Even if Moore's Law is "running out," there's still plenty of room at the bottom

1 Like

But how many of us here personally invested? If people are already so deep into a game, newcomers might be more likely to succeed by playing long odds, or even a completely different game. [quote=“nixiebunny, post:20, topic:80829”]
By the way, I found a symposium proceedings from 1965 at the local science library that discussed ways to make electronics move forward. Amidst all the tinker-toy packaging ideas, one prescient paper actually said that CMOS in silicon was the future, because of the low cost and small size of the process. It was right then and it’s still right.
[/quote]

I can concede that it may have been accurate for 1965 (ECL FTW!), but not for present day. Carbon is much smaller and cheaper than silicon. And one’s choice of transistor and logic types has a lot to do with how many can fit on a die. It’s probably good enough for the industry, but the industry is behind the times. This is why we need an R&D based economy based upon knowing how things work instead of trying to influence people by selling stuff.

Anyway, my point was that sticking with silicon is a choice, and not a technical barrier of Moore’s Law as many say.

1 Like

Parallelization is good, but not a silver bullet. Or even really copper jacketed.

Amdahl’s law shows that increasing parallelization gives you rapidly diminishing returns, and even stuff that’s highly parallelizable isn’t going to see major speed increases unless the number of processors you throw at the problem starts increasing geometrically.

The TL;DW for that video is basically: Adding more cores isn’t a good solution. You need more cores that are specialized, like an FPGA. Or even big.LITTLE type stuff. Uniform masses of identical cores aren’t going to save you.

5 Likes

What about quantum computing? My limited understanding is that it uses quantum entanglement to remove the limitation of the speed of light in connecting the results of parallel processes. For practical use, it is still pretty far away, but my understanding is that the NSA is already thinking about how it might make it much easier to decode current communications in the future.

This would constitute information travelling faster than the speed of light, which violates all current physics theories. Entanglement results in changes being propagated faster than the speed of light, but currently there’s no evidence to suggest that those changes can be used to transmit information.

The big advance of quantum computers is that their behaviour allows them to implement algorithms that just don’t work on classical computers - such as Shor’s algorithm, which scales up in time required much less than equivalent algorithms for classical computers, allowing prime factorisation much faster than in classical computers.

5 Likes

Quantum computing is not going to replace classical computers. It is basically like magic, but only for certain well-defined classes of optimization problems. It won’t help us directly with most of the things we use computers for, day-to-day. If we had enough entangled qbits, it could help with a lot of the things we use supercomputing clusters for.

7 Likes
3 Likes

What do you mean specifically by first class?

ARM chips run much of the mobile linux world, not to mention the iOS and Win mobile devices. [quote=“LDoBe, post:23, topic:80829”]
Amdahl’s law show
[/quote]
That is the weirdest episode of Law and Order ever…

4 Likes

There is quite a bit of software out there that I’ve seen that won’t compile and run on ARM and which gets no prepackaged .deb or other distribution files for it.

As to the mobile Linux world, is there really much of such a thing? Is it even 1% of 1% of the market?

i’ve seen that. i think as arm adoption increases on desktop os style devices we will see that shift.

Android uses the Linux kernel under the hood and Android accounts for 80% of the current mobile market.

3 Likes

Yeah, but I wasn’t talking about Android. I was talking about Linux. :slight_smile:

Let me know how you check into a git repro on Android.

Answer: You have to run a virtual machine on it running Linux because you can’t do it from Android itself.

I’ve used a shell on android. It runs a lot of stuff, including busybox. If you can do git with busybox you can use git with Android.

(you may need to root your device)

2 Likes

but Android doesn’t let you run things like git natively, hence my having a conversation about Linux on ARM and not doing a handwavy thing of “well, actually, Android is Linux…” It isn’t and, outside of my phone, I prefer real operating systems. It is the same reason I don’t run Chrome OS on a laptop-like.

I do a lot of stuff with raspbery pi’s and similar devices that are ARM-based. I run Debian systems on these but, as I said, a large subset of software for Linux won’t even compile for them, hence my original remark about wanting ARM to be a first class citizen on Linux.

Hell, you can’t even run a current version of MAME (WTF?) on ARM systems. It makes me cry.

2 Likes

Rooting your Android device (especially if it is your phone) is step one in the rampant malware infection process which is Android as well. I keep my phones unrooted. :slight_smile:

2 Likes

I root mine, but I don’t use my phones for games. I also use a hosts file and ublock, so I avoid a lot of bad stuff.

1 Like

The thing about Android is that so much of the software for it is pirated that it is unreal. A large percentage of the “pirated” software people download is riddled with malware. Rooting the phone in combination with this being the lay of the land for Android is a security disaster.

As bad as iOS is for being locked down, it winds up being effectively far more secure for this reason. Most of the security folks I know and work with tell people “Just get an iPhone” for this reason, even if they often own an Android phone or two.

4 Likes

So you’re saying “stupid users”?

I mean, working the helldesk, I sorely want to agree. But I believe that the vast majority of users can be trained in good computer hygiene, and that a lot of “good software” incentivizes really fucking bad habits for people who don’t know any better.

1 Like

Not stupid, just cheap and willing to pirate software if they can figure out a way to do it without paying for it. “Oh, this blog post says to root my phone and install this. Ok! Now I don’t have to pay $2 for this app!”

There is a reason why it is hard for developers to make money on Android despite its market dominance. It is because everyone on Android seems to pirate their apps.

1 Like

That sounds an awful lot like you’re saying that just because people may be able to pirate software on Android, and just because some of what they may be pirating may be malware, Android itself is insecure.

Do you think Windows is a security disaster because people may download pirated software that might have malware in it also? (Windows is in some cases a security disaster, but that’d be a novel reason to call it one)

Yes and Yes.

Android is pretty bad. I don’t care how secure someone’s hermetically sealed archetypal install is if the vast majority of users go install stupid shit and get pwned. Yes, in a perfect world, it isn’t horrible but we don’t live in the best of all possible worlds.

You realize, historically, most botnets are composed of Windows computers where people either installed stupid malware or just never bothered to run regular security updates and got owned by some security hole? Your cousin’s ancient Windows XP on their “perfectly good” 2003 computer that’s never installed a service pack but which they use to cruise porn sites…

I mean, I do security for one of the top name brand browsers and I handle incoming security bugs. The shit I see…

3 Likes