Programming Parallel Computers
The lecture videos are available both on Panopto and on YouTube, in up to 4K resolution, with English and Finnish subtitles. The slides are also available in the PDF format.
Lectures
YouTube playlist with all parts of the lecture.
Part 1A: What is this course about? · Why parallelism? (14 min)
Part 1B: Course practicalities (10 min)
- Lecture on Panopto
- Lecture on YouTube
- Slides
- This video introduces the submission system we use in this course, and also explains the basic principles related to e.g. collaboration. Please note that the video was prepared in 2021, and it also refers to some details of the previous implementation of the course. Please see the MyCourses page for the latest updates on the course arrangements in 2026. In particular, in 2026 we will use Sisu for course registration, and our lectures and exercise sessions will be held in person on campus.
Part 1C: Sample application · Memory access pattern (14 min)
Part 1D: Instruction-level parallelism (12 min)
Topics covered
- Course practicalities and the submission system
- Chapter 1: why parallelism
- Chapter 2, V0: baseline implementation
- Chapter 2, V1: linear reading
- Chapter 2, V2: instruction-level parallelism
Terminology
| English | Finnish |
|---|---|
| accumulate | koota (yhteen) |
| arithmetic operation | aritmeettinen operaatio |
| array | taulukko |
| benchmark | suorituskyvyn mittaus, suorituskykymittaus |
| benchmarking | vertailuanalyysi |
| bottleneck | pullonkaula |
| cache memory | välimuisti |
| cache memory hierarchy | välimuistihierarkia |
| chain | sarja |
| clock cycle | kellojakso |
| computational problem | laskennallinen ongelma |
| core | ydin |
| cost | hinta, kustannus |
| CPU | prosessori, suoritin |
| CPU core | prosessoriydin, suoritinydin |
| CPU register | rekisteri |
| dependency chain | riippuvuusketju |
| execution unit | suoritusyksikkö |
| floating-point multiplication | liukulukulaskenta, (yksi) liukulukulasku |
| GPU | grafiikkaprosessori, grafiikkasuoritin |
| hardware | laitteisto |
| in parallel | rinnakkain |
| input data | syöttötiedot, syötedata, datasyöte |
| input matrix | syötematriisi |
| input size | syötekoko |
| instruction level parallelism | käskytason rinnakkaisuus |
| interleaved | limittynyt |
| latency | latenssi |
| linear reading | lineaarinen lukutapa, lukeminen |
| linked list | linkitetty lista |
| machine language instruction | konekielinen käsky |
| machine language operation | konekielinen operaatio |
| main memory | keskusmuisti |
| massively parallel | massiivisesti rinnakkainen |
| matrix | matriisi |
| memory access pattern | muistista hakeminen |
| memory cell | muistisolu |
| multicore | moniytiminen |
| multithreading | usealla säikeellä tehtävä laskenta |
| operation | käsky, operaatio |
| optimization flag | optimointilippu |
| parallel computer | rinnakkaistietokone |
| parallel performance | rinnakkaisten prosessien suorituskyky |
| parallel programming | rinnakkaisohjelmointi |
| parallel throughput | rinnakkainen läpäisy |
| parallelism | rinnakkaisuus |
| performance | suoritusteho, suorituskyky |
| pipeline | liukuhihna |
| pipelining | liukuhihnoitus |
| processing power | laskentateho |
| processor | prosessori, suoritin |
| result matrix | tulosmatriisi |
| sequential | peräkkäin |
| sequential computer | sekventiaalisesti toimiva tietokone |
| sequential operation | perättäinen operaatio |
| sequential performance | peräkkäinen suorituskyky, peräkkäisten prosessien suorituskyky |
| sequential throughput | peräkkäinen läpäisy |
| simultaneously | samanaikaisesti |
| single-precision floating point | yksinkertaisen tarkkuuden liukuluku |
| superscalar | superskalaarinen |
| theoretical maximum performance | teoreettinen maksimisuorituskyky |
| throughput | suoritusteho, läpäisy |
| transistor | transistori |
| transpose | transpoosi |
| unroll (a loop) | purkaa silmukka |
| vector operation | vektorioperaatio |
| vectorization | vektorointi |
