Advertisement

Speicherarchitektur

  • Christian Müller-Schloer
  • Ernst Schmitter
Chapter
  • 42 Downloads

Zusammenfassung

Die Fortschritte in der VLSI-Technologie führen dazu, daß Mikroprozessoren bei zunehmender Komplexität mit immer höheren Taktraten betrieben werden und so Mikroprozessorsysteme in neue Leistungsklassen vorstoßen können. Einen weiteren Leistungssprung konnte man durch die Übernahme der RISC-Entwurfsphilosophie erzielen. Ungünstigerweise stellen aber besonders RISC-Prozessoren harte Randbedingungen in Bezug auf die benötigte Speicherbandbreite. Um diese bereitstellen zu können, muß man auf die bisher nur bei teueren Rechnerklassen — wie Minicomputer, Mainframes oder Supercomputer — verwendete Technik des Cache-Speichers zurückzugreifen. Dies wirkt sich jedoch auch auf andere Rechnerkomponenten aus, insbesondere auf den Hauptspeicher, der nun auch Blockzugriffe unterstützen muß. In geringem Umfang kann sogar das Betriebssystem durch die Cache-Entscheidung betroffen sein, falls der Cache bei einem Prozeßwechsel ungültig gesetzt werden muß. Mikroprozessoren der oberen Leistungsklasse haben den 64 KByte großen Adreßraum längst hinter sich gelassen. Dadurch ist es möglich, Betriebssysteme zu installieren, die früher den bereits erwähnten teueren Rechnerklassen vorbehalten waren. Ein Betriebssystem wie UNIX stellt aber neue Anforderungen an Prozessor- und Speicherarchitektur, da z. B. wegen des sehr großen Speicherplatzbedarfs und der Multi-Tasking/Multi- User-Eigenschaft ein virtueller Speicher eingerichtet werden muß. Dieses Kapitel versucht, Grundlagen zum Thema „Speicherarchitektur“ zu vermitteln. Um die Realität nicht aus den Augen zu verlieren, wird auf die Konzepte gängiger RISC- Architekturen eingegangen.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Literatur

  1. [ABDE 81]
    Abdel-Hamid, R.; Madnik, S.: A Study of the Multicache Consistency Problem in Multiprocessor Computer Systems. MIT Technical Report 16, Center for Information Research, September, 1981.Google Scholar
  2. [ALEX 85]
    Alexander, C.; Keshlear, W.; Briggs, F.: Translation Buffer Performance in a UNIX Environment. Computer Architecture News 13 (1985) 3, pp. 2–14.CrossRefGoogle Scholar
  3. [ALPE 83]
    Alpert, D.: Performance Tradeoffs for Microprocessor Cache Memories. Technical Note No. 83-239, Stanford University, Department of Electrical Engineering und Computer Science, December 1983.Google Scholar
  4. [ARCH 86]
    Archibald, J.; Baer, J.: Cache Coherence Protocols: Evaluation Using a Multiprocessor Simulation Model. ACM Transactions on Computer Sytems 4 (1986) 4, pp. 273–298.CrossRefGoogle Scholar
  5. [BERE 89]
    Beresford, R.: Partitioning Memory Control for RISC Machines. High Performance Systems, April 1989.Google Scholar
  6. [BITA 85]
    Bitar, P.: Fast Synchronisation for Shared-Memory Multiprocessors. Technical Report TR 85.11, Research Institute for Advanced Computer Science, NASA Ames Research Center, December, 1985.Google Scholar
  7. [BITA 86]
    Bitar, P; Despain A. M.: Multiprocessor Cache Synchronization. Proc. 13th Ann. Int. Symp. on Computer Architecture, Tokyo, Japan, June, 1986, pp. 424–433.Google Scholar
  8. [BORR 85]
