The exploratory thesis that I produced during my internship at the Infineon Technologies headquarters in Padua had the objective of studying analysis methodologies, useful in order to try to optimize as much as possible the cost of tests for non volatile memories (NVM). I followed two parallel paths, getting different considerations but with the same goal. The first approach is related to the test sequence where, through a statistical survey, I was able to determine if there were relationships between specific tests. I therefore worked with these tests to stress them as much as possible in order to find time-consuming cases. The goal is to find a trade-off between time spent and devices, in ordere to find a local optimum and increasing the yield. In the second approach I worked directly on the test sequences, trying to optimize them and therefore reduce the execution times as much as possible. The evaluation is made on the execution time because, if it was not too high, that time could be "invested" by inserting evaluation points, useful to improve the stability of the test and consequently to improve its yield.
La tesi di carattere esplorativo che ho prodotto durante il mio periodo di tirocinio presso la sede di Padova di Infineon Technologies ha avuto come obiettivo di studiare delle metodologie di analisi, utili allo scopo di cercare di ottimizzare il costo dei test per memorie non volatili (NVM). Ho seguito due strade parallele, ottenendo considerazioni diverse ma con lo stesso obbiettivo. Il primo approccio è relativo alla sequenza di test dove, tramite un'indagine statistica, ho potuto determinare se ci fossero relazioni tra test specifici. Ho poi lavorato con questi test per stressarli il più possibile, in modo da trovare delle casistiche onerose in termini di tempo. Lo scopo è quello di trovare un trade-off tra tempo speso e dispositivi, trovando un ottimo locale e andando ad incrementare il rendimento. Nel secondo approccio ho lavorato direttamente su sequenze di test, cercando di ottimizzarle e dunque ridurre il più possibile il tempo di esecuzione. La valutazione viene fatta sul tempo di esecuzione poichè, se non fosse troppo elevato, si potrebbe "investire" quel tempo inserendo dei punti di valutazione, utili per migliorare la stabilità del testing e di conseguenza migliorandone il rendimento. Per portare avanti il tutto mi sono servito di diversi strumenti interni.
Concetto e definizione del design dell'esperimento per ottimizzare il costo di test (rendimento e tempo di test) per un programma di test NVM
BOMBEN, RICCARDO
2021/2022
Abstract
The exploratory thesis that I produced during my internship at the Infineon Technologies headquarters in Padua had the objective of studying analysis methodologies, useful in order to try to optimize as much as possible the cost of tests for non volatile memories (NVM). I followed two parallel paths, getting different considerations but with the same goal. The first approach is related to the test sequence where, through a statistical survey, I was able to determine if there were relationships between specific tests. I therefore worked with these tests to stress them as much as possible in order to find time-consuming cases. The goal is to find a trade-off between time spent and devices, in ordere to find a local optimum and increasing the yield. In the second approach I worked directly on the test sequences, trying to optimize them and therefore reduce the execution times as much as possible. The evaluation is made on the execution time because, if it was not too high, that time could be "invested" by inserting evaluation points, useful to improve the stability of the test and consequently to improve its yield.File | Dimensione | Formato | |
---|---|---|---|
Bomben_Riccardo.pdf
accesso riservato
Dimensione
4.52 MB
Formato
Adobe PDF
|
4.52 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/11545