receive a physical address of a first cache line that includes the requested page table entry, wherein the prefetch unit is further configured to responsively generate a request to prefetch into the microprocessor a second cache line that is the next physically sequential cache line to the first cache line.
2: The microprocessor of claim 1 , wherein the second cache line is absent the page table entry implicated by the virtual address that missed in the translation lookaside buffer.
3: The microprocessor of claim 1 , further comprising:
a tablewalk engine, configured to perform a tablewalk in response to the miss of the virtual address in the translation lookaside buffer, wherein the tablewalk does not access any of the page table entries in the prefetched second cache line.
4. (canceled)
5: The microprocessor of claim 1 , wherein the request to load the page table entry, in response to which the prefetch unit prefetches the second cache line, is generated internally by the microprocessor in response to the miss of the virtual address in the translation lookaside buffer, rather than the load request being a load request of a program executed by the microprocessor.
6: The microprocessor of claim 1 , further comprising:
a request to load a page directory entry into the microprocessor generated in response to the miss of the virtual address in the translation lookaside buffer;
wherein the prefetch unit is further configured to receive a physical address of a third cache line that includes the requested page directory entry, wherein the prefetch unit is further configured to responsively generate a second request to prefetch into the microprocessor a fourth cache line that is the next physically sequential cache line to the third cache line.
7: A method for reducing tablewalk time by a microprocessor having a translation lookaside buffer, the method comprising:
requesting to load a page table entry into the microprocessor, in response to a miss of a virtual address in the translation lookaside buffer;
receiving a physical address of a first cache line that includes the requested page table entry;
generating a request to prefetch into the microprocessor a second cache line that is the next physically sequential cache line to the first cache line, in response to said receiving the physical address of the first cache line that includes the requested page table entry.
8: The method of claim 7 , wherein the second cache line is absent the page table entry implicated by the virtual address that missed in the translation lookaside buffer.
9: The method of claim 7 , further comprising:
performing a tablewalk in response to the miss of the virtual address in the translation lookaside buffer, wherein the tablewalk does not access any of the page table entries in the prefetched second cache line.
10. (canceled)
11: A microprocessor, comprising:
a translation lookaside buffer;
a request to load a page table entry into the microprocessor generated in response to a miss of a virtual address in the translation lookaside buffer; and
a prefetch unit, configured to receive a physical address of a first cache line that includes the requested page table entry, wherein the prefetch unit is further configured to responsively generate a request to prefetch into the microprocessor a second cache line that is the previous physically sequential cache line to the first cache line.
12: The microprocessor of claim 11 , wherein the second cache line is absent the page table entry implicated by the virtual address that missed in the translation lookaside buffer.
13: The microprocessor of claim 11 , further comprising:
a tablewalk engine, configured to perform a tablewalk in response to the miss of the virtual address in the translation lookaside buffer, wherein the tablewalk does not access any of the page table entries in the prefetched second cache line.
14. (canceled)
15: The microprocessor of claim 11 , wherein the request to load the page table entry, in response to which the prefetch unit prefetches the second cache line, is generated internally by the microprocessor in response to the miss of the virtual address in the translation lookaside buffer, rather than the load request being a load request of a program executed by the microprocessor.
16: The microprocessor of claim 11 , further comprising:
a request to load a page directory entry into the microprocessor generated in response to the miss of the virtual address in the translation lookaside buffer;
wherein the prefetch unit is further configured to receive a physical address of a third cache line that includes the requested page directory entry, wherein the prefetch unit is further configured to responsively generate a second request to prefetch into the microprocessor a fourth cache line that is the previous physically sequential cache line to the third cache line.
17: A method for reducing tablewalk time by a microprocessor having a translation lookaside buffer, the method comprising:
requesting to load a page table entry into the microprocessor, in response to a miss of a virtual address in the translation lookaside buffer;
receiving a physical address of a first cache line that includes the requested page table entry;
generating a request to prefetch into the microprocessor a second cache line that is the previous physically sequential cache line to the first cache line, in response to said receiving the physical address of the first cache line that includes the requested page table entry.
18: The method of claim 17 , wherein the second cache line is absent the page table entry implicated by the virtual address that missed in the translation lookaside buffer.
