Everything can interfere with performance, but here comes the question of need or not. It certainly influences, but very little and proportional. Actually proportional amortized because some transactions will not make a difference how many columns have.
There are many much more important reasons that will influence performance.
Don’t worry about it. If you need a lot of columns, put them in. If you don’t have a reason for them to exist, you don’t have to create them. If you try to gather data in a column to have fewer columns, then I would say that it will further worsen performance and create a lot of confusion. Don’t do this.
Even if it has many columns, read only the ones you need in that query. This minimizes the impact.
Having a lot of data can indirectly influence performance because it can fill up the cache faster and have to access disk more often. But note that the size of the total data is more important than the number of columns.
as the bigown pointed out, the best strategy for inheritance, will depend on your need, in any case I advise you to read this comparative between TPH, TPT and TPC: How to Choose an Inheritance Strategy
– Tobias Mesquita
@Tobymosque opa.. then, related to inheritance in EF...I never used, until I tried once... But I preferred not to use to not get engaged...
– Rod
although the article refers to tips regarding EF, this chapter in specific applies to which wants ORM, it addresses precisely how the strategy of creating the tables affects the performance, integrity, flexibility and aesthetics... due to some limitations of some Orms, it is not interesting to have a table for each object, a single table representing the base object and its derivatives will present a better performance, against starting several tables can improve the validation of the data in the database, flexibility to include subtypes and its aesthetics.
– Tobias Mesquita