dsharlet

https://github.com/dsharlet/ComputerAlgebra (disclaimer: my project) might do what you want, it lets you manipulate expressions and then compile them to "native" .net functions that can be evaluated with high performance.

As much as I can't stand it either, Trump is the president of the US, and making a lot of waves. It's not surprising that he is dominating the news, and it's not just "political blather".

I really disagree with this. I work on compilers, so I feel for the people who work on GCC, but, I think it is totally unreasonable to expect the average GCC user who might have encountered this bug to be capable of fixing this themselves.

I hacked around in clang for a bit for a compiler project. I think it would have taken me weeks or months to get to the point where I was able to fix the equivalent bug (if it existed) in Clang. Compilers are big, hard to debug projects. And, IMHO, the Clang/LLVM code is much more approachable than GCC. Not all developers are hobbyists that can afford to detour from their work for weeks or months to ramp up on GCC to fix this.

I'd also point out that this is the kind of bug that could exist and be triggering in a lot of code as we speak but no one (besides the author of this blog post) have caught it. I think it's insane to ignore a bug like this.

Maybe not all, but many of those regulations and permits are there for a reason. The vast majority of the population cares way, way more about not having construction in their yard than they do about enabling a competitive market for ISPs.

If you try to dig up all the sidewalks and driveways on any given block block, most of the residents will raise holy hell with their local elected representatives and try to shut you down any way they can. Much of this regulation is in fact the will of the people, and I'm not sure it's wrong. It is a significant disruption to people's property, and it's silly to lay redundant network infrastructure. It's a bad solution to a stupid problem that only exists because of ideological intransigence.

That "rot" likely started centuries before anyone even imagined a computer for an "array" to exist on.

> Fiber I lay in the ground isn't a shared resource

I disagree, and you are trivializing serious issues that lead to this regulation.

There's a good reason there's a lot of regulation in this space: people don't like it when companies tear up their street (or yard or driveway) to install utilities, and complain to their local elected officials when it happens. Democracy really works at this scale. For every person that is willing to have their street torn up for a second (or 10th) time to install a competing ISP's infrastructure, there are 49 people that don't understand and will keep paying comcast, and 50 people that don't care at all, but they sure as hell care about backhoes and construction workers in their yard.

Honestly, the local officials preventing competing ISPs from operating aren't even wrong, even if you view it as short sighted to represent their constituencies accurately. It's monstrously inefficient and stupid to lay competing physical networks.

It's not going to compare well. The assembly is a pretty simple port of a basic C implementation of radix 2 Cooley-Tukey FFTs. The author gets ~2x speedup from writing this in assembly.

In contrast, highly optimized FFT implementations that use higher radix transforms and other FFT algorithm optimizations, along with SIMD and good C (not necessarily assembly) will get speedups of 10-40x over basic C implementations with a radix 2 implementation: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.153... (Implementing FFTs in practice by Frigo and Johnson, of FFTW)

FWIW, it is possible to significantly beat FFTW, if you highly optimize for particular cases: https://github.com/halide/Halide/pull/977. I don't have data handy, but I've seen similar results comparing against IPP.

Uhh, no, I don't remember that. I remember this:

> The war itself played a major role in the growth of health insurance, according to a March 2002 history of health insurance benefits from the Employee Benefit Research Institute, a think tank that examines employment-based benefits. Wages had been frozen by the National War Labor Board amid a shortage of workers as many potential employees went off to fight in the war. Employers sought to get around the wage controls in order to attract scarce workers. Providing them health insurance was one way to do that, EBRI said.

> "During this time, employers petitioned, and were granted to have benefits, particularly health benefits, not be considered part of wages," Stephen Blakely, the director of communications and managing editor at EBRI, told us. "Congress agreed to exempt health insurance benefits from taxation."

http://www.politifact.com/virginia/statements/2014/oct/30/ja...

Maybe, just maybe, if there is literally nowhere on the earth that you can go to find your ideal society, that your ideal society can't exist? There are hundreds of countries and 7 billion people. If none of them have the system you want, that might suggest it isn't workable in practice.

> Perhaps Google could, by offering a no-ads subscription across their adwords networ. They are unlikely to ever offer this though, because it would cannabalise their existing, highly profitable ad business. Hardly anyone else would have the necessary traction to make it work.

https://www.google.com/contributor/welcome/

> I disagree with you about the DNC emails, they go above and beyond preference into outright campaign strategy coordination and undermine the whole idea of the primary process.