19: The method of claim 17 , further comprising:
performing a tablewalk in response to the miss of the virtual address in the translation lookaside buffer, wherein the tablewalk does not access any of the page table entries in the prefetched second cache line.
20. (canceled)
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/872,527 US20140013058A1 (en) | 2009-03-30 | 2013-04-29 | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry |
US14/790,467 US9569363B2 (en) | 2009-03-30 | 2015-07-02 | Selective prefetching of physically sequential cache line to cache line that includes loaded page table entry |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16458809P | 2009-03-30 | 2009-03-30 | |
US12/604,998 US8161246B2 (en) | 2009-03-30 | 2009-10-23 | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry |
US13/412,934 US8433853B2 (en) | 2009-03-30 | 2012-03-06 | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry |
US13/872,527 US20140013058A1 (en) | 2009-03-30 | 2013-04-29 | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry |
Related Parent Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US13/412,934 Continuation US8433853B2 (en) | 2009-03-30 | 2012-03-06 | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/790,467 Continuation-In-Part US9569363B2 (en) | 2009-03-30 | 2015-07-02 | Selective prefetching of physically sequential cache line to cache line that includes loaded page table entry |
Publications (1)
Publication Number | Publication Date |
---|---|
US20140013058A1 true US20140013058A1 (en) | 2014-01-09 |
Family
ID=42785701
Family Applications (3)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US12/604,998 Active 2030-11-23 US8161246B2 (en) | 2009-03-30 | 2009-10-23 | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry |
US13/412,934 Active US8433853B2 (en) | 2009-03-30 | 2012-03-06 | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry |
US13/872,527 Abandoned US20140013058A1 (en) | 2009-03-30 | 2013-04-29 | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry |
Family Applications Before (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US12/604,998 Active 2030-11-23 US8161246B2 (en) | 2009-03-30 | 2009-10-23 | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry |
US13/412,934 Active US8433853B2 (en) | 2009-03-30 | 2012-03-06 | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry |
Country Status (3)
Country | Link |
---|---|
US (3) | US8161246B2 (en) |
CN (2) | CN102999440B (en) |
TW (2) | TWI451334B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130262779A1 (en) * | 2012-03-30 | 2013-10-03 | Jayaram Bobba | Profile-based hardware prefetching |
US9569363B2 (en) | 2009-03-30 | 2017-02-14 | Via Technologies, Inc. | Selective prefetching of physically sequential cache line to cache line that includes loaded page table entry |
US9880842B2 (en) | 2013-03-15 | 2018-01-30 | Intel Corporation | Using control flow data structures to direct and track instruction execution |
US9904569B2 (en) | 2015-10-28 | 2018-02-27 | International Business Machines Corporation | Pre-loading page table cache lines of a virtual machine |
WO2021184141A1 (en) * | 2020-03-15 | 2021-09-23 | Micron Technology, Inc. | Pre-load techniques for improved sequential read |
Families Citing this family (34)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8417895B1 (en) | 2008-09-30 | 2013-04-09 | Violin Memory Inc. | System for maintaining coherency during offline changes to storage media |
US8838850B2 (en) * | 2008-11-17 | 2014-09-16 | Violin Memory, Inc. | Cluster control protocol |
US8442059B1 (en) | 2008-09-30 | 2013-05-14 | Gridiron Systems, Inc. | Storage proxy with virtual ports configuration |
US8443150B1 (en) | 2008-11-04 | 2013-05-14 | Violin Memory Inc. | Efficient reloading of data into cache resource |
US8788758B1 (en) | 2008-11-04 | 2014-07-22 | Violin Memory Inc | Least profitability used caching scheme |
US8650362B2 (en) | 2009-04-17 | 2014-02-11 | Violin Memory Inc. | System for increasing utilization of storage media |
