News for hackers. Stuff that matters.
Re:New two-stroke engine design could help keep internal com... by londons_explore on Tuesday May 26, 2020 @04:56AM (#23308791)
Investment into productionizing these kind of advances has pretty much stopped.

It's a bit like hard drives... As soon as the SSD arrived, everyone saw that hard drives would eventually die out, so manufacturers pretty much stopped new technology development. Sure, there have been a few small advances, but seek times and data read speeds haven't gone up for a decade, and capacities have only crept up.

IC engines are the same. It might take them 30 years to die out, but in those 30 years there won't be any significant tech advances, because every company that makes them is only optimizing for cost in a shrinking market.

Re:New two-stroke engine design could help keep internal com... by AmVess on Tuesday May 26, 2020 @07:49AM (#23309919)
That's not even remotely true. Increasing emissions standards and economy standards means that ICE development continues apace. Nissan, for instance, recently put into production a variable compression engine, and all manufacturers are pairing ICE advances with electric for even larger gains.

Honda has a nicely sized sedan that gets 55 MPG city out of a drivetrain unit that is amply powerful.

Most of them produce small 4-cylinder engines that are as powerful as the V-6's that proceeded them, all while producing less emissions and offering greater fuel economy. Even Ford, Chevy, and RAM are putting gas 4-cylinder engines in their mid-size trucks are just as capable at 8-cylinder engines used to be just not very long ago.

Emissions and fuel economy standards will continue to tighten, which means that ICE development will continue for a long time.

Re:New two-stroke engine design could help keep internal com... by pkulak on Tuesday May 26, 2020 @01:23PM (#23313874)
Anything recently put into production has been in development for years. The decision to begin development probably happened before people were even lining up to pre-order the Model 3.

This monster of mechanical complexity is brand new as of a couple days ago. What company is going to commit the next half decade to developing it while watching gasoline engines fade into the background? Especially after watching Nissan's new engine totally flop.

Re:New two-stroke engine design could help keep internal com... by londons_explore on Tuesday May 26, 2020 @09:23AM (#23310790)
Most of those improvements are not really new technology, but instead technology from European cars (where gas taxes are high) just imported into the USA (where previously people didn't care about MPG much because gas is much cheaper).
Re:New two-stroke engine design could help keep internal com... by AmVess on Tuesday May 26, 2020 @09:38AM (#23310934)
It is a global effort, and not one limited to the US. Besides, the biggest improvements in ICE fuel economy and power are from Japan, not Europe. Europe hitched its wagon to the diesel engine, and we all know how that turned out.

Sure, turbochargers and hybrid tech aren't new in an of themselves, but that doesn't matter because there's still a lot of room for development left in them.

Even just 5 years ago, the idea of getting a roomy, nicely configured sedan that gets 55MPG city was a silly notion. Now, Honda sells them all day for $24k USD.

Re:New two-stroke engine design could help keep internal com... by rsynnott on Tuesday May 26, 2020 @01:19PM (#23313818)
> Europe hitched its wagon to the diesel engine, and we all know how that turned out.

I think this is somewhat exaggerated; it never really went above half the new car market, and is obviously down a lot now.

Re:New two-stroke engine design could help keep internal com... by NullPrefix on Tuesday May 26, 2020 @12:56PM (#23313499)
>Even just 5 years ago, the idea of getting a roomy, nicely configured sedan that gets 55MPG city was a silly notion

VW diesels from the 90s were doing 5 or 6 l/100km

Re:New two-stroke engine design could help keep internal com... by AmVess on Tuesday May 26, 2020 @07:12PM (#23317537)
Those were neither roomy nor nicely configured.
Re:New two-stroke engine design could help keep internal com... by londons_explore on Tuesday May 26, 2020 @03:58PM (#23315742)
~= 50 MPG
Re:New two-stroke engine design could help keep internal com... by grecy on Tuesday May 26, 2020 @02:28PM (#23314690)
How many billions do you think will be poured into ICE R&D from 2020-2030?

My guess is virtually none.

Re:New two-stroke engine design could help keep internal com... by zaarn on Tuesday May 26, 2020 @05:21AM (#23308910)
I don't think that's entirely true; like ICEs, HDDs still have various niches were they are very valuable. Notably when storage density is the most important factory (data hoarding, backup solutions, etc.) and most of your access will be linear or random reads will not happen that often.

While not many new things in HDD space have made it to market, there is still quite a few big things in development, the new HAMR technology (and friends from other vendors) are projected to be able to double/triple capacity.

SSDs sadly still cost a lot if you need 50TB of them. Comparatively, my 50TB NAS cost me about 15$/TB. Most SSDs still hover around 90$/TB and go up to 120$/TB depending on use case.

