计算机组成与设计 硬件 软件接口-第四版-课后题答案

1SolutionsSolution1.11.1.1Computerusedtorunlargeproblemsandusuallyaccessedviaanetwork:5supercomputers1.1.21015or250bytes:7petabyte1.1.3Computercomposedofhundredstothousandsofprocessorsandterabytesofmemory:3servers1.1.4Today’ssciencefictionapplicationthatprobablywillbeavailableinnearfuture:1virtualworlds1.1.5Akindofmemorycalledrandomaccessmemory:12RAM1.1.6Partofacomputercalledcentralprocessorunit:13CPU1.1.7Thousandsofprocessorsformingalargecluster:8datacenters1.1.8Amicroprocessorcontainingseveralprocessorsinthesamechip:10multi-coreprocessors1.1.9Desktopcomputerwithoutscreenorkeyboardusuallyaccessedviaanet-work:4low-endservers1.1.10Currentlythelargestclassofcomputerthatrunsoneapplicationoronesetofrelatedapplications:9embeddedcomputers1.1.11Speciallanguageusedtodescribehardwarecomponents:11VHDL1.1.12Personalcomputerdeliveringgoodperformancetosingleusersatlowcost:2desktopcomputers1.1.13Programthattranslatesstatementsinhigh-levellanguagetoassemblylanguage:15compilerS2Chapter1Solutions1.1.14Programthattranslatessymbolicinstructionstobinaryinstructions:21assembler1.1.15High-levellanguageforbusinessdataprocessing:25cobol1.1.16Binarylanguagethattheprocessorcanunderstand:19machinelanguage1.1.17Commandsthattheprocessorsunderstand:17instruction1.1.18High-levellanguageforscientificcomputation:26fortran1.1.19Symbolicrepresentationofmachineinstructions:18assemblylanguage1.1.20Interfacebetweenuser’sprogramandhardwareprovidingavarietyofservicesandsupervisionfunctions:14operatingsystem1.1.21Software/programsdevelopedbytheusers:24applicationsoftware1.1.22Binarydigit(value0or1):16bit1.1.23Softwarelayerbetweentheapplicationsoftwareandthehardwarethatincludestheoperatingsystemandthecompilers:23systemsoftware1.1.24High-levellanguageusedtowriteapplicationandsystemsoftware:20C1.1.25Portablelanguagecomposedofwordsandalgebraicexpressionsthatmustbetranslatedintoassemblylanguagebeforeruninacomputer:22high-levellanguage1.1.261012or240bytes:6terabyteSolution1.21.2.18bits×3colors=24bits/pixel=4bytes/pixel.1280×800pixels=1,024,000pixels.1,024,000pixels×4bytes/pixel=4,096,000bytes(approx4Mbytes).1.2.22GB=2000Mbytes.No.frames=2000Mbytes/4Mbytes=500frames.1.2.3Networkspeed:1gigabitnetwork==1gigabit/persecond=125Mbytes/second.Filesize:256Kbytes=0.256Mbytes.Timefor0.256Mbytes=0.256/125=2.048ms.Chapter1SolutionsS31.2.42microsecondsfromcache==20microsecondsfromDRAM.20micro-secondsfromDRAM==2secondsfrommagneticdisk.20microsecondsfromDRAM==2msfromflashmemory.Solution1.31.3.1P2hasthehighestperformanceperformanceofP1(instructions/sec)=2×109/1.5=1.33×109performanceofP2(instructions/sec)=1.5×109/1.0=1.5×109performanceofP3(instructions/sec)=3×109/2.5=1.2×1091.3.2No.cycles=time×clockratecycles(P1)=10×2×109=20×109scycles(P2)=10×1.5×109=15×109scycles(P3)=10×3×109=30×109stime=(No.instr.×CPI)/clockrate,thenNo.instructions=No.cycles/CPIinstructions(P1)=20×109/1.5=13.33×109instructions(P2)=15×109/1=15×109instructions(P3)=30×109/2.5=12×1091.3.3timenew=timeold×0.7=7sCPI=CPI×1.2,thenCPI(P1)=1.8,CPI(P2)=1.2,CPI(P3)=3ƒ=No.instr.×CPI/time,thenƒ(P1)=13.33×109×1.8/7=3.42GHzƒ(P2)=15×109×1.2/7=2.57GHzƒ(P3)=12×109×3/7=5.14GHz1.3.4IPC=1/CPI=No.instr./(time×clockrate)IPC(P1)=1.42IPC(P2)=2IPC(P3)=3.331.3.5Timenew/Timeold=7/10=0.7.Soƒnew=ƒold/0.7=1.5GHz/0.7=2.14GHz.1.3.6Timenew/Timeold=9/10=0.9.SoInstructionsnew=Instructionsold×0.9=30×109×0.9=27×109.S4Chapter1SolutionsSolution1.41.4.1P2ClassA:105instr.ClassB:2×105instr.ClassC:5×105instr.ClassD:2×105instr.Time=No.instr.×CPI/clockrateP1:TimeclassA=0.66×10−4TimeclassB=2.66×10−4TimeclassC=10×10−4TimeclassD=5.33×10−4TotaltimeP1=18.65×10−4P2:TimeclassA=10−4TimeclassB=2×10−4TimeclassC=5×10−4TimeclassD=3×10−4TotaltimeP2=11×10−41.4.2CPI=time×clockrate/No.instr.CPI(P1)=18.65×10−4×1.5×109/106=2.79CPI(P2)=11×10−4×2×109/106=2.21.4.3clockcycles(P1)=105×1+2×105×2+5×105×3+2×105×4=28×105clockcycles(P2)=105×2+2×105×2+5×105×2+2×105×3=22×1051.4.4(500×1+50×5+100×5+50×2)×0.5×10–9=675ns1.4.5CPI=time×clockrate/No.instr.CPI=675×10–9×2×109/700=1.921.4.6Time=(500×1+50×5+50×5+50×2)×0.5×10–9=550nsSpeed-up=675ns/550ns=1.22CPI=550×10–9×2×109/700=1.57Chapter1SolutionsS5Solution1.51.5.1a.1G,0.75Ginst/sb.1G,1.5Ginst/s1.5.2a.P2is1.33timesfasterthanP1b.P1is1.03timesfasterthanP21.5.3a.P2is1.31timesfasterthanP1b.P1is1.00timesfasterthanP21.5.4a.2.05µsb.1.93µs1.5.5a.0.71µsb.0.86µs1.5.6a.1.30timesfasterb.1.40timesfasterSolution1.61.6.1CompilerACPICompilerBCPIa.1.001.17b.0.800.58S6Chapter1Solutions1.6.2a.0.86b.1.371.6.3CompilerAspeed-upCompilerBspeed-upa.1.521.77b.1.210.881.6.4P1peakP2peaka.4GInst/s3GInst/sb.4GInst/s3GInst/s1.6.5Speed-up,P1versusP2:a.0.967105263b.0.7302631581.6.6a.6.204081633b.8.216216216Solution1.71.7.1Geometricmeanclockrateratio=(1.28×1.56×2.64×3.03×10.00×1.80×0.74)1/7=2.15Geometricmeanpowerratio=(1.24×1.20×2.06×2.88×2.59×1.37×0.92)1/7=1.621.7.2Largestclockrateratio=2000MHz/200MHz=10(PentiumProtoPentium4Willamette)Largestpowerratio=29.1W/10.1W=2.88(PentiumtoPentiumPro)Chapter1SolutionsS71.7.3Clockrate:2.667×109/12.5×106=212.8Power:95W/3.3W=28.781.