    Borriello, G.; Eggers, S.; Katz, R.; McKinley, H; Perkins, C.; Scott, W.; Sheldon, R.; Whalen, S.; Wodd, D.: Design and Implementation of an Integrated Snooping Data Cache. UC Berkeley CS Technical Report UCB/CSD 85/199, January, 1985.Google Scholar
  9. [BREN 84]
    Brenza, J.: Cross-Interrogate Directory for a Real, Virtual or Combined Real/Virtual Cache. IBM Technical Disclosure Bulletin, 26 (1984) 4, pp. 6069–6070.Google Scholar
  10. [BRIG 83]
    Briggs, F; Dubois, M.: Effectiveness of Private Caches in Multiprocessor Systems with Parallel-Pipelined Memories. IEEE Transactions on Computers, C-32 (1983) 1, pp. 48–59.CrossRefGoogle Scholar
  11. [CENS 78]
    Cerisier, L.; Feautrier, P.: A New Solution to Coherence Problems in Multicache Systems. IEEE Transactions on Computers, C-27 (1978) 12, pp. 1112–1118.CrossRefGoogle Scholar
  12. [CRIS 88]
    Crisp, R.; et. al.: Designing a Cache for a fast Processor. Electronic Design, 36 (1988) 20.Google Scholar
  13. [DENN 70]
    Denning, P.: Virtual Memory. ACM Computing Surveys, II (1970) 5.Google Scholar
  14. [DUBO 82]
    Dubois, M.; Briggs, F.: Effects of Cache Coherency in Multiprocessors. IEEE Transactions on Computers, C-31 (1982) 11, pp.1083–1099.CrossRefGoogle Scholar
  15. [DUBO 85]
    Dubois, M.: A Cache Based Multiprocessor with High Efficiency. IEEE Transactions on Computers, C-34 (1985) 10, pp.968–997.CrossRefGoogle Scholar
  16. [FLET 84]
    Fletcher, R.; Stein, D.; Wladawsky-Berger, I.: Three Level Memory Hierarchy Using Write and Share Flags, US: Patent, 4, 442, 487, April 10, 1984.Google Scholar
  17. [FRAN 84]
    Frank, S.: Tightly Coupled Multiprocessor Systems Speeds Memory Access Times. Electronics 57 (1984) 1, pp. 164–169.Google Scholar
  18. [FREE 85]
    Freeman, C.: Read-Modify-Write Data Integrity. IBM Technical Disclosure Bulletin, 27 (1985) 1, pp. 5019–5020.MathSciNetGoogle Scholar
  19. [FURT 87]
    Furth, B.; Milutinovic, V.: A Survey of Multiprocessor Architectures for Memory Mangement. IEEE Computer 20 (1987) 3, pp. 498–467.Google Scholar
  20. [GOOD 83]
    Goodman, J.: Using Cache Memory to Reduce Processor — Memory Traffic. Proc. 10th Ann. Int. Symp. on Computer Architecture, Stockholm, Sweden, June 1983, pp. 124–131.Google Scholar
  21. [GOOD 87]
    Goodman, J.: Cache Memory Optimization to Reduce Processor / Memory Traffic. Journal of VLSI and Computer Systems. 2 (1987) 1-2, pp. 51–86.Google Scholar
  22. [HAMA 84]
    von Hamacher, C.; et. al.: Computer Organization, New York: Mc Graw Hill, 1984, pp. 288–329.Google Scholar
  23. [HANS 86]
    Hansen, C.; et. al.: A RISC Microprocessor with integral MMU and Cache Interface. Proceedings ICCD, IEEE, October 1986.Google Scholar
  24. [HAYE 88]
    Hayes, J.: Computer Architecture and Organization. Mc Graw Hill, 1988, pp. 404–457.Google Scholar
  25. [HENN 90]
    Hennessy, J.; Patterson, D.: Computer Architecture, a Quantitative Approach, Morgan Kaufmann Publishers, 1990.Google Scholar
  26. [HILL 84]
    Hill, M.; Smith, A.: Experimental Evalution of On-Chip Microprocessor Cache Memories. 11th Ann. Int. Symp. on Computer Architecture, 1984, S. 158ff.Google Scholar
  27. [HILL 87]
    Hill, M.: Aspects of Cache Memory and Instruction Buffer Performance. UC Berkeley CS Technical Report UCB/CSD 87/381, November, 1987Google Scholar
  28. [HWAN 85]
