`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`
`
`
`ARM Inc.
`Petitioner
`
`v.
`
`VANTAGE POINT TECHNOLOGY, INC.
`Patent Owner
`
`
`
`U.S. Patent No. 5,463,750
`Issue Date: October 31, 1995
`
`Title: METHOD AND APPARATUS FOR TRANSLATING VIRTUAL
`ADDRESSES IN A DATA PROCESSING SYSTEM HAVING MULTIPLE
`INSTRUCTION PIPELINES AND SEPARATE TLB’S FOR EACH PIPELINE
`
`
`
`
`
`
`DECLARATION OF V. THOMAS RHYNE, PH.D., P.E., R.P.A.
`
`
`
`
`
`
`
`
`
`
`Case No. IPR2014-TO BE ASSIGNED
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`ARM_VPT_IPR_00000419
`
`ARM Ex. 1009
`IPR Petition - USP 5,463,750
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`
`
`I, Vernon Thomas Rhyne, declare:
`
`1. I have been retained by Wiley Rein LLP to provide my opinions
`
`concerning the background of the computer art as related to the validity
`
`of U.S. Patent No. 5,463,750 (“the ’750 Patent”). In particular, in this
`
`Declaration I provide my opinions regarding the state of that art as of
`
`November 1993 for processor architectures that used virtual memory
`
`and mapping caches (known as “translation lookaside buffers” or
`
`“TLB’s”) to perform a translation which mapped virtual addresses to
`
`physical addresses used to access memory in a computer system. I am
`
`being compensated for my work in preparing this Declaration at the rate
`
`of $695 per hour, plus reimbursement of reasonable direct expenses. I
`
`have no other interest in this matter or the parties involved in this
`
`matter.
`
`I. QUALIFICATIONS
`
`2. My qualifications for forming the opinions set forth in this Declaration
`
`are summarized in the following paragraphs and explained in more
`
`detail in my curriculum vitae which is attached as part of Exhibit A to
`
`this report. Exhibit A also includes a list of my publications and a
`
`listing of my testimony during the past four years.
`
`
`
`1
`
`ARM_VPT_IPR_00000420
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`
`
`3. I have studied, taught, and practiced electrical engineering for over fifty
`
`years. I hold degrees from Mississippi State University (Bachelors of
`
`Science in Electrical Engineering with Honors, 1962), the University of
`
`Virginia (Masters of Electrical Engineering, 1964), and the Georgia
`
`Institute of Technology (Ph.D. in Electrical Engineering, 1967). I have
`
`been a registered Professional Engineer in the State of Texas since 1969
`
`(TX, No. 28,728). I have been a Registered Patent Agent since 1999
`
`(No. 45,041).
`
`4. I taught electrical engineering, computer engineering, computer
`
`architecture, and computer science at the undergraduate and graduate
`
`levels full-time at Texas A&M University from 1967 to 1983 and part-
`
`time at the graduate level at the University of Texas from 1983 to 1991.
`
`My twenty-plus years of industrial experience include work for the
`
`Electric Power Research Institute, Texas Instruments, Control Data
`
`Corporation, NASA, Texas Digital Systems, Inc. (a company I co-
`
`founded to produce microprocessor-based computer peripherals in
`
`1976), the Microelectronics and Computer Technology Corporation
`
`(MCC), and Motorola, Inc.
`
`5. I have extensive experience with computer technology, including
`
`design and teaching experience with a variety of computer systems,
`2
`
`
`
`ARM_VPT_IPR_00000421
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`microcomputer systems, and microcontrollers. I have participated in
`
`the design of several computer systems and microprocessors, and have
`
`designed systems which made use of those devices as control elements.
`
`I am an experienced programmer in a variety of programming
`
`languages as well as assembly-level language on a number of different
`
`computers and microprocessors.
`
`6. During my academic career I authored thirty technical papers and
`
`research reports; as I noted above, those papers and reports are listed in
`
`Exhibit 1. I also presented papers at a number of technical conferences
`
`and authored an award winning textbook, Fundamentals of Digital
`
`System Design, published by Prentice-Hall in 1973 and adopted at over
`
`thirty-five U.S. and international universities during its lifetime. My
`
`textbook has also been cited as a reference by the U.S. Patent and
`
`Trademark Office. I have also served as a technical reviewer for
`
`Prentice-Hall, the IEEE Transactions on Computing, and IEEE
`
`Spectrum.