US8667366B1 (en) | 2009-04-17 | 2014-03-04 | Violin Memory, Inc. | Efficient use of physical address space for data overflow and validation |
US8713252B1 (en) | 2009-05-06 | 2014-04-29 | Violin Memory, Inc. | Transactional consistency scheme |
US8402198B1 (en) | 2009-06-03 | 2013-03-19 | Violin Memory, Inc. | Mapping engine for a storage device |
US9069676B2 (en) | 2009-06-03 | 2015-06-30 | Violin Memory, Inc. | Mapping engine for a storage device |
US8402246B1 (en) | 2009-08-28 | 2013-03-19 | Violin Memory, Inc. | Alignment adjustment in a tiered storage system |
US9418011B2 (en) * | 2010-06-23 | 2016-08-16 | Intel Corporation | Region based technique for accurately predicting memory accesses |
US8959288B1 (en) | 2010-07-29 | 2015-02-17 | Violin Memory, Inc. | Identifying invalid cache data |
US8832384B1 (en) * | 2010-07-29 | 2014-09-09 | Violin Memory, Inc. | Reassembling abstracted memory accesses for prefetching |
US8972689B1 (en) | 2011-02-02 | 2015-03-03 | Violin Memory, Inc. | Apparatus, method and system for using real-time performance feedback for modeling and improving access to solid state media |
US8635416B1 (en) | 2011-03-02 | 2014-01-21 | Violin Memory Inc. | Apparatus, method and system for using shadow drives for alternative drive commands |
WO2013095401A1 (en) * | 2011-12-20 | 2013-06-27 | Intel Corporation | System and method for out-of-order prefetch instructions in an in-order pipeline |
CN102722451B (en) * | 2012-06-25 | 2015-04-15 | 杭州中天微***有限公司 | Device for accessing cache by predicting physical address |
US20140108766A1 (en) * | 2012-10-17 | 2014-04-17 | Advanced Micro Devices, Inc. | Prefetching tablewalk address translations |
US9201806B2 (en) | 2013-01-04 | 2015-12-01 | International Business Machines Corporation | Anticipatorily loading a page of memory |
US9251083B2 (en) | 2013-03-11 | 2016-02-02 | Via Technologies, Inc. | Communicating prefetchers in a microprocessor |
US9483406B2 (en) | 2013-03-11 | 2016-11-01 | Via Technologies, Inc. | Communicating prefetchers that throttle one another |
KR102069273B1 (en) | 2013-03-11 | 2020-01-22 | 삼성전자주식회사 | System on chip and operating method thereof |
CN104424117B (en) * | 2013-08-20 | 2017-09-05 | 华为技术有限公司 | Internal memory physics address inquiring method and device |
US9645934B2 (en) * | 2013-09-13 | 2017-05-09 | Samsung Electronics Co., Ltd. | System-on-chip and address translation method thereof using a translation lookaside buffer and a prefetch buffer |
GB2528842B (en) * | 2014-07-29 | 2021-06-02 | Advanced Risc Mach Ltd | A data processing apparatus, and a method of handling address translation within a data processing apparatus |
TWI590053B (en) * | 2015-07-02 | 2017-07-01 | 威盛電子股份有限公司 | Selective prefetching of physically sequential cache line to cache line that includes loaded page table |
US10175987B2 (en) | 2016-03-17 | 2019-01-08 | International Business Machines Corporation | Instruction prefetching in a computer processor using a prefetch prediction vector |
US10386933B2 (en) | 2016-08-30 | 2019-08-20 | International Business Machines Corporation | Controlling navigation of a visual aid during a presentation |
CN111198827B (en) * | 2018-11-16 | 2022-10-28 | 展讯通信(上海)有限公司 | Page table prefetching method and device |
US10936281B2 (en) | 2018-12-19 | 2021-03-02 | International Business Machines Corporation | Automatic slide page progression based on verbal and visual cues |
US10909045B2 (en) * | 2018-12-20 | 2021-02-02 | Arm Limited | System, method and apparatus for fine granularity access protection |
KR20220135560A (en) * | 2021-03-30 | 2022-10-07 | 삼성전자주식회사 | Electronic device for managing memory and operating method thereof |
CN114218132B (en) * | 2021-12-14 | 2023-03-24 | 海光信息技术股份有限公司 | Information prefetching method, processor and electronic equipment |
Family Cites Families (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5666509A (en) * | 1994-03-24 | 1997-09-09 | Motorola, Inc. | Data processing system for performing either a precise memory access or an imprecise memory access based upon a logical address value and method thereof |