    Hwang; Briggs: Computer Architecture and Parallel Processing: Memory and Input-Output Subsystems. New York: Mc Graw Hill, 1985,pp.52–144.Google Scholar
  29. [IBM 84]
    IBM Technical Disclosure Bulletin, 26 (1984) 3.Google Scholar
  30. [IBM 85]
    IBM Technical Disclosure Bulletin, 27 (1985) 4.Google Scholar
  31. [IBM 85a]
    IBM Technical Disclosure Bulletin, 27 (1985) 5.Google Scholar
  32. [ISCA 83]
    Proc. 10th Ann. Int. Symp. on Computer Architecture, Stockholm, Sweden, June 1983.Google Scholar
  33. [ISCA 84]
    Proc. 11th Ann. Int. Symp. on Computer Architecture, Ann Arbor, MI, USA, June 1984.Google Scholar
  34. [ISCA 85]
    Proc. 12th Ann. Int. Symp. on Computer Architecture, Boston, MA, USA, June 1985.Google Scholar
  35. [ISCA 86]
    Proc. 13th Ann. Int. Symp. on Computer Architecture, Tokyo, Japan, June 1986.Google Scholar
  36. [JONE 77]
    Jones, J.; Junod, D.: Cache Address Directory Invalidation Scheme for Multiprocessing System. IBM Technical Disclosure Bulletin, 20 (1977) 6, pp. 295–296.Google Scholar
  37. [KANE 89]
    Kane, G.: MIPS RISC Architecture, Englewood Cliffs (NJ): Prentice Hall, 1989.Google Scholar
  38. [KATZ 85]
    Katz, R.; Eggers, S.; Wodd, D.; Perkins, C.; Sheldon. R.: Implementing a Cache Consistency Protocol. Proc. 12th Ann. Symp. on Comp. Arch., Boston, MA, June, 1985, pp. 276–283.Google Scholar
  39. [KATZ 85a]
    Katz, R.; Eggers, S.; Gibson, G.; Hanson, P.; Hill, M.; Pendleton, J.; Ritchie, S.; Taylor, G.; Wood, D.; Patterson, D.: Memory Hierarchy Aspects of a Multiprocessor RISC: Cache and Bus Analyses. Report No. UCB/CSD 85/221, Comp. Science Division, University of California, Berkeley, CA 94720, January, 1985.Google Scholar
  40. [KNAP 85]
    Knapp, V.; Baer, J.: Virtually Addressed Caches for Multiprogramming and Multiprocessing Environments. Proc. of the 18th Ann. Hawaii Int. Conf. on System Sciences, 1985, pp. 477–486.Google Scholar
  41. [LAWE 89]
    Lawell, T; et. al.: Die Schnittstelle des RISC-Chipsatzes 88000. Design&Elektronik, (1984) 4.Google Scholar
  42. [LEE 80]
    Lee, J.; Weinberger, A.: A Solution to the Synonym Problem. IBM Technical Disclosure Bulletin, 22 (1980) 1, pp. 3331–3333.Google Scholar
  43. [MANO 82]
    Mano, M.: Computer System Architecture, Englewood Cliffs (NJ): Prentice-Hall Inc., 1982.Google Scholar
  44. [MATT 82]
    Mattos, P.; Frenzel, D.: Nibble Mode beschleunigt Speicherzugriff. Elektronik 31 (1982) 15.Google Scholar
  45. [McCRE 84]
    McCreight, E.: The Dragon computer system: An early overview. Tech. Rep., Xerox Corp., September, 1984Google Scholar
  46. [MEKH 89]
    Mekhiel, N.: Speed System Memory by Interleaving DRAM Accesses. Electronic Design, 37 (1989) 20.Google Scholar
  47. [MEYE 88]
    Meyer, A.: MMU Requires tailoring to meet needs of demandpaging UNIX. Computer Design, 27 (1988) 17.Google Scholar
  48. [MILE 90]
    Milenkovic, M.: Microprocessor Memory Management Units. IEEE Micro, 10 (1990) 4, pp. 70–85.MathSciNetCrossRefGoogle Scholar
  49. [MOTO 88]
    32-Bit Cache mit integrierter MMU für den 88100. Design& Elektronik, (1988) 9.Google Scholar
  50. [NORT 82]
    Norton, R.; Abraham, J.: Using Write Back Cache to Improve Performance of Multiuser Mulitprocessors. Proc. 1982 Int. Conf. on Parallel Processing, August, 1982, pp. 326–331.Google Scholar
  51. [PAPA 84]
