Computer Architecture Exercise 1 Due 02.November 2010, 12:15pm Part 1. Case Study - Chip Fabrication Cost There are many factors involved in the price of a computer chip. New, smaller technologies give a boost in performance and a drop in required chip area. In the smaller technology, one can either keep the small area or place more hardware on the chip in order to get more functionality. In this case study, we explore how different design decisions involving fabrication technology, are, and redundancy affect the cost of chips. 1.1. Table 1 gives the relevant chip statistics that influence the cost of several current chips. In the next few exercises, you will be exploring the trade-offs involved between the AMD Opteron, a single-chip processor, and the Sun Niagara, an 8-core chip. Table 1: Manufacturing cost factors for several modern processors. α = 4. Chip Die Size Estimated defect Manufacturing Transistors (mm 2 ) rate (per cm 2 ) size (nm) (millions) IBM Power5 389 0.30 130 276 Sun Niagara 380 0.75 90 279 AMD Opteron 199 0.75 90 233 a) What is the yield for the AMD Opteron? b) What is the yield for the Sun Niagara processor? c) Why does the Sun Niagara have a worse yield than the AMD Opteron, even though they have the same defect rate? 1.2. You are trying to figure out whether to build a new fabrication facility for your IBM Power5 chips. It costs $1 billion to build a new fabrication facility. the benefit of the new fabrication is that you predict that you will be able to sell 3 times as many chips at 2 times the price of the old chip. The new chip will have an area of 186mm 2, with a defect rate of 0.7 defects per cm 2. Assume the waver has a diameter of 300mm. Assume it costs $500 to fabricate a wafer in either technology. You were previously selling for 40% more than their cost. a) What is the cost of the old Power5 chip? b) What is the cost of the new Power5 chip? c) What was the profit on each old Power5 chip? Coordinator: Dipl.-Ing. Philpp Mahr (Dipl.-Inf. Benajmin Andres) 1
d) What is the profit on each new Power5 chip? e) If you sold 500,000 old Power5 chips per month, how long will it take to recoup the costs of the new fabrication facility? Part 2. Case Study - Power Consumption in Computer Systems Power consumption in modern systems is dependent on a variety of factors, including chip clock frequency, efficiency, the disk drive speed, disk drive utilization, and DRAM. The following exercises explore the impact on power that different design decisions and/or use scenarios have. Table 2: Power consumption of several computer components. Component Product Performance Power Type Processor Sun Niagara 1.2 GHz 72-79W peak Intel pentium 4 2 GHz 48.9-66W DRAM Kingston X64C3AD2 1GB 184-pin 3.7W Kingston D2N3 1 GB 240-pin 2.3W Hard drive DiamondMax 16 5400 rpm 7W read/seek, 2.9W idle DiamondMax Plus 9 7200 rpm 7.9W read/seek, 4W idle 2.1. Table 2 presents the power consumption of several computer system components. In this exercise, we will explore how the hard disc affects power consumption for the system and how to calculate the MTTF. a) Assuming the maximum load for each component, and a power supply efficiency of 70%, what wattage must the server s power supply deliver to a system with a Sun Niagara 8-core chip, 2 GB 184-pin Kingston DRAM, and two 7200 rpm hard drives? b) How much power will the 7200 rpm disk drive consume if it is idle roughly 40% of the time? c) In a single rack, the MTTF of each processor is 4500 hours, of the hard drive is 9 million hours, and of the power supply is 30000 hours. For a rack with 8 processors, what is the MTTF for the rack. 2.2. Your company s internal studies show that a single-core system is sufficient for the demand on your processing power. You are exploring, however, whether you could save power by using two cores. a) Assume your application is 100% parallelizable. By how much could you decrease the frequency and get the same performance. b) Assume that the voltage may be decreased linearly with the frequency. Using the equation Power dynamic = 1/2 Capacity load V oltage 2 Frequency switches, how much dynamic power would the dual-core system require as compared to the single-core system. Coordinator: Dipl.-Ing. Philpp Mahr (Dipl.-Inf. Benajmin Andres) 2
