As­signed to J. Weller.

Bach­e­lor’s The­sis / Stu­dent Re­search Pro­ject

Ab­stract

Ab­stract mod­el­ling of HW/SW sys­tems is a fairly new re­search topic. The goal of this tech­nique is to cap­ture only the es­sen­tial pa­ra­me­ters of soft­ware and hard­ware which in­flu­ence their tim­ing be­hav­ior. One cru­cial as­pect of any HW/SW sys­tem’s tim­ing be­hav­ior is how long a mem­ory in­struc­tion will take. Un­for­tu­nately a com­plete cache sim­u­la­tion can be very ex­pen­sive.

Goal of this stu­dent pro­ject is to im­ple­ment and train a ma­chine learn­ing model to pre­dict the cache hit/miss rate of a trace.

Re­quire­ments

• Python, Scikit, Py­Torch
• Ma­chine Learn­ing Knowl­edge
• Suc­cess­fully at­teded the lec­ture “Grund­la­gen der Rech­ner­ar­chitek­tur” and/or “Par­al­lele Rech­ner­ar­chitek­turen” (op­tional)
• Linux (op­tional)

Con­tact

Jung, Alexan­der

Lübeck, Kon­stan­tin

Bring­mann, Oliver