`
`7. I have been a member of the IEEE for over fifty years and am a Life
`
`Fellow of that professional organization. I was elected to serve on the
`
`IEEE Board of Directors for two terms representing the engineering
`
`
`
`
`
`3
`
`ARM_VPT_IPR_00000422
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`education community and the IEEE Computer Society. I was also
`
`elected to two terms as the IEEE Treasurer.
`
`8. I have extensive experience with the accreditation of engineering and
`
`computer science programs in the U.S. and abroad, an activity which
`
`provided me an excellent opportunity to become familiar with the
`
`program curricula, faculties, and graduates from a large number of U.S.
`
`and international colleges and universities. I represented the IEEE for
`
`five years on the Engineering Accreditation Commission and for six
`
`years on the Board of Directors of the Accreditation Board for
`
`Engineering and Technology (ABET). I have assisted Japanese
`
`universities and industries in the establishment of the Japanese
`
`Accreditation Board for Engineering Education, and have led several
`
`other international accreditation missions.
`
`9. I was appointed by the U.S. National Research Council to the Panel of
`
`Assessment for the Electronics and Electrical Engineering Laboratory
`
`of the U.S. National Institute of Standards and Technology. I served on
`
`that Panel for seven years, including three terms as its chair; based on
`
`that work I was invited to provide testimony before the U.S. Congress
`
`regarding the status of the Laboratory.
`
`
`
`
`
`4
`
`ARM_VPT_IPR_00000423
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`10. My experience and qualifications have been recognized by the Texas
`
`Society of Professional Engineers (Young Engineer of the Year in
`
`Texas, 1973); the American Society for Engineering Education
`
`(Terman Awardee as the “Outstanding Young Electrical Engineering
`
`Educator in the U.S.,” 1980); the Institute of Electrical and Electronics
`
`Engineers (IEEE Fellow, 1990, recognizing my contributions to
`
`“computer engineering and computer engineering education”); the
`
`Accreditation Board for Engineering and Technology (ABET Fellow,
`
`1992); and the IEEE Computer Society (Golden Core Awardee, 2000).
`
`11. Based on my academic and consulting experience, I am familiar with a
`
`variety of computer-based user interfaces, database technologies, and
`
`data-communications protocols. I have managed large and complex
`
`software-development programs and have been and am familiar with
`
`the development of software systems for supporting complex
`
`operations.
`
`12. My experience has also included the use of a variety of communications
`
`systems such as the dial-up telephone network to carry data; the use of
`
`leased-line telephone connections to carry data; the use of the
`
`ARPANET (the precursor to the Internet) to carry data; and extensive
`
`
`
`
`
`5
`
`ARM_VPT_IPR_00000424
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`
`
`use of the Internet itself as that system came into being in the early
`
`1990’s.
`
`13. While at MCC I also managed MCC’s research and development
`
`programs dealing with graphical user interfaces, natural-language
`
`interfaces, and early website development tools. I also managed
`
`distributed database development for several years at MCC, as well as
`
`MCC’s successful research program on Internet-based credit card fraud
`
`detection using neural networks. While working at MCC in 1994 I was
`
`assigned to represent MCC and its participating companies in the initial
`
`planning efforts of the U.S. Technology Policy Working Group
`
`addressing the technical challenges associated with the planned
`
`National Information Infrastructure (“NII”), a national effort being
`
`promulgated, in a large part, by Al Gore through the U.S. Department
`
`of Commerce. The report of that initial planning effort is available at
`
`www.cra.org/govaffairs/archives/niireport/.
`
`14. My teaching and industrial experience has included the architectural
`
`design of a number of computers ranging from microprocessors to
`
`supercomputers. As part of that experience I have worked with the
`
`design of high-speed instruction pipelines and with virtual memory
`
`systems which made use of translation look-aside buffers. I taught
`6
`
`
`
`ARM_VPT_IPR_00000425
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`computer architecture at the graduate level at both Texas A&M and the
`
`University of Texas. At the latter institution I used the Hennessey and
`
`Patterson textbook for my graduate-level class.