What exactly do you think the primaries are? They are not official elections. The democratic party (and the republican party) exist to help the candidates they want to support get elected. Note the mere existence of superdelegates. Primaries are designed to allow people to have some influence on the party's choice (and to get the electorate feeling invested in the party's choice of candidate... obviously the DNC messed that up), but really, the primaries are designed to maximize exposure for the candidates to help them in the general election. Trump is the perfect example of why the parties want some control over who they lend their name to in the general election. I'm sure that the emails flying around the RNC while they had a chance of influencing Trump's election put the DNC to absolute shame.

Bernie was not even a democrat at all prior to this election. Bernie has been an independent for his whole political career. He joined the democratic party just for this election, and he's now an independent again. It seems absurd to expect party support with that context. I think Bernie made a deliberate choice to run in the Democratic primary (despite not being a Democrat) to avoid the spoiler effect in the general election, knowing that it was unlikely he would get DNC support. I respect him a lot for that decision.

Don't get me wrong, I would have been very happy if Bernie won (especially over Trump), but this idea that the DNC is a rotten institution that deserves an organizational death penalty for working with Clinton over Bernie is ridiculous.

Don't miss the blog post he linked either, it's also a great read: http://zxq9.com/archives/911

This is where the "100x myth" comes from. If you write a naive scalar algorithm in C, and then you write (almost) the same thing in CUDA, you can get a ~100x speedup. The problem is that most of that speedup is coming from the lack of parallelism and vectorization in the C implementation, which CUDA gives you "for free".

The thing is, this is still a real win for CUDA, because most people aren't going to write highly optimized C. It's just not a hardware win, it's a compiler win. The thing that most people get out of CUDA is not an advantage from running on a GPU, but an advantage of a smart compiler/language design. Most of the same people would get all but 1-5x of the benefits from using something like ISPC[1], which gives a similar programming model as CUDA implemented on the CPU.

In my experience (from the Tesla/GTX 200 era, though I doubt things have changed that much now), such a small boost in performance is not worth the hassle of transferring data to/from the GPU, the (lack of) virtualization, and driver shenanigans/support issues (at one point, I had to suggest someone buy a fake monitor dongle to plug into his GPU to be able to use my code...).

Anyways, it's good to see this thread, people these days think you are crazy for not using the GPU to do big compute workloads.

1. https://ispc.github.io/

Would you still say this:

> I'm not sure Wheeler's turncoat, pro-consumer behaviour was predictable (or, for that matter, predicted) by anyone, you or otherwise.

After reading this[1], written by Wheeler?

> Back in the original analog-to-digital days I can remember AT&T’s representatives warning of catastrophic job losses and damage to the national security if innovative competitors were allowed into their business. The same echoes surround the proposed NASA changes. Many of the Old Guard, launch system contractors, and their congressional supporters are bemoaning the thought of a competitive manned launch environment. We’ve heard all that before at the time of another “analog-to-digital” conversion. The earlier warnings not only failed to materialize, but just the opposite occurred as new, innovative and less expensive services came forward and economic growth and a new generation of jobs exploded.

Is it really that surprising that someone who wrote that turned out to be a regulator with reasonable positions on things like net neutrality?

Maybe "obvious" was a bit strong of a word (Wheeler's writing is a bit heavy on long-winded metaphors...) but I also definitely do not think that the vitriol directed at him starting the day of his nomination was warranted, and I do not think that Wheeler's behavior was so unpredictable.

1. http://web.archive.org/web/20110206170617/http://www.mobilem...

I read a bunch of the posts that are still accessible, I still think it makes the point: it's not surprising that Wheeler is reasonably pro-consumer. It's just that I can't find the particular post that has stuck in my head all this time.

Amusingly, it seems that he allowed his blog's domain/hosting to lapse. You can find some of it on archive.org: http://web.archive.org/web/20110206171653/http://www.mobilem... I've only been able to access a subset of the posts, and I can't find the particular post that really convinced me that he wasn't the example of lobbyist/regulator revolving door that everyone was saying he was.

I wasn't on HN at the time. However, I did post links to his blog on reddit in the relevant threads, my posts linking to his blog just didn't get much traction. They got mildly upvoted and started a few small conversations, but absolutely nothing compared to the witty one-liners and ranting about lobbyists, shills, and Obama being a sellout.

When Wheeler was nominated, he had a blog on the topic of telecom public policy, going back several years, with a lot of very deep and nuanced commentary on issues like the natural monopoly that arises in telecom infrastructure. I read a lot of it. It was obvious that he was not a last minute "turncoat". Wheeler's behavior was very predictable, if people only read primary sources instead of letting themselves get led around by the nose by clickbait/outrage farms.

Without digging into it too much, I would bet that most of the improvement from ICC comes from their awesome implementations of transcendental functions used in this program (exp, sqrt, etc.).