    Papamarcos, M.: A Low Overhead Coherence Solution for Bus-Organized Multiprocessors with Private Cache Memories. Report CSG-29, Coordinated Science Laboratory, Computer Systems Group, University of Illinois at Urbana-Champaign, May, 1984.Google Scholar
  52. [PASZ 88]
    Paszcza, M.: Speicherverwaltungstechniken. mc 7 (1988) 9.Google Scholar
  53. [PATE 81]
    Patel, J.: A Performance Model for Multiprocessors with Private Cache Memories. Int. Conf. on Parallel Processing, August, 1981, pp. 314–317.Google Scholar
  54. [PATE 82]
    Patel, J.: Analysis of Multiprocessors with Private Cache Memories. IEEE Transactions on Computers, (C-3K1982) 4, pp. 296–304.Google Scholar
  55. [PATT 83]
    Patterson, D.; Garrison P.; Hill M.; Lioupis D.; Nyberg C.; Sippel T.; Van Dyke K.: Architecture of VLSI Instruction Cache for a RISC. Proc. 10th Ann. Int. Symp. on Computer Architecture, Stockholm, Sweden, June, 1983.Google Scholar
  56. [PNEV 90]
    Pnevmatikatos D. N.; Hill M.: Cache Performance of the Integer SPEC Benchmarks on a RISC. Computer Architecture News 18 (1990) 2.CrossRefGoogle Scholar
  57. [SCHE 87]
    Scheurich, C.; Dubois, M.: Correct memory Operation of Cache-Based Multiprocessors. Proc. of the Ann. Int. Symp. on Computer Architecture, 1987, pp. 234–243.Google Scholar
  58. [SMIT 82]
    Smith, A.: Cache Memories. Computing Surveys 14 (1982) 3, pp. 473–530.CrossRefGoogle Scholar
  59. [SMIT 83]
    Smith, J. E.; Goodman J. R.: A Study of Instruction Cache Organizations and Replacement Policies. ACM Computer Architecture News, 3, 1983, pp. 132–137CrossRefGoogle Scholar
  60. [SMIT 85]
    Smith, A.: Problems, Directions and Issues in Memory Hierarchies. Proc. 18th Ann. Hawaii Int. Conf. on Systems Science, Honolulu, Hawai, January 2-4, 1985, pp. 468–476, Also Avaiable as UC Berkeley CS Report UCB/CSD84/220.Google Scholar
  61. [SMIT 85a]
    Smith, A.: CPU Cache Consistency with Software Support and Using ‘One Time Identifiers’. Proc. Pacific Computer Communications Symposium, Seoul, Republic of Korea, October, 1985, pp. 142–150.Google Scholar
  62. [STON 87]
    Stone, H.: High Performance Computer Architecture. Reading: Addison Wesley 1987, pp. 21–94.Google Scholar
  63. [SWEA 86]
    Sweazey, P.: Cache Coherence on the Futurebus. Proceedings: Configuring Buses into Systems. BUSCON/86, January, 1986, pp. 122–127.Google Scholar
  64. [TRAT 86]
    Trattnig, W.: Virtuelle Speicherverwaltung in mc-Systemen. Elektronik 35 (1986) 2, S. 67–74.Google Scholar
  65. [WIEG 82]
    Wiegmann, A.: Dynamische 64-k-RAMs. Elektronik 31 (1982) 25.Google Scholar
  66. [WYLA 87]
    Wyland, D.: Cache tag RAM chips boost speed and simplify design, November 1, 1987.Google Scholar
  67. [YEN 82]
    Yen, W.; Fu, K.: Coherence Problem in a Multi-Cache System. Proc. 1982 Int. Conf. on Parallel Processing, August, 1982, pp. 332–339.Google Scholar
  68. [YEN 85]
    Yen, W.; Yen, D.; FU, K.: Data Coherence Problem in Multicache System. IEEE Transactions on Computers, C-34 (1985) 1, pp. 56–65.CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • Christian Müller-Schloer
    • 1
  • Ernst Schmitter
    • 2
  1. 1.Institut für Rechnerstrukturen und BetriebssystemeUniversität HannoverGermany
  2. 2.Zentralabteilung Forschung und EntwicklungSiemens AGMünchenGermany

Personalised recommendations