`
`15. I retired from full-time work as of 1997 and draw retirement benefits
`
`from Texas A&M University. In addition to the work described above
`
`and in my curriculum vitae which is attached to this expert report as
`
`Exhibit 1, I have worked part-time as a consulting engineer for the past
`
`thirty years doing computer systems design, application-specific system
`
`design, and expert witness work in intellectual property litigation.
`
`16. I have read and am familiar with the ’750 patent and the file history of
`
`that patent.
`
`17. As part of my preparation for forming the opinions set forth in this
`
`Declaration, I have reviewed and relied upon Exhibits 1001, 1002, and
`
`1006-1013 as identified in the Exhibit list attached to the Requests for
`
`inter partes Review to which this Declaration is also attached. In
`
`particular, I reviewed and relied upon the following:
`
`• The ’750 patent;
`
`• The prosecution history of the ’750 patent;
`
`• United States Patents Nos. 3,947,823 (“the ’823 patent”);
`
`
`
`
`
`7
`
`ARM_VPT_IPR_00000426
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`• 4,935,835 (“the ’835 patent”); and 4,920,477 (“the ’477 patent”);
`
`• A well-known computer textbook entitled Computer Architecture: A
`
`Quantitative Approach by John L. Hennessy and David A. Patterson,
`
`Morgan Kaufman, 1990 (“Hennessy and Patterson”);
`
`• DEC VAX 8800, Digital Technical Journal, Number 4, February 1987,
`
`ISBN 1-55558-001-7. The Digital Technical Journal is published by
`
`Digital Equipment Corporation, 77 Reed Road, Hudson,
`
`Massachusetts 01749. Copyright © 1987 Digital Equipment Corporation
`
`(the “VAX 8800” reference);
`
`• “A VLIW Architecture for a Trace Scheduling Compiler,” Colwell et al,
`
`IEEE Transactions On Computers, Vol. 37, No. 8, August 1888 (IEEE)
`
`(“Colwell”);
`
`• The “Decision Denying Institution of Inter Partes Review” for
`
`IPR2014-00467 dated September 5, 2014 (the “Board Decision”);
`
`• Vantage Point’s Infringement Contentions for the Cortex A8 and Cortex
`
`A9 processors;
`
`• The PowerPC 601 Microprocessor, Moore, IEEE Proc. Compcon,
`
`Feb. 1993, pp. 109-116. (“Moore”); and
`
`
`
`
`
`8
`
`ARM_VPT_IPR_00000427
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`
`
`• “The 68040 Processor: Part 2, Memory Design and Chip Verification,”
`
`Robin W. Edenfield, et al., IEEE Micro, June 1990. (“Edenfield”).
`
`II. BACKGROUND OF THE ART
`
`18. I have reviewed the “Technical Introduction To The ’750 Patent”
`
`presented in the Request for inter partes Review to which my
`
`Declaration is attached. I agree with and adopt that summary herein.
`
`For brevity, my declaration does not repeat that summary. However,
`
`my Declaration does add some additional information of relevance to
`
`that introduction.
`
`19. By the time of the filing of the ’750 patent in November 1993, the
`
`technology for TLB’s was well-known to the art. The ’750 patent
`
`appears to acknowledge this in its descriptions of Figures 1-4 which are
`
`all identified as being part of conventional prior art systems. However,
`
`the ’750 patent does not provide any specific details about the structure
`
`or operation of TLB’s. Instead, the patent appears to rely on a prior art
`
`patent cited during prosecution (U.S. Patent No. 4,933,835) to provide
`
`the technical details of TLB technology.
`
`
`
`
`
`9
`
`ARM_VPT_IPR_00000428
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`
`
`III. THE VAX 8800
`
`20. I have reviewed the claim charts showing the correspondence between
`
`the VAX 8800 reference and claims 8-12 of the ’750 patent which are
`
`located in the Request for inter partes Review to which this Declaration
`
`is attached. I agree that that those charts show that the architecture of
`
`the VAX 8800 discloses to one of skill in the art all of the elements of
`
`claims 8-12 of the ’750 patent and therefore either anticipates or renders
`
`obvious those claims.
`
`21. I note that some aspects of the VAX 8800 use slightly different
`
`terminology from that used in the ’750 patent. For example, as shown
`
`in the following figure taken from page 43 of the VAX 8800 reference,
`
`the VAX 8800 uses the terms “translation buffer” or “TB” for the same
`
`component referenced in the ’750 patent as the “translation lookaside
`
`buffer” or “TLB.”