Still, the general point is good. last time I tried ICC on a heavily vectorized (with intrinsics) program, ICC was a 30% boost over clang or gcc.

You must be joking. The only place I've ever gotten adequate audio quality from an FM transmitter is in a tunnel more than 1/4 mile long. They're crappy even driving through the middle of nowhere where 9/10 radio stations are noise.

It's bullshit on another axis as well: it assumes that the labels have a claim to a share of all music revenue on YouTube. The slideshare deck they linked for that claim shows the large majority of YouTube songs are "off label content" (slide 46), and it also says that more than half of the revenue is from "user generated content" (slide 43).

There's a huge difference in illiquid assets (most of that 11.2% is real estate, stock, etc.) and currency that is at least presumably intended to be liquid.

I called it that because that's where I'm from...

And I don't really consider it derogatory. It's really an accurate description. I was referring to the literal middle of nowhere, where no one lives, not even small towns. The places you drive to get to the small towns.

edit: I think maybe the confusion here is that I was not intending to make an association between the people of Oregon and the AM radio callers I was referring to. (The shows you get are nationally syndicated, aren't they?) I only said that because that's the only time I listen to talk radio, and unless you drive through places where it's hard to find any radio station, you probably won't have listened to it before. It's very different from e.g. Fox News, where you hear professional pundits talking about these issues. You'll hear real people stating opinions that others in this thread are assuming can only be the result of blackmail. That was the whole point of the post.

I wouldn't assume that the general public is strongly against this kind of legislation. A subset (our subset) is, but the HN demographic is small compared to the rest of the country.

The basic concept of strong encryption is not something that most people understand. People are projecting their ideas of physical locks and warrants onto digital systems because that's what they understand. It's flawed, but it's hardly the first or even biggest/most important issue where poor understanding by the public leads to bad legislation.

Every now and then I have to drive through bumfuck Oregon, and the only radio stations are AM talk radio. It would be "educational" to spend some time listening to that every now and then. There are real people on the other sides of these issues that seem blindingly one sided to us. I think a lot of Congress probably just thinks that passing this kind of legislation (on encryption) will make most of their voters happy. Just like shutting down the EPA, or ...

The compiler can use the known bounds of a type to perform correctness checks or optimizations. It might not be testing every possible value of an int explicitly, but the compiler certainly uses the range of an int to make lots of decisions and produce warnings/errors.

> C++ is a baroque language with way too many conflicting features, conflicting cultural standards, and general all-around cruft. Isn't it?

This is a really common sentiment, but when pressed to name a specific thing that should be removed (ignoring backwards compatibility), people really struggle.

The standard library has a lot of cruft and poor legacy decisions, but it's several decades old. Any code that's been around for several decades is going to have cruft. Any language that is as successful as C++ will have the same problem when it reaches the age of C++. It's inevitable. Anything looks elegant and free of cruft when its only a few years old and has low enough usage that the maintainers can make breaking changes without breaking billions of lines of code.

We absolutely do know that. Even without any aggressive tax avoidance schemes, if you have a significant investment income, you probably pay an effective ~15% tax rate [1] [2]. I can pretty much guarantee you that most of the people named are in this bucket.

If you earn a salary that puts you in the literal 1%, you probably pay an effective tax rate of 30-50%. I don't know how low your income needs to get to reach a 15% tax rate on income, but it's really low.

1. http://www.nytimes.com/2011/08/15/opinion/stop-coddling-the-...

2. http://www.csmonitor.com/USA/Elections/President/2012/0119/U...

Electrical engineering is a pretty wide field that doesn't necessarily involve working with actual electronics. Personally, I've built a circuit simulator software package, but I can't actually design or build circuits worth a damn. I had to copy all but the most basic test circuits for my project off of existing designs, and I'd probably fail to build a physical realization of most of them.

There's no way that those other sports are more dangerous than boxing. That's ludicrous. About 5-10 people die per year from boxing [1], while about 1 in 1.06 million people die per day of skiing [2]. That same source estimates that 1 in 1.2 million boxers die per minute in the ring.

1. http://ejmas.com/jcs/jcsart_svinth_a_0700.htm

2. https://www.nsaa.org/media/68045/NSAA-Facts-About-Skiing-Sno...

If it just gets encoded at the same semantic level as text or an AST, then what do you really gain by using modules? That doesn't seem like as big of a win as people asking for modules seem to think it is.

I think this is vastly over-estimating Tay's capabilities. Nearly all of the examples I saw of offensive behaviors were just generic responses to leading questions.

Everything that didn't fall in that bucket was clearly just regurgitated quotes from things it had been sent, which clearly needed moderation. But, that's no different than moderating a message board or comment section.