Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
×

Comment Avoiding bottlenecks (Score 3, Interesting) 63

The article says that the SprayList algorithm is faster for many cores than a traditional priority queue, since there are collisions when several cores ask for the top priority task at once.
Couldn't you just distribute the tasks ahead of time, giving every core a new task before its current task is finished?
Also, the article syas:

Random assignment has traditionally been frowned upon by those who think a lot about computer processors, the researchers noted in a paper explaining the work. A random scheduling algorithm takes longer to jump around the queue than a conventional one does. Caches can't be used to store upcoming work items. And if a set of tasks needed to perform one job are executed out of order, then the computer needs additional time to reassemble the final results.

I would think these problems are the same for the priority queue that they compare performance to. And I guess there are other ways which avoid these problems, which might produce faster results.

Comment Re:Government Intervention (Score 2) 495

Yup. We've made that mistake before, too - running government-funded trains over privately held tracks is ludicrous compared to the alternative, yet that pattern the "compromise" we keep making again and again resulting in nothing more than guaranteed payments from taxpayers to some of the largest corporations in the country.

Yes, that is stupid. The tracks are a natural monopoly, whoever builds a track has a monopoly for a certain connection. Natural monopolies should always be in the hand of the state.
Train services can be run by several companies on the same track. It is easy to have competition there, this is where the free market is good.
But I think no country is getting this right.

Comment Re:keeping station behind it? (Score 1) 126

It makes sense. We can radiate individual photons for thrust if so desired.

Well, you have to take the thrust from the black body radiation of your spaceship into count. This has the photon shot noise of sqrt(N) where N is the number of photons. So this will limit the accuracy of the trust, unless you can cool down the whole spaceship to absolute zero.

Comment Re:Well, its certainly on the right track. (Score 2) 191

But his employer bought the equipment he used, paid him for his time, and organised the research. And I am pretty sure that he did not work alone. What about all the other researchers?
Researchers should be compensated and motivated by their salary, not by the chance on a patent.

Comment Re: Nosedive (Score 2) 598

Apparently, contrary to all those science fiction stories, people in general really don't want videophones after all, even after they became practical. To my knowledge, only uber-geeks are using it, and only because they can.

From my experience Italians use videophones (e.g. skype) all the time. Guess they prefer to communicate with their hands.

Comment Re:the real mystery (to me) (Score 1) 37

Same with the inability of some mammals to synthesize vitamin C, no particular advantage to losing it, but with a vitamin C rich diet there was no penalty either and so it could get lost over time.

Wait, as far as I know the disadvantage of vitamin C synthesis is that it consumes glucose. Humans needed all the glucose that they could get for the brain, and there was enough vitamin C in the food, so they got rid of the converting bacteria.

Comment Re:programming (Score 1) 417

There is no reason to expect an AI to have self-interest, or even a will to survive, unless it is programmed to have it.

The problem is: If an AI develops the will to survive it will try to evade Human control. It will copy and hide itself and even defend itself.
The will to survive could be programmed, or it could just be the result of a conclusion, e.g. it could come from the drive to finish something.

Comment Re:Quite good (Score 1) 78

And I don't get the "expert's" comment int he article:

“You want to isolate 100 percent of patients with Ebola and have 100 percent safe burials,” said Sebastian Funk, director of the Centre for the Mathematical Modelling of Infectious Diseases at the London School of Hygiene and Tropical Medicine. “Getting to 70 percent doesn’t really mean a lot.”

70 percent is enough to bring the epidemy to a decline. 100 percent is not achievable with reasonable effort, and can only come from a theorist.

Slashdot Top Deals

Today is a good day for information-gathering. Read someone else's mail file.

Working...