`
`
`
`10
`
`ARM_VPT_IPR_00000429
`
`
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`
`22. One of skill in the art would understand that the lower box shown above
`
`is depicted with three parallel lines which together form the common
`
`mathematical symbol for congruence or equivalence.1 Based on these
`
`symbols and the usage of the box in the above diagram, one of skill in
`
`the art would understand this box to represent a comparator which
`
`compared its two inputs (portions of virtual addresses) and output a
`
`signal indicating whether or not those addresses matched (the “TB
`
`HIT” output signal which, when TRUE, indicated that a given virtual
`
`address was found in the “TB” buffer).
`
`23. On page 41 of the VAX 8800 reference there is a discussion of the
`
`“Characteristics of the Translation Buffer.” One of skill in the art
`
`
`1 See, for example, http://en.wikipedia.org/wiki/Table_of_mathematical_symbols.
`11
`
`
`
`ARM_VPT_IPR_00000430
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`
`
`would understand that “direct mapped” is a term of art used to describe
`
`a family of caches. These caches are addressed by the input address
`
`being searched for (in this instance, the “tags” for the virtual address).
`
`Direct mapped caches are addressed by selecting only a single entry
`
`based on a subset of bits from a virtual address – the “tag.” Alternative
`
`cache types include “set-associative” and “fully-associative.” Those
`
`caches use a different approach to determining whether or not a given
`
`data item is currently held within them.
`
`IV. U.S. PATENT NO. 4,920,477
`
`24. I have reviewed the ’477 patent and believe it discloses all of the
`
`elements, either implicitly or explicitly, of claims 8-12 of the ’750
`
`patent. I have also reviewed the claim charts showing the
`
`correspondence between the ’477 patent reference and claims 8-12 of
`
`the ’750 patent which are within the Request for inter partes Review to
`
`which this Declaration is attached. I agree that those charts show that
`
`the ’477 patent discloses all of the elements of those claims to one of
`
`skill in the art.
`
`25. The ’477 patent shows a system with multiple instruction pipelines in
`
`which each pipeline includes its own data TLB for performing
`
`
`
`12
`
`ARM_VPT_IPR_00000431
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`translations of virtual address to physical address for data being
`
`accessed.
`
`26. Figure 1 of the ‘477 patent as copied below shows such a system with
`
`multiple pipelines each with its own data TLB.
`
`
`
`
`27. Each of the “integer processors” denominated as 20, 22, 24, and 26 in
`
`Figure 1 is an instruction pipeline. The “Background Of The
`
`Invention” provided in the ’477 patent states that “[t]he invention
`
`relates generally to pipelined computer apparatus….” 1:7-8. That
`
`section of the ’477 patent then continues as “[i]n a Trace computer,
`
`such as that described hereinafter …, the data processor has a pipelined
`
`
`
`13
`
`ARM_VPT_IPR_00000432
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`CPU….” 1: 22-25. Also, as stated in the ’477 patent, “program
`
`execution and memory operations herein are all pipelined.” 25:24-25.
`
`These passages indicate to one of skill in the art that the processors
`
`described in the ’477 patent operate as instruction “pipelines.”
`
`28. Additionally, as shown in Figure 3 of the ’477 patent, each “integer
`
`processor” is a pipeline having two ALUs, Integer ALU 0 and Integer
`
`ALU 1, both of which are connected to a TLB. ALU0 is responsible
`
`for generating memory addresses for load and store operations.
`
`See 23:15-26. “The integer processor, in the illustrated embodiment,
`
`includes two independent arithmetic logic units 70, 72 (designated
`
`ALUO and ALU1 respectively), 64 X 32-bit register file 74, a virtual to
`
`physical address data translation lookaside buffer 76 [TLB] …”
`
`(5:9-13). Thus, each TLB is associated with an integer processor that
`
`has its own pipeline and thus each TLB is associated with its own
`
`instruction pipeline.