Similary ICEs have two big advantages over non-combustion alternatives: energy density and price. A modern ICE is complicated but not terribly hard to construct, in part because processes exist but also because the things that happen in an ICE can largely be achieved by using common materials like steel, titanium, ceramics, oil and rubber. Gasoline and Diesel are very energy dense (and bad for the environment and climate) and can be lit/triggered by a simple electric sparkplug.

Comparatively running a hydrogen (non-combustion) engine is very hard to achieve as hydrogen is an element that really wants to set itself on fire and if it does so, the result is not very conducive to safe operation (hydrogen flames tend to be invisible and it loves to explode or atleast deflagrate).

Electric engines are much more efficient than any ICE but suffer from energy density problems (solvable by using a hydrogen fuelcell but that has above mentioned issues). If not for that, it would be the next thing and people are working on that. I have no doubt that despite that, ICEs will continue to be used where alternatives aren't possible or expensive.

Re:New two-stroke engine design could help keep internal com... by greggyb on Tuesday May 26, 2020 @08:58AM (#23310527)
> Notably when storage density is the most important factor

Flash storage is denser than HDD storage. For an easy example, think about stuffing. 3.5" enclosure with 1TB microSD cards.

As for SSDs, here's a four year old article on Samsung shipping 15TB SSDs in a 2.5" form factor.

HDDs are unlikely to reclaim the density crown from flash. That said, HDDs do still maintain quite a price advantage, which is typically more relevant than storage density for archival solutions.

The $/byte figures all remain in favor of HDDs.

And, as always, a relevant xkcd tells us that one of the highest bandwidth data transfer solutions remains overnighting a milk jug of microSD cards: https://what-if.xkcd.com/31/

Re:New two-stroke engine design could help keep internal com... by zaarn on Tuesday May 26, 2020 @09:22AM (#23310776)
I think in practicality, HDDs still beat out SSD in storage density in larger clusters simply because the SSD cluster would be obscenely expensive.
Re:New two-stroke engine design could help keep internal com... by greggyb on Tuesday May 26, 2020 @09:45AM (#23311018)
These are two different metrics. Density is simply bytes/volume. Price is $/bytes. The first has no factor for price. The second has no factor for volume.

If you are optimizing for volume, you will have all flash-based storage. If you are optimizing for price, you will have all HDD (I think. Not sure on tape pricing).

If we want to optimize for another measure, that's fine. But we can't redefine density just because there are price differences. It's like saying my Mazda is faster than a Ferrari, because the Ferrari is so expensive.

So, practicality is a different measure to optimize for, and a much more complex and situational one. We can observe that many bulk storage solutions use HDDs rather than flash storage. This provides evidence that HDDs are practical for bulk storage. It tells us very little about density.

Re:New two-stroke engine design could help keep internal com... by zaarn on Wednesday May 27, 2020 @01:35AM (#23319674)
Even then, flash storage that is practical to access (ie, not a milk jug full of microSD cards) isn't as dense as HDDs quite yet.
Re:New two-stroke engine design could help keep internal com... by greggyb on Wednesday May 27, 2020 @12:50PM (#23326088)
I think we're talking past each other. Both today, and in terms of potential, flash-based storage has a higher storage density, where storage density is the simple math of bytes/volume.

You keep coming back to "practicality", and this is harder to define. I welcome a proposed definition, so we can talk about this more clearly.

As a first pass, I might consider "practical to access" as being a combination of available at retail, and able to be readily attached to consumer hardware through a combination of onboard motherboard IO connectivity and expansion cards.

Here's a list of hard drives available at retail[0] that shows the largest retail HDD at 16TB and 3.5" form factor, vs SSD at 15.36TB and 2.5" form factor. These are available at retail today, and in standard form factors, with standard physical interfaces (SAS and SATA).

Without even getting into volume, but just looking at a single dimension, width, we can see that the HDD is 40% larger, physically and only 4% larger in storage capacity. The story only gets worse if we move to volume, because the 2.5" form factor is smaller in every dimension than 3.5". An enclosure to hold equivalent storage based on these two drives would be smaller for the SSDs than for the HDDs.

Perhaps practicality just means something that you could find without needing specialty hardware? Datacenter stuff is huge:

If we move to things you might find in a datacenter, we can find some articles (both from 2018). One talks about an existing 100TB 3.5" SSD - still in a standard form factor and with a standard interface (SATA/SAS).[1] In another article from the same year, Seagate was introducing the largest HDD, a 16TB model, and discussing the potential to hit 60TB by 2030.[2] Again, this is at the same time that a 100TB SSD became available.

Price is also something that has come up a couple times. I do not know if price is synonymous with practicality to you? I would appreciate clarification on the matter.

SSDs are, as I have mentioned previously, more expensive per unit of storage than HDDs. This is not up for debate, but I feel almost as if you think I am debating it.

