stephencanon

159 ms·

1561.
by stephencanon 15y ago
Ok, then you need to ask for a raise. $9/hour is a wage you pay someone who is completely replaceable; and even then only if you're a greedy bastard who doesn't care about your employees. A raise could come in the form of a much higher wag
1562.
by stephencanon 15y ago
The point still stands. Why is a 20-year-old engineering company paying you $9/hour? The coffee shop in town here pays new counter staff $14/hour. Generally speaking, one does not pay someone who is as important to the business as you cl
1563.
by stephencanon 15y ago
> Removing freedom to see the source might not effect peoples choice it's self I agree, but really why should it also effect the quality of the software? We are in violent agreement. Freedom to see the source is entirely orthogonal to
1564.
by stephencanon 15y ago
Yes, that would be a false dichotomy. However, the post you quote does not set it up as a dichotomy. My reading of it is that it asks the question "which of these two things is more important?" I would suggest that it would be better fra
1565.
by stephencanon 15y ago
It also installs "gcc-4.2", which is actually the gcc-4.2 driver.
1566.
by stephencanon 15y ago
This is true of every field. There are vast differences in training, ability, and focus between practitioners, but you're not aware of them because you don't know enough about most fields to be able to tell the difference. Yes, "programmer
1567.
by stephencanon 15y ago
What you describe is pretty common on many parts of the US-Canada border, too. There are towns that are split down the middle, and it's normal for locals to cross the border several times a day without any hassle. When I was growing up (1
1568.
by stephencanon 15y ago
First: not all surgeons start at 5; it varies somewhat with the specialty. The more "hard-core" specialties tend to start earlier in the morning because the cases can run longer (12-16 hour surgeries, while not typical, are not unheard of
1569.
by stephencanon 15y ago
Exactly. Most people are far more distractible in the evening (partially because there are simply more distractions available), so shifting your schedule earlier results in less time spent on distractions.
1570.
by stephencanon 15y ago
My wife is a surgeon. I'm up at 5am with her, drive her to work, then sit down and start working myself around 5:30. I find that I'm fantastically productive from then until about 9 or 10am. I take a long lunch break (and go running or x
1571.
by stephencanon 15y ago
Two reasons: first, the matrix multiply algorithms with exponents less than three do not have the same numerical stability properties, which can introduce subtle bugs into software that was developed with the expectation of a "usual" O(n^3)
1572.
by stephencanon 15y ago
Besides the ATL-prefixed symbols, there are are numerous APL-prefixed symbols, as well as many _block_ symbols which indicate extensive use of GCD in the level 3 routines (and which ATLAS doesn't use).
1573.
by stephencanon 15y ago
The OS X BLAS was once ATLAS all the way down (and I see no reason to believe that any effort was made to hide that fact), but that's no longer true. A glance at nm on Lion shows that there's some ATLAS in there, but quite a lot of not-ATL
1574.
by stephencanon 15y ago
The gemm -> syrk optimization is very common (I would expect most tuned BLAS libraries to take advantage of it -- all that I have inspected do so).
1575.
by stephencanon 15y ago
I'm not sure how meaningful recruiter pings are; before I moved out of the bay area, I was averaging about 2/month from Google, and only once did the recruiter have a clear description of a position that was actually a good fit for my skill
1576.
by stephencanon 15y ago
The median really is a spectacularly bad choice when there are so few data points, and especially when thinking about performance. If A and B get 3x faster, but C gets 3x slower, then your overall workload gets slower, even though the "med
1577.
by stephencanon 15y ago
They conducted a series of experiments, and reported their results. If that's not a "study", I don't know what is. The fact that the authors acknowledge that it's a "game" doesn't change that. What is misleading is the title on the link:
1578.
by stephencanon 15y ago
This is a really terribly done study, with a misleading title to boot. First, even if we take the conclusion as given, "Fortran" isn't faster than "C". The median of one set of programs, compiled with one Fortran compiler with one set of o
1579.
by stephencanon 15y ago
You would keep a frame pointer, which would tell you where to find the caller's stack frame. Most systems already do this, to aid debugging and tracing tools, to support variable-length stack allocations, or for any number of other reasons
1580.
by stephencanon 15y ago
I write high-performance numerical software for a living. There are a lot of baseless claims in this post. You can write high performance software in C, C++, Fortran, Assembly, or a whole host of other languages. There are syntactic reas
1581.
by stephencanon 15y ago
This is just totally false. You can lay out your C arrays in exactly the same way that Fortran does, if you choose to. Or you can do something else, if that will give better performance. Tools are just tools; they don't determine what y
1582.
by stephencanon 15y ago
Underscore-capital has been a reserved identifier prefix for many, many years. It's rather late to begin complaining about it. The side note about name-mangling on some platforms also using a leading underscore only shows that you can use