mmf

Also known for woodgears.ca Matthias is an international treasure. Any googlers around please help him.

If you get a chance check out his recounting of RIM rise and fall.

Edit: it seems he got it back :)

I think the main reason they did not expose lower level semantics is that the wanted a drop in replacement for hdds. The second is liability: unfettered access ti arbitrary location erases (and writes) can let you kill (wear out) a flash device in a really short time.

“Eutectic alloys are also notable because they often have the lowest melting point”. Isn’t the very definition of eutectic the mix proportion that has _the lowest_ melting point?

Optical computation will never become relevant at scale. There are fundamental reasons for this: first, particle size. A photon at usable wavelengths is extremely large, much larger of any modern electron based _devices_ This makes it imossible to scale to usable density. Second, optic-optic (as opposed to electro-optic) non linear effects are based on interaction with electrons, in particular with electron decay from an energy state to another which is tipically extremely slow.

If you feel that something is trivial and potentially wasting other people's time by means of wasting pixels, I think it's fine to speak up. (Explaining why something is trivial may not be that effective, in general).

I was clearly wrong in that many people disagreed and found this to be an interesting and nontrivial subject.

Ok, I guess I could have provided more info. Do this if you own a bicycle: hold it by the saddle perfectly straight, front of the bike forward. Make it lean on the left (no forward movement). Look at the handle, it steers left. When you're moving this means that when you lean left there will be a torque, due to your circular path you'll be following, that will try to make the bike lean right.

Bend left -> torque to the right, bend right -> torque to the left. Here you have the basic ingredient for stability: the one torqueless angle is the bike going straight perfectly vertical wrt the ground, while any perturbation leads to a torque in the opposite direction.

I felt it was basic stuff maybe it's not?

How complex can it be? If you have a bare minimum background in basic physics and control theory, and a bicycle in your hands, the qualitative reasons for balance and stability will jump at you in under 10 minutes...

I absolutely agree, all those negative comments sound like rants from grumpy old guys. I have worked in HPC distributed storage and the problems the CoreOS people are trying to solve is a valid and present one: deploying flexible distributed storage today is an absolute pain.

Plus all the arguments about local storage failure being hard are moot because this is a distributed solution and the hardship of local storage reliability can be abstracted away almost completely.

That's great news but: 900MBps != 5GB/3sec. Plus last time I checked (2 years ago) write power for flash was about 10W/GBps and growing over time. So I expect these speeds not to be sustainable for extended periods of time.

This effectively makes forgetting one's password a crime with a lifetime jail sentence... Not bad, not bad

And will they share the data they collect so that everybody has a shot at making discoveries?

Little known implies that if you ask 100 randomly selected people if they know if it exists, what it is and what it does, a small number, say less than 10 or so (little) will know.

Did you take your educated guess at that fraction?

I do not think he was implying that he did not fit because he's old. He implied that he did not fit because he's mature...

At this stage of things, I think it's more forward looking to open source trained models. Not only they are beginning to be the real core of future building blocks (see, e.g., trained word2vec vectors) but also the contain the real complexity in a NN, i.e., the are the "real function" you would want in a library.

I don't know why you're being down voted. The fact that the human won, makes it now so much more interesting! It surely is the best outcome possible both for human go players and Google as well (having won the first 3 already).

On the spot. I love wood and I'm a woodworker but I have yet to see a convincing way of treating wood that is exposed to weather.

Finally a post on Concept Insights :)

I work on this project, feel free to ask anything.

No they cannot, and you cannot either :)

I think you missed the point. He said that the starting idea was not important, and made examples of actually important starting ideas (page rank). His point is that the execution mattered way more, given that the idea was not really mind blowing.

IMO once all barriers (psychological and not) to quick and easy small online payments are gone, ads will not be useful anymore to content providers. A click through rate of 1/1000 will be easily compensated by a rate of donation of 1/10000.

Awesome: "fed to cattle or schoolchildren". :-)

I read it as meaning that it should be a superconductor. Why they do not mention it, I don't understand...

It's W/GBps, or if you prefer simplifying J/GB. I like the former better, power and bandwidth are handier in real life computer science.

A few corrections:

-refresh period is more like 32-64ms

-refreshing is way (way) cheaper than streaming data out since you refresh an entire word line, and you do not pay io cost, which is a significant fraction of the cost of accessing dram.

Here:

BS: "going into production this year"

Possibly not BS: "going into production this year at 2x nam technology node with a y Gb part"

Many nvm memory technologies, including mram and PCM have been "in production" for quite many years, with << 1Gb parts on old technology nodes, that is.

There are many ways of suggesting that a technology is ready when it's not. I have not seen anything published with 2x nm and a cell size that is anywhere near 4F2 to 8F2 (on a decently large array)...

Re the latency caveat I was suggesting BS spotting rules. You're right that this article did not claim the memory was faster than dram.

RE power consumption, it's true that it you don't use it an NVM will not consume power, yet that is not what people refer to when talking about power consumption. Can you quote a technology paper that shows lower than dram, or lower than flash for that matters, active power consumption? (The proper unit is W/GBps here)

As I did not have the numbers fresh in memory (apparently I needed a refresh ;) ) I googled for dram data sheet and got the first ddr3 PDF (Hynix). Quick, and very dirty math suggests that

-refresh power is on the order of milliwats per Gigabyte

-memory access is on the order of 0.1W per GB per second

For refresh power, let's face it: it's puny. Regarding the active power, I have yet to see anything (off chip and this size) that is better than that.

P.S. I did say active in my original post :)

Not shits and giggles, it's access to the array: subarray and word line decoding plus sense amp time. Find me a technology that does not need all these and I'll bow to you (and invest in your startup :D).

I did research in the field from 2008 to 2013. The claims are always the same: faster than flash, denser than dram, lower power than both.

Here's how to spot BS: first, the density claims are immaterial until they prove yield at a technology node same as dram's today. There are multiple billions investment between 180nm and high yield (and low mask count) 2xnm, no matter how cool is the new memory technology.

Second, speed claims must be explicitly about latency: flash bandwidth is as large as you want it, latency is ~100us (read). Even so, the moment anybody claims that latency is faster than Dram, you know they're feeding the hype and lying to you: dram latency does not depend on the memory technology, rather it depends on the array size. So an 8Gb chip of any memory technology that is fast enough, is likely to be just as fast as dram.

Third: power consumption. Dram's active power is as low as it gets, the memory cell in particular stores information with really little energy. The array interconnect and circuitry consume most energy and a different memory technology won't change that.

It's so easy to forget that your salary is not everybody's salary. In Italy (a first world country afaik) a post doc makes a grand total of about 30 euros per day. An assistant professor probably around 45, which is similar to the entry salary of a software engineer. If you subtract food, housing and transportation, that $100 book starts looking very (very) expensive.

For a large fraction of the world's population time is not money...