-
Notifications
You must be signed in to change notification settings - Fork 0
Home
NORGES TEKNISK-NATURVITENSKAPELIGE UNIVERSITET (NTNU) Fakultet for informasjonsteknologi og elektroteknikk Institutt for datateknologi og informatikk
Prosjektrapport Lagerhierarki - Datamaskinarkitektur
Emne: IDATT1004 – Teambasert Samhandling Semester: Høst 2025 Gruppe 1: • Johanna Krogh Andersen • Per Eric Trapnes • Vanessa Efakor Akosua Amedi Veileder: Geir Ove Rosvold Innleveringsdato: 10 Oktober 2025
This project aimed to demonstrate understanding of the concept of memory hierarchy in computer architecture by producing an educational video. The task required presenting the topic creatively yet technically accurately while explaining how different storage technologies work together in a computer system. The video explored the relationship between cache, RAM, SSD, and hard disk in terms of speed, capacity, and cost. The report documents both the final product and the working process. It also reflects on what was learned, the challenges faced, and how collaboration contributed to the outcome. The project provided experience in teamwork, communication, and using digital tools to organize and share work efficiently.
The memory hierarchy is a fundamental concept in computer architecture. It describes how data is stored and accessed at different levels depending on the required speed and the cost of each storage technology. Fast memory is expensive and small, while slower memory offers larger capacity at lower cost. This hierarchy aims to create an optimal balance between performance and efficiency. For this project, the group designed a short video that explains memory hierarchy in a relatable way. The video combines theory with practical analogies to show how a computer's central processing unit (CPU) communicates with different storage levels. The content was based on course materials and supported by external sources such as Uniweb and academic texts.
The video begins by posing a simple question: "Have you ever wondered why your computer uses many different components to store data?" It then introduces the hierarchy visually and conceptually, starting from the CPU and cache to RAM and SSD. The presentation uses an analogy of the CPU as a person retrieving information from different locations — the hand (cache), a book on the table (RAM), and a library (SSD). The video also includes a price and speed comparison between the storage levels: • Cache: approximately 100 NOK per MB • RAM: approximately 0.2 NOK per MB • SSD: approximately 0.01 NOK per MB This exemplifies how performance and cost are traded off. SSDs offer large capacity at a significantly lower cost than cache memory, which is high speed but costly. The video concludes by demonstrating how computers can function more effectively thanks to this hierarchy structure, which guarantees that active data is kept in fast memory and inactive data is kept in more affordable storage.
The filming took place in a park in front of the main school library using a mobile phone. Roles were divided as follows: • Johanna: Responsible for the script's theoretical content and overall structure. • Vanessa: Co-writer and actor explaining the RAM and SSD concepts. • Per Eric: Writer and the editor responsible for the final video production. In order to make the explanation intriguing, Per Eric edited the video and added transitions, background music, and captions. Each participant was able to view and evaluate the materials since the script and video were distributed via Messenger and a shared Word document.
A traditional memory hierarchy consists of several levels, from the CPU registers and cache at the top to magnetic disks or cloud storage at the bottom. According to Patterson and Hennessy (2017), this structure ensures that frequently used data is kept close to the processor, minimizing delays caused by slower storage. Modern systems have introduced new technologies that extend this hierarchy. For instance, when compared to older SATA-based drives, NVMe SSDs drastically lower latency (Kingston Technology, 2023). Similarly, high-bandwidth DRAM and 3D-stacked cache enhance performance in sophisticated CPUs and GPUs. These developments show how memory hierarchies are still changing to accommodate rising performance requirements. The group's comparison in the video shows that while cache provides nanosecond access times, RAM operates in tens of nanoseconds, and SSDs in milliseconds. This exponential difference highlights why a layered approach is essential. The CPU constantly waits for data without intermediate storage like cache and RAM, leading to poor system performance.
At the start of the project, the team decided on a clear task division. Johanna maintained technical accuracy while concentrating on theoretical research. Vanessa acted in the video, contributed creative ideas, and helped with the writing process. Eric was in charge of post-production and video editing. Every member took part in the filming and made an equal contribution to the decision-making process.
For this project, the team didn't use GitLab during the work process but for post-production. Rather, Messenger was used for rapid communication and shared Word documents were used for the script and report so that drafts and the report could be uploaded. However, the manuscript, the video and this report have been uploaded to GitLab with all necessary information. This system operated well because of the small group size and the straightforward project requirements.
A number of difficulties arose throughout the procedure. When filming various scenes, the group first had trouble with timing and synchronisation. Simplifying complex technical ideas without sacrificing accuracy presented another difficulty. To get around this, the team chose examples and analogies that are familiar to everyday users. The project was successfully finished by the deadline in spite of these difficulties. This project illustrated the value of communication, patience, and preparation for teamwork.
This project provided a practical understanding of how computer memory systems are organized and why a hierarchical model is necessary. Translating theoretical knowledge into a creative format helped reinforce learning. It also demonstrated how digital collaboration tools can replace traditional programming platforms such as GitLab for simple multimedia projects. Working in a team developed important skills in project coordination, technical communication, and time management. The group learned to adapt to each member's strengths and schedule while clearly focusing on the final goal. Overall, the assignment combined creativity with technical accuracy, deepening the group's understanding of core computer architecture principles.
