Antingen stödjer din webbläsare inte javascript, eller är javascript inaktiverat. Denna webbplats fungerar bäst om du aktiverar javascript.

Thomas Lundqvist

Universitetslektor

Anställd på Avdelningen för matematik, data och lantmäteri.

Forskar inom området teknik, elektroteknik och elektronik.


Forskningsintresse/forskningsområde

Forskningsfokuset de senaste åren har varit på effektivare beräkningar i datorkluster. Moderna processorer har många kärnor som kör flera jobb samtidigt. Med hjälp av rätt kunskap om applikationerna man vill köra och hur dessa använder olika resurser i ett datorsystem kan man placera jobben smartare på processorkärnorna och därmed effektivisera körningarna så att vi kan köra fler jobb på samma tid eller helt enkelt utnyttja hårdvaran bättre. Tidigare har jag även forskat om verifiering av tidskrav i hårda realtidssystem. Detta var det ämne jag doktorerande inom 2002. Jag byggde här ett nytt verktyg som kunde analysera datorprogram och hur lång tid de kräver för att köras.

Undervisning/handledning

Jag undervisar främst i kurser som handlar om datorarkitektur, operativsystem och programmering. Dessutom är jag ansvarig för exjobbskursen kopplad till kandidatexamen i datateknik. 

Nyckelord

Datorkluster, HPC, multicore.

Publikationer


Node Sharing for Increased Throughput and Shorter Runtimes : an Industrial Co-Scheduling Case Study

Publicerat

Node Sharing for Increased Throughput and Shorter Runtimes : an Industrial Co-Scheduling Case Study

The allocation of jobs to nodes and cores in industrial clusters is often based on queue-system standard settings, guesses or perceived fairness...

Disallowing Same-program Co-schedules to Improve Efficiency in Quad-core Servers

Publicerat

Disallowing Same-program Co-schedules to Improve Efficiency in Quad-core Servers

Programs running on different cores in a multicore server are often forced to share resources like off-chip memory,caches, I/O devices, etc. This...

Flipping the Data Center Network : Increasing East-West Capacity Using Existing Hardware

Publicerat

Flipping the Data Center Network : Increasing East-West Capacity Using Existing Hardware

In today's datacenters, there is an increasing demand for more network traffic capacity. The majority of the increase in traffic is internal to the...

Increasing Throughput of Multiprogram HPC Workloads : Evaluating a SMT Co-Scheduling Approach

Publicerat

Increasing Throughput of Multiprogram HPC Workloads : Evaluating a SMT Co-Scheduling Approach

Simultaneous Multithreading (SMT) is a technique that allows formore efficient processor utilization by scheduling multiple threadson a single...

Initial Formulation of Why Disallowing Same Program Co-schedules Improves Performance

Publicerat

Initial Formulation of Why Disallowing Same Program Co-schedules Improves Performance

Co-scheduling processes on different cores in the same server might leadto excessive slowdowns if they use the same shared resource, like a memory...

Thing-to-thing electricity micro payments using blockchain technology

Publicerat

Thing-to-thing electricity micro payments using blockchain technology

Thing-to-thing payments are a key enabler in the Internet of Things (IoT) era, to ubiquitously allow for devices to pay each other for services...

Terrible Twins : A Simple Scheme to Avoid Bad Co-Schedule

Publicerat

Terrible Twins : A Simple Scheme to Avoid Bad Co-Schedule

Co-scheduling processes on different cores in the same server might lead to excessive slowdowns if they use a shared resource,like the memory bus. ...

Addressing characterization methods for memory contention aware co-scheduling

Publicerat

Addressing characterization methods for memory contention aware co-scheduling

The ability to precisely predict how memory contention degrades performance when co-scheduling programs is critical for reaching high performance...

A methodology for estimating co-scheduling slowdowns due to memory bus contention on multicore nodes

Publicerat

A methodology for estimating co-scheduling slowdowns due to memory bus contention on multicore nodes

When two or more programs are co-scheduled on the same multicore computer they might experience a slowdown due to the limited off-chip memory...

A light-weigh non-hierarchical file system navigation extension

Publicerat

A light-weigh non-hierarchical file system navigation extension

Drawbacks in organising and finding files in hierarchies have led researchers to explorenon-hierarchical and search-based filesystems, where file...