Given the interest in diversity, it's perfectly reasonable to rule out the probability that there were any because the interest in diversity would have almost invariably lead to them being given hiring priority if they applied.

You can keep on repeating the same rubbish: it doesn't become true if you say it enough.

This is not about favouring unqualified "minorities", it's about actually getting them to apply in the first place.

Oh and equal opportunity so that when they do apply they're not being judged based on race/gender/whatever and instead on actual ability.

If

You're entire premise is based on an unsupported supposition, mixed with other random crap likethe feeble attempt at some sort of racism-reflection.

Rcruiting through peer networks (what a lot of tech companies do) tends very strongly to hit exactly the same demographic they've already got. FBs numbers are below the numbers on the courses at uni and in the industry so it implies strongly that their recruiting message isn't getting outside a rather select demographic.

IOW there are talented people who they're never going to recruit because their recruitment strategy is flawed.

I don't think that pointing out a fact to another person falls into "calling yourself something" .

Ah, so it's a fact that you are/were a C++ guru, eh? So my man, what precisely defines guru that makes such thing a fact? Calling yourself a guru is akin to insisting your nickname is Mr Awesome.

I don't think he's an intel fanboi as much as an Intel hater, which is fine, because they're pretty despicable. They're crooks but the legal system seems to love large companies so, for example when they dealt an illegal yet crippling blow, they got away with a $1bn payoff which is certainly less than they've made from their illegal activities.

When fines for bad behaviour have a strongly positive ROI, then there's something deeply broken with the system.

It's also funny that on Linux, with fully open benchmarks on phoronix, the AMD chips trade blows with the Intel ones and the top end ones of each are actually pretty close, with AMD being a bit slower on average than the top intel ones, but not far off. You don't have to be a fanboi to read the benchmark results.

These days I buy AMD if at all possible because they're fast in most cases and I'd rather not give my money to a crook if I can avoid it.

It doens't have 8 separate 256 bit FPUs. It can however split the 256 bit ones into separare 128 bit ones. So unless you're hammering on AVX instructions, then it effectively has one FPU per core.

Um, that's utter crap.

No it ain't. AMD at that point had an actually scalable architecture using hyper transport and could scale in multi socket boxes way, way better than Intel. It made a huge difference.

I never call myself a master in martial arts, however I am ...

you just did.

Apache can serve webdav shares. It can back them with SVN.

WebDAV can be mounted as a FS by all major OSs, and SVN also has a web based history. You need to provide some auth system to hook into (obviously), but otherwise it works really very well.

You're shifting your claim.

From what to what? You may wish tocheck what my claim actually was.

PTAM is decent C++ code with probably pretty decent performance, but it isn't "high performance numerical code" or even high performance embedded code; it contains none of the performance tuning that such code usually should have.

It's not pure numerics, as you're thinking of them. It's certainly not embedded code. It is certainly high performance however. It is a very CPU bound algorithm and it uses what was at the time state of the art algorithms for pretty much every step. Not only that but very fast implementations of them. It is a very, very well written piece of code.

I think it is dishonest on your part to recommend Eigen, when you know (or ought to know) that it actually isn't high performance

OK, you're determined to completely ignore a large facet of computing. For large amounts of small matrices the fixed sized numerics libraries won't be beaten. I've examined the output of the ASM code. For such things, the cache locality is good and the compiler knows about aliasing. So unless you provide benchmarks, I simply won't believe you.

Also, the eigen benchmarks seem to show it matching BLAS and LAPACK for speed.

and is deeply mired in C++ idiosyncracies.

I think personally, that TooN has a nicer interface than Eigen. I suspect others don't agree because Eigen is more popular. In any case, the idosynchracies are in the function defintion, which frankly doesn't matter nearly as much.

The actual maths code reads like maths. That's the important bit.

Likewise, I think it is dishonest of you to portray C++ as the only high performance computing language in town,

Except I didn't. I said I don't know anyone who uses anything else. I also asserted C++ has the best combination of performance and expressibility. Or course you could write in C or FORTRAN, they're both high performance and Turing complete.

C++ is still more expressive though.

You seem very angry though. I think you should go back and carefully re-read my posts and the context and sift out what I actually said and whatever random things you've hallucinated.

Excuse the lack of quoting, on phone.

First, business what now? There is so much more to C++ than that. I doubt I've encountered one this year. But when I do, they can either be shoved into a container, or instantiated with make share or make unique.

There's no need for new anywhere.

But it seems like you're massively misinterpreting what I say as no news happen under the hood. Certainly they do, but they're all safely hidden away, so lifetime is managed cleanly and safely for you. Not sure how you came to that conclusion, but whatever.

And dude, if you call yourself an ex guru, you're implying that you think you were a guru. It's rather tacky to describe oneself as such, not to mention lacking in credibility. Stroustrup and etc never call themselves gurus.

Where did you force me to do anything at all?

I didn't force you to do anything, and apparently I didn't force you to read my post. To give a sane and coherent answer would have forced you to actually think about what you're saying.

You can't be aresd though which is why you're not answering my question.

I don't care about your friend and I don't bother anybody and they shouldn't bother me.

You care enough to (in the hypothetical case) switch from calling her "her" to "him".

