Choosing the Right Lab Upgrade Under a Fixed Budget
Riverside College's open-access lab has 40 desktop computers used for programming classes. Each machine has a 4-core CPU, 8 GB of RAM, and a 1 TB 5400 rpm hard disk, and both the operating system and the compiler run from that disk. Students report that starting the IDE takes about 90 seconds, switching between a browser and the compiler freezes the machine for several seconds, and the disk activity light stays on almost continuously during lessons. Monitoring shows average RAM use of 7.6 GB during classes, CPU use averaging 35% with short peaks of 70%, and sustained page-file activity. The college can fund one upgrade this year: 16 GB of RAM for every machine, a 500 GB SSD for every machine, or a faster 6-core CPU for every machine. The lab manager wants the largest improvement in student experience per dollar and has asked for a recommendation with evidence.
1. Using the evidence, identify the primary bottleneck and recommend which single upgrade should be bought first. Explain why the other two options would produce smaller gains.
Read the worked answer
The evidence points to the memory and storage paging path, not the CPU. RAM use averages 7.6 GB out of 8 GB, so the system is nearly saturated before caches and applications peak. Sustained page-file activity shows the operating system constantly moving pages between RAM and the 5400 rpm disk. That explains the multi-second freezes when switching applications: the needed pages must be read back from a slow disk. CPU use averages 35%, with peaks of 70%, which leaves substantial headroom, so a 6-core CPU would upgrade a resource that is not the constraint.
An SSD would cut boot and application-load times and would make paging faster, but it would not remove the cause of paging. More RAM removes the need to page in the first place. Therefore buy 16 GB of RAM first. If a second phase is funded, add SSDs, because they improve cold-start time and make any remaining paging less painful. If the only complaint were cold boot time, the SSD would be the better single purchase, but the reported lesson-time freezes point to RAM.
2. A colleague argues that the page file already provides the extra memory, so RAM is unnecessary. Evaluate that argument and recommend what the lab should measure before and after the upgrade.
Read the worked answer
The colleague is confusing capacity with performance. A page file extends virtual address space, but it lives on the same slow hard disk and is orders of magnitude slower than RAM. Heavy use of it is thrashing, not a solution. It can also mask a memory leak and shorten disk life. The argument would only be reasonable if the workload rarely exceeded physical RAM and the disk were fast.
Here, an average of 7.6 GB means many students cross the limit at the same time. Recommend 16 GB of RAM as the single upgrade, with a managed page file kept as a safety net rather than as primary memory. Before and after the change, measure committed bytes, hard page faults per second, disk queue length, IDE launch time, and the number of freeze complaints. If page faults remain high after the upgrade, investigate a memory leak in the IDE or browser extensions rather than buying more hardware. Set a review trigger, for example if average use exceeds 12 GB after the upgrade.