static_noise

It's just another corporation doing its thing to stay in business.

They ran into a dead end and changed the strategy. They can change it again. Utimately, money decides.

They get an average parallelism of 4-10 with some standard algorithms assuming unlimited resources. Where are we now with modern intel processors on those algorithms? 2-3?

I see you are very knowledgable and an expert in the field. Could you please give us a short introduction into how spintronic transistors work?

I'll have to repeat my question in other words. How many MIPS do you think, remain, if every third instruction is a mispredicted branch and every second memory access is a cache miss?

Modern CPUs use pipelining which executes many instructions parallel. This only works well if everything goes as predicted. If you have an algorithm which works contrary to what the branch prediction thinks and a cache which does not hold the data you need, your performance goes down the drain. Those MIPS mean nothing if not put into the right context.

That linked thread on Anandtech digs in quite a bit. The real physics involved is, of course, more complex than my simple model. Instead of a square dependence, they find the 5th power of the supply voltage. I believe this has to do with how the transistors are designed which creates a lot of wasted current at higher voltages. The square dependence would show with perfect transistors.

I was wondering why the graphs stopped like a decade ago with a Pentium 4. Then I saw the date of the article which is 2008 a.D.

Does this hold true for algorithms which are hard to parallelize because everything depends on intermediate results which are hard to predict?

> The only problem is that the anmount of electrons you have to push through the transistors has a quadratic dependence on frequency.

I don't see how this is ultimately true.

The amount of electrons needed to flip a capacitor is constant given that the voltage is constant. The more often you flip, the more current you need per time.

The amount of electrons needed to flip a capacitor is proportional to the voltage. So, lower voltage means a) less power per electron and b) less electrons. Therefore the power is quadratic with the operating voltage.

Third, if you want to increase the frequency, at some point, you have to increase the voltage or you get errors. Then you increase frequency and voltage at the same time and your power consumption shoots up.

While I can't speak from first hand experience, I have an open ear sometimes.

There are these developers who just made an awesome webapp or something. It's really great but there is a problem. Noone knows about it. But they have a solution for that problem. They buy some "promotion". It's articles in newsletters or outright spam directed at their target audience. I cannot tell you exactly who their target audience is but I can tell you who it is not. It's not you, me or most readers on hackernews. It's less tech savyy people with more decision power. Old people. Their site is made for old people. And they don't care about you and me or any other passer-by who won't pay a dime. They care about conversion rates and that is what they monitor all day long.

It's a bit like in those scam-emails. They are so obvious that anybody with half a brain will click delete. But those who do reply, oh boy, they are so ready to be milked off by that nigerian prince.

If you start running EC2 clusters, you're more in the "cloud storage" dominion and not really in the "self hosting" area anymore.

Not everybody needs that and not everybody who does need it also realizes it before a major failure strikes out one of their machines and there is no working failover and their one admin is not reachable and when they finally reach him he has to be flown in, then they need to wait for replacement parts, which all in all delays recovery by several days.

Do you need to store data?

If yes, how is it replicated over multiple datacenters?

If yes, does failover work flawlessly?

So, you have redundancy and failover at the application level. You don't need reliable servers then.

What if your hardware fails?

What if you need to scale up (or scale down again)?

Making servers run as reliable as in cloud datacenters is really hard work and imposes additional cost. I.e. you need not one, but two or three datacenters with 2-3 times the amount of servers you actually use. Then you need admin staff who knows their work and keeps things running smoothly and reliably. You need not one but at least 3 of that staff because they might get sick or go on vacation.

Your cost savings probably comes from cutting corners. Depending on the structure of your applications, downtime requirements and recovery plans, that may even be OK. You don't always need a fleet of tanks to deliver a box of milk bottles.

What most people fail to realize is that we are pieces in the game of life. Our emotions have been shaped by eons of evolution. Emotions are part of the game, leading the mind into specific directions governed by genetically programmed and environmentally activated reflexes. At times they overrule the rational mind and result in behavior that is "not my real me" that the subject identifies with.

We should at least try understand the basic rules that shape our lives and not give in to simplified (and often misleading) models presented by ideology.

Society and its picture of the human person will (and must) change fundamentally upon the results we concieve by proper scientific study. This will not go down easy.

The game has changed in major ways and contraception is one key part. Sex is detached from reproduction.

Absence of evidence is not evidence of absence, correct. Some theory that has not been tested, however, is equally pointless. Unfortunately we're opertating in a space here where it's more an economy of attention than hard scientific discourse which is very hard to attain in that field. You seem on the right track, though.

