CSCE 614 (Spring 2015) Eun Jung KimComputer ArchitectureHomework # 1 COVER SHEET(Due: Beginning of class on 02/27/15)Name :ID Number :Directions: Write your answers on the sheets provided. Submit with the COVER SHEET. If youneed additional sheets for any of the problem, add as many blank papers as you require. Printyour name clearly. No late homework will be accepted. You are expected to write up yoursolutions on your own, without referring to other students' works or to solutions you may find onthe web. The total score is 170 points. This homework is due at the beginning of class onFriday, Feb 27, 2015.Speedup1. Consider two different implementations, I1 and I2, of the same instruction set. There are threeclasses of instructions (A, B, and C) in the instruction set. The clock rates of I1 and I2 are 6GHzand 3 GHz, respectively. The average number of cycles for each instruction class on I1 and I2 isgiven in the following table: Class CPI on I1 CPI on I2 C1 usage C2 usage C3 usage A 2 1 30% 40% 60% B 3 2 50% 20% 15% C 5 2 20% 40% 25% The table contains a summary of average proportion of instruction classes generated by threedifferent compilers. C1 is a complier produced by the makers of I1, C2 is produced by themakers of I2, and the other compiler is a third-party product. Assume that each compiler uses thesame number of instructions for a given program but that the instruction mix is as described inthe table. (a) Using C1 on both I1 and I2, which machine is faster and by how much? (5)(b) Using C2 on both I1 and I2, which machine is faster and by how much? (5)(c) If you purchase I1, which compiler would you use? (5)(d) If you purchase I2, which compiler would you use? (5)2. Suppose that we support an enhancement mode to a computer by a factor of 20. Enhancedmode is used 60% of the time, measured as a percentage of the execution time when theenhanced mode is in use. Recall that Amdahl’s law depends on the fraction of the original,unenhanced execution time that could make use of enhanced mode. Thus, we cannot directly usethis 60% measurement to compute speedup with Amdahl’s law.a. What is the speedup obtained from the enhanced mode? (5)b. What percentage of the original execution time has been converted to the enhanced mode? (5)3. When parallelizing an application, the ideal speedup is speeding up by the number ofprocessors. This is limited by two things: percentage of the application that can be parallelizedand the cost of communication. Amdahl’s law takes into account the former but not the latter. a. What is the speedup with 8 processors if 90% of the application is parallelizable, ignoring thecost of communication? (5)b. What is the speedup with 8 processors if, for every processor added, the communicationoverhead is 1% of the original execution time. (5)c. What is the speedup with 8 processors if, for every time the number of processors is doubled,the communication overhead is increased by 1% of the original execution time? (5)4. You will run two applications on this dual core, but the resource requirements are not equal.The first application requires 70% of the resources, and the other only 30% of the resources.Assume that when you parallelize a portion of the program, the speedup for that portion is 2.a. Given that 40% of the first application is parallelizable, how much speedup would you achievewith that application if run in isolation? (5)b. Given that 99% of the second application is parallelizable, how much speedup would this application observe if run in isolation? (5)c. Given that 40% of the first application is parallelizable, how much overall system speedup would you observe if you parallelized it? (5)d. Given that 99% of the second application is parallelizable, how much overall system speedup would you observe if you parallelized it? (5)Power5. Cooling is crucial to operate massive number of servers in a data center. Use the followingtable for your power calculations.ComponenttypeProduct Performance PowerProcessorSun Niagara 8-core 1.2 GHz 72-79W peakIntel Pentium 4 2 GHz 48.9 – 66WDRAMKingston X64C3AD2 1GB 184-pin 3.7 WKingston D2N3 1GB 240-pin 2.3 WHard driveDiamondMax 16 5400 rpm 7.0W read/seek, 2.9W idleDiamondMax 9 7200 rpm 7.9W read/seek, 4.0W idlea. A cooling door for a rack dissipates 14 KW. How many servers with an Intel Pentium 4processor, 1 GB 240-pin DRAM, and a single 5400 rpm hard drive can you cool with onecooling door? (5)b. Suppose that you want to build fault-tolerant RAID 1 storage system which doubles the number of disks. Now how many systems can you place on a single rack with a single cooler? (5)c. Suppose that a data center can dissipate a maximum of 300 W per square foot. Given that a server rack requires 15 square feet, how many servers from part (a) can be placed on a single rack, and how many cooling doors are required? (5)6. In order to save power, the servers could be turned off. However, they would take too long torestart in response to more load. Thus, a new system supports “barely alive” state that enables aquick restart but requires 20% of the maximum power while in this state.a. How much power savings would be achieved by placing 60% of the servers in the “barelyalive” state? Compare it to power savings achieved by simply turning off 60% of the servers? (5)b. How much power savings would be achieved by reducing the voltage by 30% and frequencyby 50%? (5)c. How much power savings would be achieved by placing 30% of the servers in the “barelyalive” state and 40% off? (5)Reliability7. Availability is the most important consideration for designing servers. a. We have a single processor with a failures in time (FIT) of 150. What is the mean time tofailure (MTTF) for this system? (5)b. If it takes two days to get the system running again, what is the availability of the system? (5)c. What is the MTTF for a system with 2000 processors? Assume that if one fails, they all fail.(5)8. A single failure does not cause the entire system to crash. Instead, it will reduce the number ofrequests that can be satisfied at any one time.a. Suppose that a company has 20,000 computers, each with a MTTF of 40 days, and itexperiences catastrophic failure only if 1/3 of the computers fail, what is the MTTF for thesystem?(5)b. If it costs an extra $1000, per computer, to double the MTTF, would this be a good