`
`29. In Fig 12 of the ’477 patent, Element 76, a “TLB Ram” as discussed at
`
`23:33-43 and 24:10-29, is a translation buffer, and as shown in Fig. 1 of
`
`the patent, and the item marked “DTLB” in Integer Processor 3,
`
`Element 20, is a first translation buffer that is associated with the
`
`
`
`
`
`14
`
`ARM_VPT_IPR_00000433
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`instruction pipeline contained in the Integer Processor, an example of
`
`which is shown in Fig 3.
`
`30. Similar to the ’750 patent, the ’477 patent relies upon the knowledge of
`
`one skilled in the art regarding some of the details of TLB’s. For
`
`example, the ’477 patent simply notes that “the translation lookaside
`
`buffer translates virtual memory addresses from the ALU's to physical
`
`memory addresses using a table lookup mechanism well known to those
`
`practiced in the art, … ” 5:20-23.
`
`31. By the time of the filing of the ’477 patent in April 1987, and certainly
`
`by the time of the filing of the ’750 patent in November 1993, the
`
`structure and functionality of TLB’s was well known to one of skill in
`
`the art. See, e.g., the ’477 patent at 5:20-23 (“Functionally, the
`
`translation lookaside buffer translates virtual memory addresses from
`
`the ALU's to physical memory addresses using a table lookup
`
`mechanism well known to those practiced in the art”)
`
`32. For example, one of skill in the art would have understood that the TLB
`
`used in the ’477 patent would necessarily include storing a subset of
`
`translation data from the master translation memory, e.g., a page table
`
`located in main memory, because the whole purpose of a TLB is to act
`
`as a cache memory for the translation data in order to accelerate the
`15
`
`
`
`
`
`ARM_VPT_IPR_00000434
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`virtual-to-physical translation process. This is noted in many textbooks
`
`including Hennessy and Patterson, which describes a TLB as a “special
`
`cache” of the “page tables” that “are stored in main memory.” See
`
`pages 437-48. This is particularly true given the reference to “a table
`
`lookup mechanism well known to those practiced in the art” made in
`
`the ’477 patent as noted above.
`
`33. Additionally, one of skill in the art would understand the reference to
`
`“TLB RAM 76” to include a plurality of addressable storage locations –
`
`just as in the ’750 patent. “RAM” (a well-known acronym for
`
`“Random Access Memory”) is an addressable memory. The ’477
`
`patent identifies TLB 76 as “[a] high speed cache random access
`
`memory which can store for example 4,096 … thirty-two bit entries.”
`
`23:26-42. Similarly, the ’477 patent, in several locations discusses a
`
`“TLB entry” or the “TLB RAM entry.” See 24:10-29. The discussion
`
`of such entries indicates to one of skill in the art that the TLB includes
`
`multiple entries (as would be understood to be in a TLB by one of skill
`
`in the art) which are searched for the pending virtual-to-physical
`
`address translation and which can be addressed by the update
`
`mechanism to store the cached virtual-to-physical address translations.
`
`
`
`
`
`16
`
`ARM_VPT_IPR_00000435
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`
`
`34. The instruction pipelines (the “integer processors”) in the ’477 patent
`
`all receive their instructions from a common instruction issuing unit. In
`
`the Background of the Invention, the ’477 patent indicates that the
`
`system described therein “provides for parallel processing a very long
`
`instruction word having a length of, for example, 1000 or more bits….”
`
`1:28-33. This indicates that the different integer processors are separate
`
`pipelines working on different instructions (or portions of the larger
`
`1000-bit instruction) received from the same instruction unit.
`
`35. The ’477 system describes a “four cluster” system. “This is referred to,
`
`in the illustrated embodiment, as a “four-wide system” representing the
`
`four clusters which each have an integer processor. See 5:59-65. The
`
`’477 then continues on to describe how the 1024-bit instruction “very
`
`long” words for the four-wide system can be reduced proportionately
`
`for one-wide and two-wide systems to 256 and 512 bits respectively.
`
`See 6:4-23. Thus, each integer processor would receive 256 bits of the
`
`wide instruction word and the integer processors would all receive the
`
`instructions from the same instruction issuing unit.