So, for a specific use case, it is very easy for one to make a judgment call that HDDs are a better investment. But I do not see any argument to be made that HDDs are denser than flash-based storage. Again, density is storage capacity per unit volume.

A specific example, archival storage, would likely not benefit from an all-flash storage solution, but would be better served with HDDs. The tradeoffs for archival typically favor cost and often don't need the read and write performance of an SSD. As we agree, HDDs have better capacity per unit price.

If practicality and price are synonymous, it becomes a bit of a weird measure. We'll be taking a double ratio, of price/bytes/volume. But lets do it:

    HDD dimensions(mm):
    26.11*101.85*146.99
    390891.021465
    
    HDD density (TB/mm^3):
    16/390891.021465
    .00004093212461118814
    
    SSD dimensions (mm):
    15*70.1*100.45
    105623.175
    
    SSD Density (TB/mm^3):
    15.36/105623.175
    .00014542263097090198
    
    HDD $/density ($/TB/mm^3):
    499.99 / .00004093212461118814
    12215100.11389283589280755309

    SSD $/density ($/TB/mm^3):
    4295 / .00014542263097090198
    29534605.24902343799726872620
So, we've got an HDD at $12,215,100/TB/mm^3 and an SSD at $29,534,605/TB/mm^3. If this is the measure that practicality represents, then yes, HDDs have a lower number. If this is your definition, then I am sorry for misunderstanding you.

On the point of density, though, it seems clear to me, and I hope to anyone else looking at the data, that flash-based storage is denser (bytes/volume) than HDDs.

Dimensions are from Newegg[3][4].

[0] https://pcpartpicker.com/products/internal-hard-drive/#sort=...

[1] https://www.zdnet.com/article/worlds-largest-ssd-hits-100tb/

[2] https://www.alphr.com/storage/1010285/seagate-reveals-world-...

[3] https://www.newegg.com/seagate-nytro-3330-15-36tb/p/1Z4-002P...

[4] https://www.newegg.com/seagate-ironwolf-pro-st16000ne000-16t...

Re:New two-stroke engine design could help keep internal com... by wtallis on Tuesday May 26, 2020 @05:30AM (#23308961)
Hard drive development hasn't stopped. It just changed direction, because pursuing better seek times is pointless when it is impossible to increase RPMs enough to catch up to even low-end SSDs.

Perpendicular magnetic recording hit the market around the same time SSDs started to go mainstream. Since then, hard drives have advanced with the introduction of helium filled drives, multi-stage actuators and shingled magnetic recording. This has enabled an order of magnitude increase in drive capacity, and substantial increases in sequential transfer speed. Dual-actuator drives and HAMR/MAMR are in the process of moving from the lab to commercialization, and will respectively bring an increase in random and sequential IO throughput (but not latency) and more density increases (which always bring some sequential throughput improvement, too).

What's really been missing from the hard drive R&D scene is more Nobel-class discoveries like giant magnetoresistance to allow for big jumps forward.

Re:New two-stroke engine design could help keep internal com... by dogma1138 on Tuesday May 26, 2020 @05:24AM (#23308925)
There have been quite a few advancements in mechanical drives post SSD.

The “cloud” and all your data still primarily runs on mechanical drive and their density is now higher than ever.

And the development road map for mechanical drives is still expanding.

HAMR drives will be on the market in 2021-2022, then you have BPR, MAMR and even GMR drives on roadmaps and research papers.

The only reason you think development has stopped is because other than for home NAS mechanical drives have pretty much disappeared from the consumer market.

Re:New two-stroke engine design could help keep internal com... by Hamuko on Tuesday May 26, 2020 @05:22AM (#23308913)
SSDs have been mainstream products for at least a decade now. Ten years ago, the largest HDD we had was like 2 TB. We currently have 16 TB HDDs and people are expecting us to have 50 TB HDDs in 2026.

And not only that, 16 TB HDD is quite affordable. Just 500 dollars for a Seagate IronWolf. Meanwhile a 3.2 TB Seagate Nytro will set you back about $1800. SSD mass storage is an option for only those with deep pockets.

Hell, not even tape drives have gone away yet. Amazon Glacier is just one example.

Re:New two-stroke engine design could help keep internal com... by eropple on Tuesday May 26, 2020 @11:11AM (#23312137)
I was under the impression that Glacier is hard drive backed, not tape backed?
Re:New two-stroke engine design could help keep internal com... by jasonwatkinspdx on Tuesday May 26, 2020 @11:42AM (#23312527)
Yeah, it's spinning disk backed. What's different is the drives are aggregated up on hosts under the design assumption that most of the drives will be powered off most of the time, allowing a lower cost per byte.
Re:New two-stroke engine design could help keep internal com... by Hamuko on Tuesday May 26, 2020 @11:56AM (#23312763)
Some industry insider has claimed that it's tape-backed. There's also been speculation that it's actually based on optical media.
Re:New two-stroke engine design could help keep internal com... by londons_explore on Tuesday May 26, 2020 @12:27PM (#23313121)
It's totally backed on "we have loads of hard drives with spare storage capacity but no IO capacity, so sell it off cheap, but make sure it doesn't perform too well or it'll eat into our other products".