c) Now assume the voltage may not decrease below 30% of the original voltage (minimum 70%of the original voltage). This voltage is referred to as the voltage floor, and any voltage lower than that will lose the state. What percent of parallelization gives you a voltage at the voltage floor. d) Using the equation for the dynamic power Power dynamic, how much dynamic power would the dual-core system require from part a) compared to the single-core system when taking into account the voltage floor. Part 3. Home Work - Chip Fabrication Cost [8 pt.] 3.1. Table 1 gives the relevant chip statistics that influence the cost of several current chips. In the next few exercises, you will be exploring the effects of different possible design decisions for the IBM Power5. a) What is the yield for the IBM Power5? [1 pt.] b) Why does the IBM Power5 have a lower defect rate than the AMD Opteron and the Sun Niagara? [1 pt.] 3.2. It costs $1 billion to build a new fabrication facility. You will be selling a range of chips from the factory, and you need to decide how much capacity to dedicate to each chip. Your Woods chip will be 150 mm 2 and will make a profit of $20 per defect-free chip. Your Markon chip will be 250 mm 2 and will make a profit of $25 per defect-free chip. Your fabrication facility will be identical to that for the Power5. Each wafer has a 300 mm diameter. Assume that the W af er yield = 1 and α = 4. a) How much profit do you make on each wafer of Woods chip? b) How much profit do you make on each wafer of Markon chip? [1 pt.] c) Which chip should you produce in this facility? [1 pt.] d) If your demand is 50,000 Woods chips per month and 25,000 Markon chips per month, and your facility can fabricate 150 wafers a month, how many wafers should you make of each chip to maximize the profit, without taking into account the yield? Part 4. Home Work - Power Consumption in Computer Systems [16 pt.] 4.1. Table 2 presents the power consumption of several computer system components. In this exercise, we will explore how the hard drive affects power consumption for the system. Coordinator: Dipl.-Ing. Philpp Mahr (Dipl.-Inf. Benajmin Andres) 3
a) Assuming the maximum load for each component, and a power supply efficiency of 80%, what wattage must the server s power supply deliver to a system with an Intel Pentium 4 chip, 2 GB 240-pin Kingston DRAM, and one 7200 rpm hard drive? b) How much power will the 7200 rpm disk drive consume if it is idle roughly 60% of the time? [1 pt.] c) Given that the time to read data of a 7200 rpm disk drive will be roughly 75% of a 5400 rpm disk, at what idle time of the 7200 rpm disk will the power consumption be equal, on average, for the two disks? Hint: Describe a system of equations. Take into account the idle time and read/seek time in a time frame. 4.2. Your company s internal studies show that a single-core system is sufficient for the demand on your processing power. You are exploring, however, whether you could save power by using two cores (on one die). a) Assume your application is 80% parallelizable. By how much could you decrease the frequency and get the same performance? b) Assume that the voltage may be decreased linearly with the frequency. Using the equation Power dynamic, how much dynamic power would the dual-core system require as compared to the single-core system? c) Now assume the voltage may not decrease below 25% of the original voltage. This voltage is referred to as the voltage floor, and any voltage lower than that will lose the state. What percent of parallelization gives you a voltage at the voltage floor? d) Using the equation Power dynamic, how much dynamic power would the dual-core system require as compared to the single-core system when taking into account the voltage floor? 4.3. Your company has just bought a new dual Pentium processor, and you have been tasked with optimizing your software for this processor. You will run two applications on this dual Pentium, but the resource requirements are not equal. The first application needs 75% of the resources, and the other only 25% of the resources a) Given that 60% of the first application is parallelizable, how much speedup would you achieve with that application if run in isolation? [1 pt.] b) Given that 95% of the second application is parallelizable, how much speedup would this application observe if run in isolation? [1 pt.] c) Given that 60% of the first application is parallelizable, how much overall system speedup would you observe if you parallelized it, but not the second application? [1 pt.] Your solutions need to be traceable! Coordinator: Dipl.-Ing. Philpp Mahr (Dipl.-Inf. Benajmin Andres) 4
Die yield: Die yield = Wafer yield ( 1 + ) α Defects per unit area Die area (1) α Dies per wafer: Dies per wafer = diameter π (Waver ) 2 2 π Waver diameter (2) Die area 2 Die area Cost per die: Cost per die = Wafer Cost Dies per waf er Die yield (3) Mean Time To Failure: MTTF = 1 Failure rate (4) Amdahl s Law: Speedup overall = 1 (5) (1 Fraction enhanced ) + Fraction enhanced Speedup enhanced Fraction enhanced is the enhanced fraction of the application and Speedup enhanced is the amount of speedup of the enhanced fraction. Coordinator: Dipl.-Ing. Philpp Mahr (Dipl.-Inf. Benajmin Andres) 5