CPU Architecture and the Fetch–Decode–Execute Cycle: Question 3
Syllabus 3.1
A gaming café is comparing two PCs for customers who edit video. PC Nimbus has a quad-core processor, a clock speed of 3.0 GHz and a 6 MB cache. PC Zephyr has a dual-core processor, a clock speed of 4.2 GHz and a 2 MB cache.
(a) The owner runs a test where each PC exports a video while also scanning for viruses in the background. PC Nimbus finishes both tasks noticeably sooner than PC Zephyr. Explain, in terms of cores, why PC Nimbus can handle these two tasks faster. [2]
(b) When only the video export is run, with no background scanning, PC Zephyr completes the export slightly faster than PC Nimbus, despite having fewer cores. Explain, in terms of clock speed, why this can happen. [2]
(c) Both PCs are then tested exporting a long video built from many small effect files that are reused throughout the project. PC Nimbus completes this export faster than its clock speed alone would suggest. Explain, in terms of cache, why a larger cache can give this advantage. [3]
Show worked solution Hide worked solution
Worked solution
Part (a): Why more cores help with two simultaneous tasks
A core is a processing unit capable of independently fetching, decoding and executing instructions. With four cores, PC Nimbus can dedicate one core to the video-export instructions and a separate core to the virus-scan instructions, so both run truly in parallel. PC Zephyr’s two cores give it less spare capacity: when both tasks are demanding CPU time, they end up competing for the same limited number of cores, so one task is delayed while the other is serviced, slowing the combined job down.
Part (b): Why higher clock speed helps with a single task
Clock speed measures how many fetch-decode-execute cycles a core can get through each second, a 4.2 GHz core completes more cycles per second than a 3.0 GHz core. When there is only one task running, the number of cores stops being the limiting factor, because there is nothing else that needs a second core. In that situation, the PC with the faster per-core cycle rate, Zephyr, can push through the export instructions more quickly, even with fewer cores in total.
Part (c): Why a larger cache helps with reused data
Cache sits between the CPU and RAM and is far faster to access than RAM, but much smaller. When a task keeps reusing the same small files, as this project does with its effect files, a larger cache (Nimbus’s 6 MB versus Zephyr’s 2 MB) can hold more of that reused data at once. Every time the CPU needs one of those effect files again, it can retrieve it almost instantly from cache instead of waiting for a much slower round trip to RAM. Over many repeated fetches this adds up to a real time saving, on top of whatever the clock speed alone would produce.
Final answers
- (a) More cores let Nimbus run the export and the virus scan in true parallel; Zephyr’s fewer cores force the two tasks to compete for CPU time.
- (b) Zephyr’s higher clock speed completes more cycles per second, which matters once cores are no longer the bottleneck for a single task.
- (c) Nimbus’s larger cache stores more of the reused effect files close to the CPU, avoiding repeated slow trips to RAM.