高級體系結(jié)構(gòu)期末考試課后習(xí)題匯總

1、Homework1,1 Explain the Concepts 系統(tǒng)結(jié)構(gòu)(CA: Computer Architectute)、高級系統(tǒng)結(jié)構(gòu)(Advanced CA)、Amdahl law、SCALAR PROCESSING、LOOK-AHEAD、PVP、SMP、MPP、DSM、COW、GCE、CISC、RISC、VMM、SUPERCOMPUTER、SVM、MAINFRAME、COMPUTER SYSTEM ON CHIP、PARALLEL ARCHITECTURE INTO SINGLE CHIP、MOORE定律、UMA、NUMA、COMA、CC-NUMA、NORMA、SHELL ARCH

2、ITECTURE, PRAM, BSP, Log P 2 Explain the Flynn classify and its semantics. 3 Where are the difficulties of Parallel Processing ? What is Parallel(并行)、concurrent(并發(fā)) and Simultaneous(同時)separately ? 4 Assume that a system component be speeded to 10 times, its past processing time accounts for 40% of

3、the system, how much will be the system performance improved? 5 Please draw the memory Hierarchy chart.,homework2,1 Explaining Conceptions RISC, CISC, VLIW, SUPER-SCALAR, SUPER-PIPELINE, SUPER-SCALAR-SUPER-PIPELINE, IPC, SINGLE ISSUE,MULTIPLE ISSUE, OOO, Multi-Threading 2 In ideal case, please give

4、the performances for SUPER-SCALAR, SUPER-PIPELINE, SUPER-SCALAR-SUPER-PIPELINE, and give N=8 examples and their average IPC. 3 Please describe the CPU technology in recent years, and give an example to illustrate their technical parameters. Please give computing trace of ai+7=bj+1+ck+8+dm+10 for T90

5、00 How to solve the multicore memory wall problem? 6 Simple computer design test: Supposed there are a simple CPU with A15-A0,D7-D0,read and write, an register and Tri-state buffers for I/O device, a 8k*8 ROM with A12-A0,D7-D0,CE and OE, and a 8k*8 RAM with A12-A0,D7-D0,CE,RD and WR. Please you desi

6、gn the circuit of a simple computer.,Homework,1) There is a program in a 40MHz processor，its number of integer arithmetic instructions is 45000 and the clock is 1.The number of instructions with data operations is 32000 and the clock is 2. The number of float instructions is 15000 and its clock 2. T

7、he number of instructions such as JMP etc is 8000 and its clock is 2.Please give its CPI、MIPS and the CPU execution time. 2) If the communication cost in a SP2 machine is t（m）=46+（0.035）m, please give its (漸進帶寬 )r and its (半峰值信息長度) m1/2. 3) If a N*N matrix A is multiplied by a N*N matrix B, its time

8、 is T1=CN3 s, but the execution time in a machine with n nodes is Tn=(CN3/n+bN2/square-root(N) s，here,C,N and b is a constant. Please you give its Speedup under fix-workload, fix-time and fix-memory usage.,4) (等效率)see :kumar V, Rao V N. Parallel depth firsh search, part ii: Analysis. IntI J of Paral

9、lel Programming,1987,16(6):501-519 5) (等速度)see:Sun X H,etc. scalability of parallel algorithm-machine combina-tions. IEEE transaction on parallel and distributed system,1994,5(6):519-613 6) (等延遲)see: Zhang X D,etc. latency metric: an experimental method for measuring and evaluating parallel program

10、and architecture scalability.J of parallel and distributed Computing,1994,22:392-410.,2021/3/29,5,HOMEWORK,1解釋Concepts State network(靜態(tài)網(wǎng)絡(luò))、Daymic network(動態(tài)網(wǎng)絡(luò))、crossbar(交叉開關(guān))、Multi-buses(多總線) 2 Draw N=25 MESH 3寫出用混洗交換網(wǎng)絡(luò)模擬單級立方體網(wǎng)絡(luò)的互連函數(shù)表達式please give the single cube function by using SHUFFLE-EXCHANGE N