US5613083A (en) * | 1994-09-30 | 1997-03-18 | Intel Corporation | Translation lookaside buffer that is non-blocking in response to a miss for use within a microprocessor capable of processing speculative instructions |
US5963984A (en) * | 1994-11-08 | 1999-10-05 | National Semiconductor Corporation | Address translation unit employing programmable page size |
US5752274A (en) * | 1994-11-08 | 1998-05-12 | Cyrix Corporation | Address translation unit employing a victim TLB |
CA2280441A1 (en) * | 1998-09-16 | 2000-03-16 | Morton International, Inc. | Combination of an organothio compound and a zinc mercapto ester as heat stabilizer in pvc processing |
US6681311B2 (en) * | 2001-07-18 | 2004-01-20 | Ip-First, Llc | Translation lookaside buffer that caches memory type information |
US6832296B2 (en) * | 2002-04-09 | 2004-12-14 | Ip-First, Llc | Microprocessor with repeat prefetch instruction |
US7194582B1 (en) * | 2003-05-30 | 2007-03-20 | Mips Technologies, Inc. | Microprocessor with improved data stream prefetching |
US7099999B2 (en) * | 2003-09-30 | 2006-08-29 | International Business Machines Corporation | Apparatus and method for pre-fetching data to cached memory using persistent historical page table data |
WO2005091146A1 (en) * | 2004-03-24 | 2005-09-29 | Matsushita Electric Industrial Co., Ltd. | Cache memory and control method thereof |
US7383418B2 (en) * | 2004-09-01 | 2008-06-03 | Intel Corporation | Method and apparatus for prefetching data to a lower level cache memory |
US20060136696A1 (en) * | 2004-12-16 | 2006-06-22 | Grayson Brian C | Method and apparatus for address translation |
US7383391B2 (en) * | 2005-05-18 | 2008-06-03 | International Business Machines Corporation | Prefetch mechanism based on page table attributes |
US7409524B2 (en) * | 2005-08-17 | 2008-08-05 | Hewlett-Packard Development Company, L.P. | System and method for responding to TLB misses |
US7822941B2 (en) * | 2006-06-05 | 2010-10-26 | Oracle America, Inc. | Function-based virtual-to-physical address translation |
US7949834B2 (en) * | 2007-01-24 | 2011-05-24 | Qualcomm Incorporated | Method and apparatus for setting cache policies in a processor |
US8103832B2 (en) * | 2007-06-26 | 2012-01-24 | International Business Machines Corporation | Method and apparatus of prefetching streams of varying prefetch depth |
US8078827B2 (en) * | 2007-07-05 | 2011-12-13 | International Business Machines Corporation | Method and apparatus for caching of page translations for virtual machines |
US7793070B2 (en) * | 2007-07-12 | 2010-09-07 | Qnx Software Systems Gmbh & Co. Kg | Processing system implementing multiple page size memory organization with multiple translation lookaside buffers having differing characteristics |
US7958316B2 (en) * | 2008-02-01 | 2011-06-07 | International Business Machines Corporation | Dynamic adjustment of prefetch stream priority |
US7996650B2 (en) * | 2008-07-14 | 2011-08-09 | Via Technologies, Inc. | Microprocessor that performs speculative tablewalks |
US7958317B2 (en) * | 2008-08-04 | 2011-06-07 | International Business Machines Corporation | Cache directed sequential prefetch |
US8291202B2 (en) * | 2008-08-08 | 2012-10-16 | Qualcomm Incorporated | Apparatus and methods for speculative interrupt vector prefetching |
-
2009
- 2009-10-23 US US12/604,998 patent/US8161246B2/en active Active
-
2010
- 2010-03-12 TW TW101143382A patent/TWI451334B/en active
- 2010-03-12 TW TW99107221A patent/TWI437490B/en active
- 2010-03-23 CN CN201210460154.XA patent/CN102999440B/en active Active
- 2010-03-23 CN CN 201010151433 patent/CN101833515B/en active Active
-
2012
- 2012-03-06 US US13/412,934 patent/US8433853B2/en active Active
-
2013
- 2013-04-29 US US13/872,527 patent/US20140013058A1/en not_active Abandoned
Non-Patent Citations (1)
Title |
---|
