I read the resumes, and if the code description is interesting, I'll pull up the github repo. I read a lot of them because I do a lot of interviews, and the resumes that get passed to me for interviews all look fantastic and incredibly similar. Because good programmers have remarkably similar resumes as bad programmers -- one was made by sweat & blood, the other with bullshit. It's just text on a page, there's no way of knowing which one's which.
As for their code, I learn a little bit about their programming style -- and it's often unflattering and unhelpful, as these are usually personal projects where people write code below the standard they'd use at the office -- but it tells me little of what parts were hard, what parts involved real insight from the candidate, and what parts were pasted from stack overflow. Given a github URL, how much time do you think it takes to figure out where the interesting parts are? And what's the information payback of that time, given the little information you get about a person's actual capabilities from a bunch of text saved on a server?
I know that everyone's their own special snow flake, but I don't have more than the 45 minutes to give them a fair shake. And it has to be fair - a question that they should probably have enough context to know, and who's answer can't be bull-shitted. I've got 45 minutes to give as objective an interview I can, balancing the Type 1/2 errors of missing someone good, and hiring someone bad.
1 a week was an example of what I was doing when I went interviewing. I mentioned that as slow enough to read up as much as you need. Did you want more? Less?
[Edit: not speaking for my employer, whose name I don't want to mention. ]