11、ETWORK 4 How much have the different bus Arbitrate Alogrithms ? What is its merit and demerit? 總線仲裁有幾種，各有何優(yōu)缺點？,HOMEWORK,1 THERE ARE P1,P2,P3,P4 AND M1,M2,M3,M4,PLEASE USE 2*2 CROSSBAR SWITCH BOX TO DESIGN A CONNECTION NETWORK OF P1M2,P2M4,P3M1,P4M3. 2 WHAT IS DS-LINK? 3 WHAT IS DIFFERENCE BETWEEN ME

12、SSAGE-PASSING AND SHARED MEMORY? 4 WHAT ARE WORMHOLE COMMUNICATION AND ITS PERFORMANCE? 5 Explaining that Myrinet,HiPPI,FDDI,ATM,SCI,and 100BaseT。,6采用蟲洞尋徑的超立方體多機系統(tǒng)中，如果相鄰節(jié)點間有一對方向相反的單向通道，試證明在該系統(tǒng)上實現(xiàn)立方體編碼下的尋徑不會死鎖。(In a supercube network with wormhole communication,if there is a pair opposite one-way lin

13、k between any two neighbour nodes,please try to proof that the network can not be deadlocked when looking for path),自由任選（課外大作業(yè)）,任選一個具有代表性的BENCHMARK小程序,要求: 1) 請分析其源程序,寫出其中的數(shù)據(jù)執(zhí)行流程(包括:數(shù)據(jù)類型、具體值和時序關(guān)系); 2) 在1)的基礎(chǔ)上,對其目標代碼再進行數(shù)據(jù)流分析，寫出其中的數(shù)據(jù)執(zhí)行流程(包括:數(shù)據(jù)類型、具體值和時序關(guān)系);觀察并寫出與1）中間結(jié)果的差別表。 3）在1）和2）基礎(chǔ)上，對其數(shù)據(jù)執(zhí)行的流程進行數(shù)據(jù)預(yù)取優(yōu)化

14、,并在具有CACHE的計算機上進行實際求解。要求給出具體數(shù)據(jù)預(yù)取的優(yōu)化方法和改進的測試時間效果表。 （可加5-10分）,用PVM或MPI并行軟件工具編寫一個并行程序，要求至少有2或3個并行任務(wù)進行協(xié)同求解某問題。 What are Architecture？What areKey Tech and Theory？What areMeeting Problems？ 1、Earth-Simulator 2、Blue Gene 3、Beowulf With PoPC Cluster 4、Grid Computing 5、Pervasive Computing 6、P2P Computing 7、Sp

15、ecial PoPC Cluster Such as Web Cache Cluster,作業(yè)homework1,1 H0(n)=nmH0(-n)/(-)+n(1-m)H0; Hc(n)=nmH0(-n)/(-)+n(1-m)H0 please delete the and by using ,then draw the function figure when m=0.5, =0.2(畫出其函數(shù)關(guān)系圖),Homework2,MESI Protocol,Can you fill the states?,關(guān)于可用性中的檢查點問題 CHECKPOINT(a,b,c),可在內(nèi)核、庫、應(yīng)用程序三級發(fā)生

16、；,a,b,d,c,x,y,z,P,Q,R,Process,一致性快照Checkpoint Consistency Snapshot（a-Consistency一致;b-No Consistency不一致）,如果進程之間不存在一個進程的檢查點已接收了消息，而另一進程的檢查點還未發(fā)送這個消息。稱一致性快照。,a,b,x,y,z,P,Q,R,C ?,If there is not the process which receives a message at checkpoint and another process has not sent the message at checkpoint,

17、 then we say that the related checkpoints are the Consistency Snapshot,homework,1. What are the differences about MPP and SMP? Answer: MPP A large parallel processing system with a shared-nothing architecture Consist of several hundred nodes with a high-speed interconnection network/switch Each node consists of a main memory & one or more processors Runs a separate