There are males who look female, that's nothing new.

You know, try actually reading the post I made previously.

The i stands for imaginary. Which is really telling about your world...

lol you made a mistake and you're trying to cover it up. The complex field is well accepted and widely used.

Ah, yes, how bigoted and parochial of me...

Bigoted? Perhaps. Flat out dumb-ignorant, yes. If you keep insisting that 1+1 must always equal 2 then yes, bigoted, because it's not true, you've been shown it's not true, much smarter people than you have worked extensively with it and you would in that case keep on insisting on it.

It's kind of funny for you to deny it as you're using it right now even though you don't realise. Modern error correction mostly runs on GF(2), and all modern networking kit relies on modern error correction.

https://en.wikipedia.org/wiki/...

This is the third time you claim evidence exist, but do not quote any. Full of shit much? I think, we are done here...

Apparently the title of a wikipedia page doesn't count according to you. Well, yes, if you make up lies like I didn't present any evidence you can "prove" anything you like. Doesn't make you a raving nutcase though.

One is enough, because it was a counterexample. You claimed, ancients did not "rigidly" define homosexuals nor considered the practice wrong. The quote I provided destroyed both of these claims and your only defence was to accuse of me "cherry-picking" - a line patently idiotic. You made a falsifiable statement and I managed to falsify it... Case closed, you were wrong.

Haaaaa hahahahaha!

That has to be the stupidest thing I've read on the internet today! Congratulations!

A counter example doesn't disprove a counterexample. Mine was a counterexample against yours. You've just shown my counterexample isn't a universal truth, but it doesn't need to be. All I had to do was show you one scoiety where attitudes were wildly different to disprove your claim that society has nothing to do with it. Proving that my counter example doesn't apply to all societies is meaningles.

Yes, "obviously" was my own statement

Yep and completely without evidence. You are so wrapped up in the lens of your own culture that you cannot even see where the error is.

How about you actually read the wikipedia page I quoted at you (which you for some reason denied that I quoted).

I suppose, in some "nuanced" world 0 may equal 1 and result of cos() may reach as high as 5.

cos(2.29243i) = 5

So, uh, yes, the result of cos() can indeed reach as high as 5.

And if you insist that 1+1=2, I invite you to try it on a Galois Field, specifically GF(2) where 1+1=0

It's astonishing how people with such firmly held opinions know they are right and revel in their ignorance.

Nor is it possible for a trait to be both imposed by society and inborn at the same time.

Oh gee, I don't see how both could have influence of varying degrees in people!

Citations?

The wikipedia page on homosexuality in ancient rome and the same for Greece.

I offered two citations from the top of my head - which is two more than you did. Put up or shut up.

You quoted a line from one, but your conclusion of "Obviously..." was in fact yours alone and not a citation.

So, you agree then that, numerical performance of Eigen sucks for large matrices.

Not explicitly, since I've not benchmarked it or read any benchmarks but I'm prepared to believe it. I know that BLAS people put a lot of effort into all sorts of cunning blocking stuff.

For small, fixed size matrices it likely is no better than Fortran either.

Depends on how much hand-coding you're prepared to do. C++ has the advantage that you can bake the type into the type, which aids the compiler.

Syntactically, Eigen is pretty lousy as well: you can write some array expressions, but defining functions on matrices is messy, and slicing and similar operations also are restrictive.

I agree there, which is why I don't use Eigen. I'm holding it up since it's for some reason the most popular library, rather then the best.

and parallelization support in it is close to non-existent.

I don't really care much for parallelisation at that level. For my stuff, large matrices are donw with LAPACK which seems to have it and large amounts of small matrices are solved with a #pragma parallel for

Fortran these days has built in matrices that automatically use the best possible BLAS/LAPACK backends.

Maybe, but a BLAS backend is never going to do anything but a wretchedly slow job on a 2x2, 3x3 or 4x4 matrix :) If you have a few billion of them, then the performance matters.

Anyway, we were talking about the comparison to C, not FORTRAN originally. I don't know all that much about FORTRAN except that C++ does seem more expressive for the bits which aren't purely numerical.

Large parts of the numerical and high performance computing community never touch C++ at all. C++ is mostly popular in computer vision, embedded systems, and some areas of machine learning

It seems pretty popular with the supercomputer folks for new codes. But yes, it is popular for computer vision. I can't imagine something like PTAM being written in anything other then C++. Not only do you need high performance numerics, you need high performance book-keeping and other bits and bobs as well. C++ is the only one that gives you the good performance and expressivity across the board.

Anyway, when it comes to array-banging as opposed to linear algebra, I rather prefer using explicit for-loops to array notation. I can do the array notation---I've spent enough time with MATLAB---but I find it clumsy and one often has to work rather hard to solve the problems compared to a for loop. That's personal preference.

FYI, if you only need to get the best N at the end, it's more efficient to stick everything in a vector and use partial_sort to get the first N. Or nth_element if you don't even need them sortedbut just split into less than Nth and greater than or equal

Depends on the algorith. For a VP-tree or KD-tree, you need the N-th best at every test since you use that to bound the cost and avoid searching as many bins as possible. Without that, you wind up with a linear cost not a logarithmic one.