- References
Kingston Technology. (2023). SSD vs. HDD: What’s the difference? Retrieved October 8, 2025, from https://www.kingston.com/en/blog/pc-performance/ssd-vs-hdd
OpenAI. (2025). ChatGPT (Version GPT-5) [Large language model]. https://chat.openai.com/
Patterson, D. A., & Hennessy, J. L. (2017). Computer Organization and Design RISC-V Edition: The Hardware/Software Interface. Morgan Kaufmann.
Uniweb. (n.d.). RAM and CPU. Retrieved October 8, 2025, from https://www.uniweb.no/webhotell/ram-og-cpu/
Wikipedia contributors. (2025, September). Memory hierarchy. In Wikipedia, The Free Encyclopedia. Retrieved October 8, 2025, from https://en.wikipedia.org/wiki/Memory_hierarchy
(Kamera: De tre står i et klasserom eller laboratorium. Per Eric står midt i, Vanessa og Johanna på hver side.)
Per Eric: Har du noen gang lurt på hvorfor PC-en din bruke så mange forskjellige component for å lagre data?
Johanna: Eller hvorfor 1TB SSD er billig mens 1TB RAM er så dyrt?
Vanessa: Det handler om lagerhierarki! Hvordan data flyttes mellom forskjellige lager i datamaskinen.
Johanna: Lagerhierarki er et system som organiserer dataminnet og lagringsenheter etter tilgangshastighet, kapasitet og kostnad. Øverst fra de raskeste, dyreste og minste komponentene Til nederst der de tregeste, billigste og største.
Vanessa (peker på Per Eric): Per Eric, du er CPU-en vår.
(Kamera: Zoomer inn på Per Eric.)
Per Eric (spiller stolt): Jeg, CPU er serverens hjerne som utfører alle beregninger, kommandoer og beslutninger som trengs for at en nettside skal fungere. Jeg er kjempe rask I hodet, men kjempe treig på fots! Jeg mottar instruksjoner fra RAM-minnet, tolker hva som er oppgaven og utfører deretter oppgaven. CPU (jeg) kan også håndtere flere oppgaver samtidig. Hastigheten måles i GHz og angir hvor mange operasjoner per sekund prosessoren kan utføre.
Johanna (peker på Per Eric sin hånd): Her har vi cache, den superraske lagringen som ligger rett ved siden av hjernen. Den har bare plass til litt data, men er ekstremt rask.
Per Eric : Så jeg kan bare lagre det jeg skal bruke akuratt nå i cache. (Kamera panorerer til en bok på bordet.)
Vanessa: Denne boka her er RAM. Den har mye mer plass enn cache, men er litt tregere og lengre unna. RAM brukes som et lager for midlertidig informasjon som serveren trenger rask tilgang til. Denne hastigheten er superrask i forhold til en harddisk eller SSD (som vi snart kommer tilbake til).
Johanna: Et eksempel er når man besøker en nettside så henter serveren nødvendige filer og data fra harddisken og plasserer dem i RAM-minnet for øyeblikkelig bruk. Dersom serveren måtte hente alt fra harddisken hver gang hadde nettsiden lastet mye tregere. Det som har blitt lagret i RAM-minnet tømmes når serveren slås av.
Per Eric (later som han rekker ut hånda og åpner boka): Så hvis jeg ikke finner informasjonen i hånda mi (cache), må jeg strekke meg til boka (RAM). Det tar litt lenger tid, men det går fortsatt ganske fort. (Kamera beveger seg mot et bokskap eller bibliotek i bakgrunnen.)
Johanna: Og der borte, det er SSD. Der lagres alt du ikke bruker akkurat nå, men som du vil hente senere. SSD har masse plass, men den er mye tregere.
Vanessa: CPU, hva er hovedstaden i Norge?
Per Eric (ser på hånda si først): Hmm… ikke her i cache’n. (Ser i boka på bordet) Aha! Her i RAM står det: “Oslo!”
Vanessa: Perfekt. Du trengte ikke gå helt til SSD.
Johanna: Men hvis det ikke hadde stått i boka, måtte du gått helt bort til biblioteket for å finne svaret. Det tar mye lengre tid, selv for en rask CPU!
Johanna: Men hva skjer hvis datamaskinen ikke har cache eller RAM? Bare CPU og SSD?
Per Eric (sukker dramatisk): Da må jeg… gå… helt bort til biblioteket hver eneste gang!
Vanessa (ler): Og det tar lang tid, fordi SSD er mye tregere å lese fra. CPU må vente hele tiden og maskinen føles treg.
Johanna: Akkurat derfor finnes cache og RAM, for å redusere ventetid. (Kamera zoomer ut og de tre står samlet igjen.)
Vanessa: Men hvorfor har vi ikke bare kjempemasse cache da?
Per Eric : Fordi cache er ekstremt dyrt å lage. Det koster 100kr per mb Det er superraskt, men tar plass på selve CPU’en.
Johanna: RAM er billigere med ca 0.2kr per mb og du får mye mer plass for prisen, men den er litt tregere.
Vanessa: Og SSD er billigst av alt, på bare 0.01kr per mb , men den er også mye tregere enn både cache og RAM.
Per Eric: Så, kort sagt. Jo raskere lagring, jo dyrere.
Johanna: Ja det stemmer, men med å bruke spå hva som cpu’en trenger i neste omgang kan vi danse linjen mellom hastighet og økonomi.
Alle sammen: Det er lagerhierarki!
(Kamera zoomer ut mens de ler, musikk spiller ut.)