`
`36. The ’477 patent both depicts and discusses a single “ITLB” which is the
`
`TLB used for fetching instructions. See Figure 1, Element 40; 4:57-5:4;
`
`and 17:35-65. Moreover, a single program counter is used for fetching
`17
`
`
`
`ARM_VPT_IPR_00000436
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`
`
`instructions for all of the clusters. See 15:30-60 and 16:38-68. These
`
`portions of the ’477 patent indicate to one of skill in the art that the
`
`“integer processors” all receive their instructions (in the case of the
`
`preferred embodiment, their portion of the overall 1024-bit instruction)
`
`from a common instruction issuing unit.
`
`V. U.S. PATENT NO. 4,933,835
`
`37. I have also reviewed the ’835 patent. I agree with the statements the
`
`applicant made during the prosecution of the ’750 patent that the ’835
`
`patent describes TLB technology that was well known by the filing date
`
`of the ’750 patent. See the Amendment dated 9/24/94 at pp. 1-2)
`
`(“address translators are well known and may be constructed in
`
`accordance with [the ’835 patent]”).
`
`38. I have reviewed the claim charts showing the correspondence between
`
`the ’835 patent reference and certain elements of claim 8 of the ’750
`
`patent which are within the Request for inter partes Review to which
`
`this Declaration is attached. I agree that as explained in those claim
`
`charts, the ’835 patent discloses those corresponding elements of those
`
`claims to one of skill in the art.
`
`
`
`18
`
`ARM_VPT_IPR_00000437
`
`
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`VI. COLWELL
`
`39. I also have reviewed the claim charts showing the correspondence
`
`between Colwell and certain elements of claim 8 of the ’750 patent
`
`which are located in the Request for inter partes Review to which this
`
`Declaration is attached. I agree that those charts show that Colwell
`
`discloses those corresponding elements of those claims to one of skill in
`
`the art.
`
`40. Colwell appears to describe a preferred embodiment of the ’477 patent.
`
`It uses much of the same terminology and was written by the inventors
`
`of the ’477 patent. It also appears to reference the same underlying
`
`computer design (the “Trace”) as is referenced in the ’477 patent.
`
`Thus, one of skill in the art would have readily combined the teachings
`
`of the ‘477 and Colwell as they appear to describe the same system.
`
`VII. COMBINING THE ‘835 PATENT WITH EITHER THE VAX 8800
`OR THE ‘477 PATENT IN COMBINATION WITH COLWELL
`
`41. In my opinion, one of skill in the art would readily combine the
`
`teachings of the ’835 patent with either the VAX 8800 reference or the
`
`’477 patent.
`
`42. During the course of the prosecution of the ’750 patent, the examiner
`
`rejected the pending claims and noted that “[i]t would have been
`
`
`
`19
`
`ARM_VPT_IPR_00000438
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`obvious to a person of ordinary skill in the art at the time the invention
`
`was made that employing plural units of identical function to assist
`
`plural function is commonly practiced, and the person would have
`
`employed separate TLB's for each pipeline to assist in address
`
`translation.” See the Office Action dated 7/28/94 at p. 4. I agree with
`
`the examiner’s statement that employing plural units of identical
`
`function (“parallelism”) to assist with the performance of plural
`
`functions was commonly practiced as of the date of the ’750 patent
`
`application. Moreover, one of skill in the art would have understood
`
`that separate TLB’s could be used in each instruction pipeline to assist
`
`in address translation when an architecture provided multiple
`
`instruction pipelines.
`
`43. The VAX 8800 and the ’477 patent provide good examples of using
`
`“plural units of identical function to assist plural function” by their use
`
`of separate TLB’s in each pipeline as discussed in the Requests for inter
`
`partes Review to which my Declaration is attached. As noted by the
`
`examiner, this was common practice at the time of the invention.
`
`44. United States Pat. No. 3,947,823 also provides another good example of
`
`the use of multiple pipelined processors, each containing their own
`
`TLB. The ’823 patent discloses multiple CPU’s where “[e]ach CPU
`20
`
`
`
`
`
`ARM_VPT_IPR_00000439
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`includes … a dynamic address translation mechanism (DATM) which
`
`includes a translation lookaside buffer (TLB).” 2:57-64. Using separate
`
`TLB’s in separate instruction processing pipelines merely reflects the
`
`common, long-used, practice of replicating identical units of identical
`
`function within a parallelized computing system.