Then should we consider Windows or other commerical software for anything designed to outlast a 5 year period?

Does this mean

a) older Windows versions will not run at all on those chips?

b) older Windows versions will still run but cause some problems?

c) older Windows versions will run without problems or performance impediment compared to old chips but some new features are not supported?

d) everything will work just as well but when problems occur Microsoft won't provide official support?

Replace the average (mean) with the median and you will have better results in almost all cases.

When the data is good the result ist virtually identical. When there are outliers, the median just ignores them unless there lots of them. When the data is that bad, that the median doesn't work anymore you need to up your statistics quite a few notches.

So the solution the author proposes is both absolutely correct and absolutely useless in practice?

You are arguing from a high horse because that's how the human mind came to be and look where it's gotten us.

It's not an issue either, it's a business opportunity.

It's about managing your resources. You can simply buy the fastest processor for cheap money. You can buy systems with many cores or even build a cluster for considerable money.

But you can not buy time.

Developing python applications in many cases is much quicker than developing comparable applications in C++, C or Assembler. Therefore it makes a lot of sense for many companies to choose tools with quick development time to get the product running and care for performance later.

Since there may be so much code be written in python, it cannot simply be converted into a faster language since most critical parts are already accelerated using the proper libraries provided by the language ecosystem. Porting a complex application to another language in many cases is a really hard problem and might even not be possible to be done in smaller steps.

Therefore we need every bit of performance we can squeeze out of python. The whole world will gain from that. If we manage great strides like JavaScript it would be glorious!

Very short half-life-times in the minute regime and below are actually pretty great because you can just keep that stuff in a container for some time and it decays away exponentially.

> Can't imagine that AWS bill...

That's not how things work over here.

We can and do imagine. We even calculate estimate values so we don't have to stay in the dark and reason without a solid basis.

I'd encourage you to find some traffic numbers provided by imgur and some S3 pricing to do a rough estimate. Maybe someone more knowledgable will pop in and improve on that. When we argue everyone involved shall be wiser after. That's how it works around here.

Dear lamb, we had a great time but now I have to eat.

They started as an image host for those freeloaders. That made them big. Now they stopped providing that service.

Yes, imgur can do what it wants. But so can we. Also we may complain as much as we want for being deceived for so long.

Thanks for the free bandwidth and farewell imgur.

In some fields the winner takes it all. Either you win or you die.

Also money is like a drug. What's wrong with taking a little cocaine in the morning and a little heroin in the evening? Nothing, except there are very few people who can sustain that over an extended period of time. You also wouldn't know about them because they neither become very rich nor very famous. ("Hey look there is the most average guy in the whole country who has no exceptional qualities.") The world belongs to the candles who burn from both ends.

Reddit needs to grow its revenue to be sustainable.

It looks to me as if all recent changes work towards the goal of making reddit a mainstream site platable to the average customer who does not know or care about adblock and innovative marketing techniques.

RES is for the power user who does not generate any revenue.

It won't be very long before we need a new reddit for power users where we can move on to with our guns, cocaine and hookers.

Bad news for the current shareholders of imgur.

All we want is an easy and reliable image hoster. You upload an image, they host it, you can link it, fin.

Perfect plan!

Except for the plot hole of not generating any money this way. That's where the suits creep in. First they finance you to grow and grow quickly you must. Imgur did it. Then you experiment with some methods of generating ad views or other revenue. People realize they hit on a gold mine and business starts to grow like cancer. Then there are some minor changes to optimize revenue. Then some founders start to cash out and jump the ship. Then the suits, loudmouths and ginas slowly but surely take over the company and due process replaces the dictatorship of the personal union of owner and leader. There's guitars and barbecue as the corporate drones dance around the firepit of burning money as they hurry to please the gods of corporate capitalism. A spiral is forming like a siphon circling around itself pulling everything in like a black hole. And it's great.

If you listen carefully in the storm of excitement you can hear a faint voice who still remebers the core of the business: Provide the service and don't fuck with the customer.

Or as some politicians said: You can fuck over some of your customers all the time and you can fuck over all of your customers for some time. Both is a great business plan. But you cannot fuck over all of your customers all of the time. Especially not if you don't have a stranglehold on their necks like a dictator or monopoly has.

The imgur we have today deserves to die.

It may redeem itself by remembering its roots and not wanting to become more than it is. That would be great!

One thing that is more important over the whole debacle is to watch reddit closely and be ready to jump ship quickly as they stumble deeper into the corporate trap. Having a plan B "just in case" for everything important will cost you resources but it pays out for itself in the long run.