I suspect the slowdown when opening a 100mb .txt file is not really in syntax parsing - it's "Plain Text", no syntax definition attached.
I was unclear in the description - my datadump gb's tend to be sql or xml.
Turning syntax off helps; simply it's a real pain to switch it on/off manually every time, as I'm working both with small and large files of the same type. I'd like the software to do that automagically based on file size or loading time estimates.
If I work with many small SQL/XML/whatever files, then I want it to look pretty; but if I open a 100+ mb datadump, then Sublime should be able to open it as plain text without waiting for half an hour.
There are valid text files floating around with sizes above 1gb - why shouldn't I be able to open it, run searches and browse the contents? The computing power can do it, you can map it to memory if needed, I can spare 10gb ram if working on that size of data; but it needs to work reasonably fast.