Energy Practice Worksheet
| # | Answer |
|---|---|
| 1 | Energy is the capacity to do work, often measured in joules, and in computing it refers to the power used by hardware and software systems to perform tasks. |
| 2 | An electricity meter or power meter is commonly used to measure the electrical energy used by a computer system. |
| 3 | Power is the rate at which energy is used or transferred, while energy is the total amount used over time. |
| 4 | The display, CPU, storage devices, and wireless adapter are all major sources; one example is the display. |
| 5 | Because efficient systems reduce electricity costs, improve battery life, and lower environmental impact. |
| 6 | Watts (W) are commonly used to measure electrical power. |
| 7 | 100 watt-hours. |
| 8 | It can dim the display, put the system to sleep, or schedule work to reduce idle power usage. |
| 9 | Idle power is the energy a computer uses when it is on but not doing much work. |
| 10 | Because inactive components still consume power, so disabling them reduces unnecessary energy use. |
| 11 | An energy-efficient algorithm performs the needed task while using less computational resource time and therefore less power or energy. |
| 12 | Lower CPU utilization usually reduces energy consumption because the processor spends less time at high power states. |
| 13 | SSDs usually consume less energy because they have no moving mechanical parts. |
| 14 | Data centers run many servers and cooling systems, so they consume large amounts of electricity and have significant operating costs. |
| 15 | PUE stands for Power Usage Effectiveness. |
| 16 | A lower PUE indicates better energy efficiency in a data center. |
| 17 | Reducing background activity, batching network requests, or optimizing code to run faster can reduce energy use. |
| 18 | Because the display and graphics system must work more often, which increases power usage. |
| 19 | Wireless communication requires radio transmission and reception, which use electrical power. |
| 20 | For many programs, longer computation time leads to more energy use because the hardware remains active for a longer period. |
| 21 | 3 kilowatt-hours. |
| 22 | DVFS is a technique that changes a processor's voltage and frequency to balance performance and energy use. |
| 23 | Lower frequency can reduce power consumption, especially when combined with lower voltage. |
| 24 | The trade-off is that energy savings may come with reduced performance or slower responsiveness. |
| 25 | Parallel computing can use more hardware at once, and extra coordination or overhead may increase energy use. |
| 26 | Run the program on measured hardware and record power use over time using a power meter or energy monitoring tool. |
| 27 | Sleep mode reduces power consumption while keeping the system state ready for quick wake-up. |
| 28 | They can reduce cache misses and unnecessary processor work, lowering total energy use. |
| 29 | Memory operations often cost significant power, so fewer accesses can reduce energy consumption. |
| 30 | Load balancing can consolidate work onto fewer servers so unused servers can be turned down or powered off. |
| 31 | Virtualization lets multiple workloads share fewer physical machines, improving server utilization and reducing idle hardware. |
| 32 | Cooling systems consume electricity and add to the overall energy demand of the facility. |
| 33 | Its energy use is reduced by about half because energy equals power multiplied by time. |
| 34 | It reduces radio use on the client and processing, storage, and network traffic on the server, lowering power consumption on both sides. |
| 35 | Energy-aware optimization considers power and energy along with speed, while performance-only optimization focuses mainly on execution time. |
| 36 | To improve efficiency, estimate costs, and make scheduling decisions that reduce total power consumption. |
| 37 | Large models can require heavy computation and memory usage, which increases energy consumption. |
| 38 | It can require fewer calculations and less memory access, lowering the energy needed for training or inference. |
| 39 | The team can profile the application, remove unnecessary work, optimize hotspots, and test energy use on real devices. |
| 40 | Because software design, algorithms, operating systems, networking, and hardware all affect how much energy a computing system uses. |