`
`45. As noted above, the technology for TLB’s and the structure and
`
`operation of TLB’s were known for at least 20 years before the filing of
`
`the ’750 patent. While different companies used different names for the
`
`functionality of the TLB’s, the basic functionality of those TLB’s was
`
`essentially the same. The TLB’s were used to translate virtual
`
`addresses into physical addresses. In a virtual-memory system this
`
`translation was performed by searching through a virtual-to-physical
`
`translation table which matched the virtual addresses to physical
`
`addresses. When an entry was located that matched the virtual address,
`
`then the physical address was read from that entry of the table. TLB’s
`
`merely employed one long-known computer idea (using small “cache”
`
`memories to speed access to particular data elements) to store portions
`
`of the translation table. While minor variations may exist between
`
`different virtual memory systems (e.g., the number of bits used to store
`
`the virtual and physical addresses, or the techniques used to search the
`21
`
`
`
`
`
`ARM_VPT_IPR_00000440
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`table), the basic operations of TLB’s involved a finite number of
`
`identified, predictable solutions known long before the filing of
`
`the ’750 patent. The claims of the ’750 patent merely recite well-
`
`known structures which were used to perform the translation of virtual
`
`addresses to physical addresses and to update the translation cache
`
`known as the TLB.
`
`46. To the extent that the VAX 8800 alone did not disclose all of the
`
`structural elements of the TLB’s recited in the claims of the ’750 patent,
`
`one of skill in the art would have readily considered, and substituted,
`
`the structural details of the TLB described in the ’835 patent for the
`
`“translation buffer” (“TB”) used and discussed in the VAX 8800
`
`reference. The VAX 8800 reference merely reflects the common
`
`functionality of TLB’s and does not require any particular functionality
`
`of the “TB” outside of that which was known in the prior art. A person
`
`skilled in the art in 1993 would have understood that the basic structure
`
`for the TLB disclosed in the ’835 patent could readily have been
`
`adapted to the VAX 8800. The substitution of the TLB’s disclosed in
`
`the ’835 patent for the “TB’s” disclosed in the VAX 8800 would have
`
`been obvious to one of skill in the art and could readily be
`
`implemented. One would have been motivated to make such a solution
`22
`
`
`
`
`
`ARM_VPT_IPR_00000441
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`by the very small number of different solutions known and used to
`
`solve this problem which had been solved for 20+ years before the ’750
`
`patent.
`
`47. While the specific algorithm for translating virtual addresses to physical
`
`addresses might be different between the ’835 patent and the VAX
`
`8800 (meaning, the specific manipulation of bits used to calculate the
`
`correspondence between virtual and physical addresses), the claims of
`
`the ’750 patent do not pertain to this difference. Rather those claims
`
`pertain to a specific higher-level architecture and the basic process of
`
`checking the TLB for a TLB “hit,” searching the page table for a TLB
`
`“miss,” and updating the TLB cache memory as part of the TLB “miss”
`
`processing. The ’835 patent and the VAX 8800 reference disclose
`
`those same basic processes, and the structure used to perform these
`
`processes would have been known by one of skill in the art to be
`
`interchangeable.
`
`48. To the extent that the disclosure of the different execution units in the
`
`VAX 8800 does not meet the construction of the term “instruction
`
`pipeline,” it would have been obvious to apply the teachings of the
`
`VAX 8800 to separate “instruction pipelines.” The point of using
`
`separate TLB’s in the VAX 8800 execution units was to permit each
`23
`
`
`
`
`
`ARM_VPT_IPR_00000442
`
`
`
`
`U.S. Patent No. 5,463,750
`Petition for Inter Partes Review, Declaration of V. Thomas Rhyne, Ph. D
`
`
`
`execution path to operate independently and efficiently with respect to
`
`the addressing of virtual memory. Such an approach of using a
`
`plurality of known, commonly used functional blocks for known
`
`functions in different portions of a computer architecture would have
`
`been obvious to one of skill in the art by November 1993.
`
`49. To the extent that the ’477 patent in combination with Colwell does not
`
`disclose all of the structural elements of the TLB’s recited in the claims
`
`of the ’750 patent, one of skill in the art would have readily considered,
`
`and substituted, the TLB described in the ’835 patent for “TLB 76” as
`
`disclosed and discussed in the ’477 patent and the TLB disclosed in
`
`Colwell. The ’477 patent merely reflects the common function