Nearly all cloud providers have more storage bytes than they have IO (ie. each individual hard drive receives so many read/write requests per second that if they filled the drive all the way to full some read/write requests couldn't be handled and would be unacceptably delayed)

Re:New two-stroke engine design could help keep internal com... by pg-gadfly on Tuesday May 26, 2020 @07:50AM (#23309930)
It's also to note that SSD capacity hasn't strictly improved, it's been gained through massive compromises that severely impact their lifetime and performance.

More than that, a 5 year old nas drive can outperform a modern cheap ssd in throughput easily.

Re:New two-stroke engine design could help keep internal com... by wtallis on Tuesday May 26, 2020 @02:04PM (#23314405)
Most, but not all of the capacity and affordability increases of SSDs have been due to compromises elsewhere. If you plot the characteristics of "the cheapest flash you can buy" over the years, few of the curves are monotonic.

Physical memory cell dimensions decreased steadily until the arrival of 3D NAND, whereupon cell sizes jumped back up and have been relatively stable since. SSDs incorporate more advanced error correction with each generation of controllers, so today's drives can get more usable write endurance out of the same physical cell size.

Number of bits that can be stored per cell has been increasing. Each time we cram another bit into each memory cell, performance and write endurance drop. But in between those major shifts, successive generations of flash tend to get faster and more energy-efficient over time. Today's 3 bit per cell TLC NAND is much faster than the TLC from 5 years ago, and has better write endurance thanks to the transition to 3D NAND. Today's SLC NAND has comparable performance and write endurance to the planar SLC that disappeared from the market over 5 years ago, but now it's several times cheaper.

And most of the compromises that have been made with NAND flash over the past 10+ years are purely academic. The vast majority of use cases do not need flash rated for 100k P/E cycles, and excess endurance beyond your needs provides no tangible benefit. Most of the endurance that has theoretically been sacrificed was never needed in the first place. Most of the performance lost has been offset by using more flash in parallel and by eliminating bottlenecks elsewhere.

The only metric by which a 5 year old NAS hard drive—or any hard drive—will outperform a modern cheap SSD is in long-term sustained sequential write throughput. Even the slowest modern cheap SSD will offer at least twice the sequential read throughput, and orders of magnitude better random read or write throughput. And it's not like the write throughput of modern SSDs is abysmal: gigabit Ethernet is a more significant bottleneck when filling a 1TB or larger dirt-cheap consumer SSD. So unless you're trying to use cheap consumer SSDs in an expensive NAS with 10GbE, the poor write performance of low-end SSDs won't come into play.

Re:New two-stroke engine design could help keep internal com... by pg-gadfly on Wednesday June 03, 2020 @06:42AM (#23401964)
Of course there has been improvement in absolute terms, but our current state is much more than just that, which skews the results when thinking about where we are. The opinion is definitely not popular on hackernews, but I still want to throw it out there.

Cheap SSDs can tank below 70MB/s after cache, which is less than a gigabit connection, so it can totally affect your performance even in the smallest local networks.

You don't have to sustain writes for long at all to start seeing the compromises hidden in cheap SSDs, especially if they're using their TLC as SLC cache, in which case you run out faster and faster.

Re:New two-stroke engine design could help keep internal com... by Already__Taken on Tuesday May 26, 2020 @07:24AM (#23309729)
I don't think that's true so much, the free valve innovation from koenigsegg looks like a big deal. the myriad of hybrid adaptations to the ice will keep it very competitive for a long time.
Re:New two-stroke engine design could help keep internal com... by Theodores on Tuesday May 26, 2020 @05:30AM (#23308964)
This is a very good analogy. Perhaps the best I have heard. Also SSD was with its own 'range anxiety' in that space was always a lot less in your configured system than the spinning disk equivalent.

However, advances in automotive are also driven by legislation. The three cylinder turbo charged engines you get in regular European cars are there because of the legislation and that cheating diesel doesn't sell any more.

With this continual ratcheting up of legislation there is an incentive for manufacturers to improve the ICE technology. Particularly when efficiency is across the fleet so those inefficient SUVs have to be offset by lean hatchbacks.

This means there is incentive for golden bullets, solutions that promise the earth and allow ICE to linger on for longer. Not many of the legacy auto makers can compete in EV sales. They could make the product but the Tesla is hard to beat. They all fall short except when in the subcompact segment where Tesla don't have a product offering.