Csaba Andras Moritz, An Introduction to Prefetching, November, 2007, C A Moritz, Yao, Guo, and SSA group; Software System & Architecture Lab, Electrical & Computer Engineering * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9569363B2 (en) | 2009-03-30 | 2017-02-14 | Via Technologies, Inc. | Selective prefetching of physically sequential cache line to cache line that includes loaded page table entry |
US20130262779A1 (en) * | 2012-03-30 | 2013-10-03 | Jayaram Bobba | Profile-based hardware prefetching |
US9880842B2 (en) | 2013-03-15 | 2018-01-30 | Intel Corporation | Using control flow data structures to direct and track instruction execution |
US9904569B2 (en) | 2015-10-28 | 2018-02-27 | International Business Machines Corporation | Pre-loading page table cache lines of a virtual machine |
US9910780B2 (en) | 2015-10-28 | 2018-03-06 | International Business Machines Corporation | Pre-loading page table cache lines of a virtual machine |
WO2021184141A1 (en) * | 2020-03-15 | 2021-09-23 | Micron Technology, Inc. | Pre-load techniques for improved sequential read |
US11868245B2 (en) | 2020-03-15 | 2024-01-09 | Micron Technology, Inc. | Pre-load techniques for improved sequential memory access in a memory device |
Also Published As
Publication number | Publication date |
---|---|
US8433853B2 (en) | 2013-04-30 |
US8161246B2 (en) | 2012-04-17 |
CN101833515A (en) | 2010-09-15 |
TW201312461A (en) | 2013-03-16 |
CN101833515B (en) | 2013-01-02 |
TW201035867A (en) | 2010-10-01 |
TWI437490B (en) | 2014-05-11 |
US20100250859A1 (en) | 2010-09-30 |
CN102999440A (en) | 2013-03-27 |
CN102999440B (en) | 2016-08-24 |
TWI451334B (en) | 2014-09-01 |
US20120198176A1 (en) | 2012-08-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8433853B2 (en) | Prefetching of next physically sequential cache line after cache line that includes loaded page table entry | |
KR101623514B1 (en) | Iommu architected tlb support | |
KR102448124B1 (en) | Cache accessed using virtual addresses | |
CN106537362B (en) | Data processing apparatus and method of processing address conversion in data processing apparatus | |
US9569363B2 (en) | Selective prefetching of physically sequential cache line to cache line that includes loaded page table entry | |
US7383391B2 (en) | Prefetch mechanism based on page table attributes | |
US7284112B2 (en) | Multiple page size address translation incorporating page size prediction | |
US6681311B2 (en) | Translation lookaside buffer that caches memory type information | |
US7493452B2 (en) | Method to efficiently prefetch and batch compiler-assisted software cache accesses | |
US7996650B2 (en) | Microprocessor that performs speculative tablewalks | |
US9563568B2 (en) | Hierarchical cache structure and handling thereof | |
EP2159706B1 (en) | Operation processing apparatus and operation processing method | |
KR20030097871A (en) | Method and system for speculatively invalidating lines in a cache | |
KR20170139659A (en) | A computer processor having separate registers for addressing memory | |
US11520585B2 (en) | Prefetch store preallocation in an effective address-based cache directory | |
CN112753024A (en) | External memory based translation look-aside buffer | |
CN112416817A (en) | Prefetching method, information processing apparatus, device, and storage medium | |
US20140289469A1 (en) | Processor and control method of processor | |
CN114238167B (en) | Information prefetching method, processor and electronic equipment | |
US20090158003A1 (en) | Structure for a memory-centric page table walker | |
CN113448879A (en) | Memory management | |
CN114218132B (en) | Information prefetching method, processor and electronic equipment | |
JP2022548219A (en) | Disable conversion lookaside buffer |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: VIA TECHNOLOGIES, INC., TAIWAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:HOOKER, RODNEY E.;EDDY, COLIN;REEL/FRAME:031251/0847 Effective date: 20130502 |
|
STCB | Information on status: application discontinuation |
Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION |