LWN: Comments on "Opt Green: KDE Eco's New Sustainable Software Project"

What about the gas guzzlers?

joib — Sun, 22 Dec 2024 07:37:08 +0000

As the sibling comment says, it's about NOx formation. Higher combustion temperature leads to higher efficiency, but higher temperature also leads to increased NOx. Hence the latest generation diesels tend to have worse BSFC than the previous generation, in order to meet the emissions regulations.

What about the gas guzzlers?

farnz — Sat, 21 Dec 2024 19:20:22 +0000

My engine was affected by the recall; like most engines in its family, it gets a little bit over its rated power, and is a little more efficient than promised when on the drive cycle, but the recall letter warned me that, while I would still meet the power and efficiency levels Škoda promised when I bought it, I would lose both power and efficiency compared to pre-recall.

What about the gas guzzlers?

sammythesnake — Fri, 20 Dec 2024 23:41:54 +0000

True, but they're also *newer* engines, and that's meant better efficiency almost every person of the development cycle so far, so I get the impression that the euro 6 emissions requirements might be deleterious to the efficiency of the engines.

Remember the "diesel gate" kerfuffle? The amount of effort that went into fooling the tests suggests that *some* unpalatable compromise was being avoided, and one that could be avoided without removing the whole AdBlue mechanism altogether - efficiency and power (which are related anyway) would be the obvious guesses. I'd be curious to see some careful measurements of engines before vs. after the recall to see what changed other than emissions. I'd wager a shiny penny that they lost a handful of ponies...

I did recently have a conversation with my garage, though, as they're fighting with the exhaust treatment stuff on my euro5 engine that the euro 6 engines are more *reliable* in their experience. I'm also curious about what's behind that...

What about the gas guzzlers?

atnot — Mon, 17 Jun 2024 07:59:34 +0000

Specifically mostly PM and NOx emissions. If you haven't heard the term "acid rain" in a decade, this is why. Or even smog in most cities, outside of the growing numbers caused by wildfires.

What about the gas guzzlers?

gioele — Sun, 16 Jun 2024 23:20:07 +0000

> the Euro 6 engines universally suck to drive compared to their Euro 5 predecessors, and they don't even get better mpg... :-(

The Euro 0...6 standards are _emission_ standards: they regulate how much a car is allowed to pollute, not how much it will consume. Pollution and consumption are strongly connected, but are not the same thing.

What about the gas guzzlers?

sammythesnake — Sun, 16 Jun 2024 23:03:10 +0000

Sadly, based on my experience making deliveries in a variety of vans, the Euro 6 engines universally suck to drive compared to their Euro 5 predecessors, and they don't even get better mpg... :-(

No, nuclear insurance is required and available

malmedal — Sat, 15 Jun 2024 19:13:49 +0000

Corbet has asked us to stop, apologies for replying yet again.

This is basically zero compared to a damage assessment of a realistic accident scenario, a symbolic sum that the industry agreed to for forms sake.

Elon could pay that much for a joke.

No single person is rich enough to pay for the damage a serious nuclear accident could cause.

No, nuclear insurance is required and available

sdalley — Sat, 15 Jun 2024 18:22:06 +0000

> Not a single plant would be built, ever, if the owners had to buy insurance.

Countries vary, but typically require the utilities to buy insurance, contrary to antinuclear myth. See, for example,
https://nuclear-risk.com/products-and-services/ and
https://www.amnucins.com/about-ani/anis-insurance-pools/

USA Price-Anderson Act requires utilities supplying nuclear power to have primary insurance of $450m per reactor, and also, in case of an accident above this level, requires themselves, not the government, to pay up to $122m per reactor into a mutual no-fault liability pool, bringing total cover up to ~$13 billion for a single accident, after which the government would step in. This provides a powerful financial incentive to the industry to run a clean ship.

Total accident claims over the history of the Price-Anderson Act are around $151m to date. The Three Mile Island accident resulted in about $75m of this.

What about the gas guzzlers?

kleptog — Fri, 14 Jun 2024 22:10:24 +0000

> What should happen (and happens accidentally on my home array) is I have an old-fashioned electricity meter that runs backwards when my panels are generating a surplus. So I actually only pay for my net consumption.

Absolutely not. The grid has one requirement: power generated = power consumed at all times. If you are generating and consuming power and not making any effort to correlate the two, then you are asking someone else to do it for you and they obviously want to be paid for that. Which happens now via the difference in pricing between your production and consumption.

The different between (predicted) production and (predicted) consumption has to be rectified by the energy imbalance markets, because while we can predict production to some extent, predicting consumption is obviously also not 100% perfect. In NL households without solar panels cost energy companies €4-8/MWh in imbalance costs, but with solar panels €32-65/MWh. These costs used to be distributed over all customers, but there are now moves to charge solar panel owners directly, unless they take steps like disconnecting their solar panels when the electricity price goes negative, or installing a battery.

There's no doubt we could be much smarter about it, with community batteries and allowing you to match your production with your neighbours consumption, but those technologies are only just starting to be rolled out. People who think they can just install solar panels and call it a day massively underestimate the complexities of the electricity grid.

What about the gas guzzlers?

malmedal — Fri, 14 Jun 2024 17:50:58 +0000

> I think, on all three of my solar arrays, I get paid for feeding in to the system roughly 1/3rd what I get paid for taking out. Which encourages me to provide my own (inefficient) batteries etc to try and "save" energy. DAFT!

Sounds reasonable, or even generous depending on your local neighbourhood. If your neighbour is using the power you create, that's reasonably efficient. But if it has to go back out through the nearest transformer most of it will be lost...

> Given sufficient notice, there's no reason why nuclear can't be ramped up and down even over a short time period.

No, current nuclear plants can't ramp up and down even once per day.

> switch the domestic gas network to hydrogen

You can add maybe 20% extra hydrogen to most existing gas networks. Hydrogen molecules are small and will leak from pipes that are designed for natural gas.

However, there's a company planning to generate hydrogen and capture CO2 from the air and create methane and even heavier hydrocarbons from that. We'll see if they can do it cheaply enough.

What about the gas guzzlers?

malmedal — Fri, 14 Jun 2024 17:29:34 +0000

> We have pilots, but we have _0_ countries where this technology is proven to be able to store sufficient energy from renewables to be able to provide base-load capacity to grids.

We have many countries, one of which is the US, where this is proven. In terms of instantaneous power they are already as large as needed, you don't want a single connection to be too big because the grid needs to be table to tolerate any single site dropping out due to a transmission line being cut.
In terms of energy stored, there is of course no limit.

> Decades of ignoring the fact we have a near-0 carbon (most of the carbon is in the concrete of the build probably) energy source that is capable of providing _abundant_ reliable energy. Decades of continuing to burn coal and gas instead.

Obama gave the nuclear industry everything they said they needed, nothing happened(almost, a few got started if you want to be picky) , then nothing happened in four years under Trump.

Currently the nuclear issue is just used as a partisan cudgel by the republicans. They have absolutely zero interest in actually doing anything.

> France has quite reasonably priced electricity.

Around 0.20 euro per kWh, I believe, similar to the EU average. Less than e.g. the UK and more than hydropower-countries to the north.

A solar plant in France should deliver power at 0.01 euro/kWh or less. Nuclear simply cannot compete.

> And, have you even factored in the cost of the large scale battery banks you advocate for into your solar costs?

Yes. It keeps dropping, currently around 75 dollars per kWh capacity, if I remember correctly it was 140 two years ago.

Nuclear in France

farnz — Fri, 14 Jun 2024 16:55:42 +0000

EDF are now operating on a commercial basis, but their fleet of nuclear plants was gifted to them at privatisation, and the state has agreed to cover the cost of insuring the plants against major incidents other than those caused by operations, and the cost of decommissioning the plant at end of life. This is a major subsidy to nuclear.

Your point about hydrocarbon plants is reasonable, but note that once you account for the subsidy they're given in terms of being permitted to emit CO2 without penalty at the price that Climeworks and other such companies can currently remove CO2 at a profit, hydrocarbon plants come off more expensive than nuclear, but note that renewables and sufficient battery storage to cover the base load comes out cheaper than nuclear for most inhabited parts of the planet (some parts of Siberia are an exception).

And that's the problem with nuclear as it exists today; if you account for the subsidies, while it's cheaper than fossil hydrocarbon plants, it's more expensive than the combination of wind and solar with NiMH or Li battery chemistries to cover peak use. This only gets worse when you take into account the ability to site batteries near both demand and supply, reducing the need for expensive transmission network upgrades, whereas siting nuclear or hydrocarbon plants in residential areas is problematic simply because they're industrial plants (ignoring fears around nuclear waste etc).

Note, too, that the energy industry has a huge blind spot when it comes to small-scale battery and renewable installations; in most of the world, any battery or renewable installation that's not connected to the high voltage grid (132 kV and above for my location) is treated as part of demand for energy, not part of supply. As a result, we're fairly confident that the projections that say you "need" nuclear for base load are completely and utterly wrong - they're based on the assumption that every single house you see with solar panels on it does not generate any electricity locally, and that batteries attached to houses (Tesla Powerwalls and similar) charge from the grid and turn that energy into heat, rather than supplying it (minus losses) to the house at a later point in time. These are known to be bad assumptions, but actually changing them runs into quagmires about how exactly to account for them.

Disclaimer: I actually work in the electricity sector, and I'm basing some of the above on industry insider information, not on public sources.

Nuclear in France

paulj — Fri, 14 Jun 2024 16:31:59 +0000

Are there any major power generation systems that do not get build-out subsidies? E.g., solar in the market I'm familiar with currently has extensive subsidies available to cover initial capital costs, both for consumers and commercial projects.

Further, if we want to compare a hydrocarbon plant built without subsidies, to a nuclear one paid for by the state, are we taking into account the fact that the market as it stands generally have no way to price in the cost of the CO2 emissions. I.e., the HC plant is getting that for free - effectively a subsidy too, against the ongoing and future costs incurred to the rest of society by climate change from CO2. Many arguments I've been in do not account for that cost.

If CO2 is going to cause catastrophic damage to large parts of the planet, then the fair present cost of CO2 emissions is basically infinite. If a near-0 carbon electricity system requires nuclear to achieve it, then any subsidy TO ANY LEVEL is _worth it_!

For France, in terms of the annual operational basis I thought EDF were running on a commercial basis. They have an agreement with the french state to provide a certain amount of electricity at minimum, capped to a certain price, and they have to meet their operational costs and extract profits within those parameters. That was my understanding when I went and read EDF reports on this on their costs, as part of a similar debate I was having with someone else.

Who cares if it took subsidies to build the plants. Unless you believe there exists some monetary value past which the damage of CO2 emissions is OK, then this is irrelevant - given the technology available today that is proven to be able to provide reliable, abundant base-load capacity.

On the US market question, you're saying basically everyone else is an outlier. Maybe that's true on "unsubsidised" energy markets, but.. who cares about that? If:

- nearly all energy markets across the world work on a "state subsidises builds in order to get sane planning" (not least because true competition in energy networks and physical delivery is basically impossible; along with market difficulties in building plants in a pure market-competition way)
- every energy market simply is dysfunctional anyway simply on the basis of having no (or near no) ability to correctly price in the societal/global cost of CO2 emissions (which then can only be corrected by state planning, likely by use of subsidies if it wishes to devolve implementation to commercial companies)

Then the "pure" market with no subsidies is indeed both dysfunctional and inefficient, on that view. And the US is indeed the outlier.

What about the gas guzzlers?

corbet — Fri, 14 Jun 2024 16:25:35 +0000

We are getting increasingly off-topic here, perhaps it's time to wind this one down.

What about the gas guzzlers?

Wol — Fri, 14 Jun 2024 16:19:36 +0000

Depends which greens.

There was an almighty fuss over the Peace lot and a defunct oilrig, and while I've no doubt the oil company was looking to save money, there was also an extremely good eco argument for what they wanted to do. The Greens' attitude came over as "we don't care why it might be a good idea, you can't do that!"

I've always been extremely wary of them since.

It's like the anti-hunting brigade. Most active conservationists hate them, not because they agree with hunting (they don't), but the hunting brigade are prepared to PAY for conservation, precisely because they want stuff to hunt! They make unlikely bedfellows, but pragmatism pushes hunters and conservationists together.

Cheers,
Wol

What about the gas guzzlers?

rschroev — Fri, 14 Jun 2024 15:27:33 +0000

In my experience it's not the greens who are betting on unproven technologies; it's more the people who think we all can carry on without any chance, and technology will save us.

Greens, in my experience, advocate using as few of our resources as possible, and as efficiently as possible.

Nuclear in France

farnz — Fri, 14 Jun 2024 14:43:31 +0000

I happen to know a bit about the French nuclear market, and it's also an outlier when it comes to the cost of nuclear, because the vast majority of their reactors were subsidized by the DGA - France wants to be self-sufficient when it comes to producing nuclear weapons, and in the 1960s and 1970s that meant having domestic reactors that can be repurposed to provide bomb-making materials. It also has a second tier of subsidy for nuclear power because it wants to be self-sufficient in energy, and doesn't have enough fossil fuels to make that practical without nuclear.

Similarly, Japan and China are both outliers because their national governments are willing to heavily subsidise nuclear power in order to avoid having to import power or fuel from outside the country; in Japan's case because it's nuclear or imports, and in China's case because they're aware that burning fossil fuels is a massively time-limited exercise and want to be ready for when it's no longer practical to do so.

Indeed, the only country I can find that has no direct state subsidy for nuclear power, and yet a significant civilian nuclear power industry, is the USA. Even other countries with a dysfunctional nuclear industry (like the UK and Germany) have state subsidy for nuclear.

I have theories as to why this is, but this is very off-topic for LWN, so I'll stop here.

What about the gas guzzlers?

Wol — Fri, 14 Jun 2024 14:18:20 +0000

And I'll simply throw in the crazy dysfunctional (anti)incentives for solar in the UK.

I think, on all three of my solar arrays, I get paid for feeding in to the system roughly 1/3rd what I get paid for taking out. Which encourages me to provide my own (inefficient) batteries etc to try and "save" energy. DAFT!

What should happen (and happens accidentally on my home array) is I have an old-fashioned electricity meter that runs backwards when my panels are generating a surplus. So I actually only pay for my net consumption.

The other thing in the UK (and I'm sure other countries could build them) is we have one or two hydroelectric "batteries" where excess base load at night is used to pump water up a mountain, and then it's released at times of peak demand. Given that those times used to be commercial advertisement breaks when everybody put the kettle on, and those days are long past, I'm sure they would go a long way to reducing our need for gas or whatever to iron out fluctuations in wind and solar.

(I think they are 500MW stations, and can go from 0 to full power in 30 seconds ...)

Given a decent weather forecast we can predict renewable generation. We can predict electric demand. Given sufficient notice, there's no reason why nuclear can't be ramped up and down even over a short time period. And then we've got one or two of these huge batteries to allow for differences between forecast and reality.

(Of course, the other thing to do is switch the domestic gas network to hydrogen (or in the interim a hydrogen/methane mix), and store surplus generation as hydrogen in the existing gas storage network!)

Cheers,
Wol

What about the gas guzzlers?

paulj — Fri, 14 Jun 2024 14:00:09 +0000

We have indeed wasted far too much time - many decades now, given I've been discussing climate change with people since the 90s. And a recurring theme I get from "greens" is they propose some _future_, *unproven* technology as being the reason for not building out nuclear _now_. From "carbon capture" to (now) mass battery storage, these are _unproven_ technologies at scale. We have pilots, but we have _0_ countries where this technology is proven to be able to store sufficient energy from renewables to be able to provide base-load capacity to grids.

Nuclear is proven. We have countries with energy surpluses, cause they built-out nuclear decades ago, instead of going with this "But but <handwave about some latest great hope technology that, at best, has PoCs but no proven at scale deployment> means we can don't nuclear!". Along with "But nuclear would take a decade to build!".

And what is the end-result. The end-result is that I have been listening to this for _decades_ now. Decades of ignoring the fact we have a near-0 carbon (most of the carbon is in the concrete of the build probably) energy source that is capable of providing _abundant_ reliable energy. Decades of continuing to burn coal and gas instead.

The "Nuclear is expensive" argument is flawed in 2 ways:

1. It is typically based, in whole or part, on the US market. Except the US market appears to be incredibly dysfunctional and inefficient, for whatever reason. The US is an _outlier_ on cost. If you instead look at France or China - whose energy companies are the ones most of the rest of the world would bring in to build and manage new nuclear - the cost is quite competitive.| France has quite reasonably priced electricity.

2. Second flaw is comparing nuclear energy to cheap renewables. Yes, solar is cheap. But the whole reason to look at nuclear is that solar (and other renewables) are unreliable and hence _require_ some other energy source for base-load generation. You can not compare the two. Cheap solar power is no good at night.

And, have you even factored in the cost of the large scale battery banks you advocate for into your solar costs?

What about the gas guzzlers?

malmedal — Fri, 14 Jun 2024 12:59:23 +0000

> If you are serious about being green, your answer can not be "use gas forever".

That is indeed true, now. We have wasted too much time, it would have been feasible it we had acted as reasonably as soon as the science was clear, 40 years ago.

Given your previous comments, I assume you suggesting nuclear as base-load. Apologies if I'm wrong.

Unfortunately, nuclear is a very bad complement to renewables. The main issue that they are expensive, second is that current designs are very bad at ramping up and down which is what you need to compensate for variation in renewables. They simply are not capable of changing their power output daily, much less multiple times per day.

It is theoretically possible to build reactors which can vary their output frequently, but nothing that has worked has been built yet. Various promising designs have been built, e.g. molten salt reactors and pebble bed reactors, but as far as I can tell nothing is ready for commercial operation.

Reactors in US hangar-ships and subs are better, though still not good enough, they are even more expensive per kWh and they operate on weapons-grade fuel which adds considerably to security requirements.

Fortunately it looks like batteries are going to get good enough to handle it all, they are already good enough and cheap enough to even out the solar day to day cycle and I expect they will eventually be good enough to handle long-term seasonal storage.

What about the gas guzzlers?

malmedal — Fri, 14 Jun 2024 12:10:37 +0000

> We import fossil fuels for the energy grid.

The price of solar plus batteries is so low that you will save money running the grid off that.

I can't think of a possible country where this wouldn't be true, countries with land in the far north, US, Canada, Norway, Russia are all oil-exporters.
Also except for Russia not considered part of the Global South.

Let's see, Greenland? No apparently they export oil. Sweden, Finland? Still not Global South, and they run their grids mostly on renewables and nuclear.

Argentina? Even in Ushuaia solar is still feasible, but I recall it being rather windy, so wind-power might very well be the best option.

What about the gas guzzlers?

paulj — Fri, 14 Jun 2024 11:59:58 +0000

Solar is amazing, and should be a big part of future energy production. However, it's not perfect either. It is land-intensive. It's also not reliable - some parts of the world worse than others. There is daily variation, and there is seasonal variation - especially in higher latitudes.

We will need a reliable energy source to provide a base capacity for our grids, to complement solar and other renewables (wind, etc.). If you are serious about being green, your answer can not be "use gas forever".

What about the gas guzzlers?

malmedal — Fri, 14 Jun 2024 11:41:03 +0000

> FWIW, nuclear is and will remain the greenest energy source

This is not true, but also it is not the right metric.

The metric you want is cost per kWh. By that metric solar is the clear winner.

> The fact some environmentalists don't like [nuclear]

You know who *really* don't like nuclear? Try asking an insurance-company how much it would cost to insure your nice new nuclear power plant.

Not a single plant would be built, ever, if the owners had to buy insurance.

Fortunately for them, many governments, including the US, are actually willing to waive the insurance-requirements and assume the risk themselves.

Unfortunately, even with that massive subsidy the nuclear plants are still way more expensive than solar.

What about the gas guzzlers?

mrugiero — Fri, 14 Jun 2024 10:01:41 +0000

> This is going to be extremely beneficial for most of the Global South, that actually are in the south, they can stop importing expensive fuel for their cars.

The global south would need more than electric cars to be anywhere near benefited by this. I live in one of those countries.
We import fossil fuels for the energy grid, in part because of self-inflicted problems, in part because nuclear treaties put a cap on Uranium purity unless you are one of the few countries that actually used it for weapons, which means our plants are more expensive to run and less efficient. We also use gas for heating because our electricity production is no match to just burning gas. Last but not least, you need wealth to buy new vehicles, and we lack that.

What about the gas guzzlers?

mrugiero — Fri, 14 Jun 2024 10:01:35 +0000

FWIW, nuclear is and will remain the greenest energy source. The fact some environmentalists don't like it has more to do with panics about accidents and waste management (I guess it's worse to put it in specific restriction zones than it is to simply blow it to the air and water as most other sources, heh).
Re: EVs, just use trains and walk for short distances. EVs really solve self-imposed problems for people who is either too lazy, too hasty or live in poorly designed cities. That said, it _may_ be better that any _new_ vehicles are electric, but forcing the change on working ones only makes things worse for things already mentioned.
Re: subsidies: no idea about Europe, but I live in a country that heavily subsidizes fossil fuels (they do subsidize electricity too, I don't happen to know which one is more subsidized).

What about the gas guzzlers?

mrugiero — Fri, 14 Jun 2024 09:40:25 +0000

Sometimes it's about specs. If you have a 10 years old system that was top of the line (even mid-range might do the trick) back then you'll be fine today. If you have a more average/budget/third world setup, you may experience the bloat more strongly.
It also may have to do with expectations (for some people waiting a minute is an opportunity to go fetch water and for other an insult) or what specifically you browse. Social networks, for example, are big offenders. In my case the worst ones tends to be newspapers that won't let you see anything without tons of JS. One in particular would fire up the fans of my then-new-high-end laptop. It's cooler to run Unity than it is to read the news, not exaggerating. Some people choose to or are mandated to use heavy sites for work as well. For example, a friend of mine appreciates the functionality of Notion, while at my job it's the official documentation platform, and at a previous job it used to be Jira. I certainly don't appreciate the bloat/functionality ratio of Jira. Notion is heavy, but I haven't seen anything as flexible before, so I guess it's not _pointless_ JS, though I don't really use it as more than a markdown platform in practice.

I applaud this, but...

mrugiero — Fri, 14 Jun 2024 09:24:39 +0000

That is true, but even then there can be mitigations your system can apply if it knows you're running in older hardware (for some definition of older). For example, a browser for older hardware will assume less available RAM, so the tradeoff memory vs CPU that leads to using several stages of JIT compilation when memory abounds may be more biased toward either baseline or interpreter only, or something in between (e.g. a threaded/context threading interpreter with a single instance for opcodes).
Also, considering many local applications are now just a browser running a web app, one mitigation may be not running a browser with a web app when you don't mean to. Some of the popular ones have alternatives (some better, some worse), e.g. for Discord there are a few native apps that are more sympathetic to hardware (although the one that worked the best for me is proprietary, it's not like the original is free software either).

Opt Green: KDE Eco's New Sustainable Software Project

mathstuf — Fri, 07 Jun 2024 14:06:16 +0000

Note that Microsoft also seems to have a form of triskaidekaphobia in that, AFAIK, they've never released software with major version number "13". Maybe that's the end of the road unless they change back to year-based versions after "12" to avoid having to explain why "14" is its followup?

Opt Green: KDE Eco's New Sustainable Software Project

taladar — Thu, 06 Jun 2024 07:18:47 +0000

Windows 9 was specifically unused to avoid issues with software checking for Windows 9* in version strings to check for Windows 95 and 98

Opt Green: KDE Eco's New Sustainable Software Project

NYKevin — Thu, 06 Jun 2024 01:19:09 +0000

If you want an official statement, you can find one in the article linked upthread:

> "Recent comments at Ignite about Windows 10 are reflective of the way Windows will be delivered as a service bringing new innovations and updates in an ongoing manner, with continuous value for our consumer and business customers," says a Microsoft spokesperson in a statement to The Verge. "We aren’t speaking to future branding at this time, but customers can be confident Windows 10 will remain up-to-date and power a variety of devices from PCs to phones to Surface Hub to HoloLens and Xbox. We look forward to a long future of Windows innovations."

No, they did not promise that Windows 11 would never happen ("We aren't speaking to future branding at this time"), but that's a totally different question from whether Windows 10 would remain supported indefinitely. Microsoft's statement outright says that Windows 10 "will remain up-to-date." That is a specific promise, and while it was probably unwise of Microsoft to make such an open-ended statement, that's their problem.

Opt Green: KDE Eco's New Sustainable Software Project

cesarb — Tue, 04 Jun 2024 19:49:35 +0000

> There's very little the browser can do given that the information being leaked isn't restricted to processes controlled by the browser. Heck, some of these attacks can leak information across *virtual machines*.

The browser is not alone, it works in concert with the kernel, and there are lots of tricks the kernel can use to mitigate these vulnerabilities (both between userspace and the kernel, and between the sandboxed JS process and the other browser processes) even without firmware help.

> Even at the OS level, these "mitigations" generally consist of completely disabling hardware features, usually with _severe_ performance impacts. If said features can even be disabled at all.

The main "completely disabling hardware features" mitigation is disabling SMT. Other than that, most non-microcode mitigations are AFAIK code to block speculation (or make it harmless) at key points, which does have some performance impact, but not severe enough to make the hardware unusable. If you can accept some performance degradation, you can have good enough security even after the hardware maker stopped support for that hardware.

Opt Green: KDE Eco's New Sustainable Software Project

cesarb — Tue, 04 Jun 2024 19:35:27 +0000

> Just wanted to follow up on this -- If you have to disable SMT to mitigate against a vulnerability, that's going to result in about a 33% performance impact, meaning you'll need roughly 50% more resources [1] to achieve the same overall performance. 50% more servers, 50% more datacenter space, 50% more power (and cooling), etc.

Most servers aren't running untrusted code, and so far I have heard of no demonstration of speculative execution vulnerabilities being exploited solely through a network connection. The impact on web browsers (which do run untrusted code) is lower, since client computers are idle most of the time and thus have resources to spare; there might be an impact on power use, but there's a difference between "the battery lasts less time" and "hopelessly insecure".

Opt Green: KDE Eco's New Sustainable Software Project

farnz — Tue, 04 Jun 2024 15:16:46 +0000

A 50% performance boost for SMT would be very surprising on a well-designed core. 5% to 20% is a more normal range for an Out of Order Execution (OoOE) design, since a good OoOE design is set up so that one thread can utilize all of the resources available, and your workload will tend towards bottlenecking on the same resource for all threads (with SMT's performance boost coming from the fact that you hit the bottleneck at different points on different hardware threads).

Opt Green: KDE Eco's New Sustainable Software Project

adobriyan — Tue, 04 Jun 2024 14:18:15 +0000

> [1] Assuming SMT gives you a 50% performance boost on average, losing it means your net performance will be approximately 2/3rds of what it was with SMT. (SMT gains are highly workload dependent)

Kernel compile is only ~10% faster with HT.

Opt Green: KDE Eco's New Sustainable Software Project

pizza — Tue, 04 Jun 2024 14:09:14 +0000

> Even at the OS level, these "mitigations" generally consist of completely disabling hardware features, usually with _severe_ performance impacts. If said features can even be disabled at all.
> Whereas with these microcode patches, the OS has the option of using considerably higher-performance mitigations, if anything is needed at all.

Just wanted to follow up on this -- If you have to disable SMT to mitigate against a vulnerability, that's going to result in about a 33% performance impact, meaning you'll need roughly 50% more resources [1] to achieve the same overall performance. 50% more servers, 50% more datacenter space, 50% more power (and cooling), etc.

Or you can use a combination of microcode patches (possibly in combination with OS awareness) to bring the net performance hit to ~5%. It's a win-win for everyone. [2]

[1] Assuming SMT gives you a 50% performance boost on average, losing it means your net performance will be approximately 2/3rds of what it was with SMT. (SMT gains are highly workload dependent)
[2] If you run truly private/trusted workloads, you're going to disable all of these mitigations as part of your performance tuning anyway.

Opt Green: KDE Eco's New Sustainable Software Project

pizza — Tue, 04 Jun 2024 13:36:49 +0000

> But Spectre (and similar) is limited; it cannot do much more than leaking in-memory data.

When that "in-memory data" includes crypto and session keys, authentication credentials to external third party-run services, leaking said "in-memory data" can give an attacker the literal keys to your non-local kingdom.

> mitigations in web browsers and operating systems can reduce some of its impact (at a performance cost) even without help from the CPU microcode or platform firmware.

There's very little the browser can do given that the information being leaked isn't restricted to processes controlled by the browser. Heck, some of these attacks can leak information across *virtual machines*.

Even at the OS level, these "mitigations" generally consist of completely disabling hardware features, usually with _severe_ performance impacts. If said features can even be disabled at all.

Whereas with these microcode patches, the OS has the option of using considerably higher-performance mitigations, if anything is needed at all.

Opt Green: KDE Eco's New Sustainable Software Project

cesarb — Tue, 04 Jun 2024 13:08:06 +0000

> However in more recent times, many of these data-leakage exploits require no elevated privileges, and real-world attacks have been demonstrated using nothing but javascript running in a web browser.

Yes, Spectre is one of the "rare exceptions" I was thinking of (vulnerabilities in out-of-band management systems like Intel AMT is the other, since they allow remote access which completely bypasses the operating system, and it's hard to even know whether they are enabled or not).

But Spectre (and similar) is limited; it cannot do much more than leaking in-memory data. It cannot modify data, it cannot leak "cold" data which is only on disk, and mitigations in web browsers and operating systems can reduce some of its impact (at a performance cost) even without help from the CPU microcode or platform firmware.

What about the gas guzzlers?

paulj — Tue, 04 Jun 2024 10:33:26 +0000

Light that detect cyclists.... cry.

Here in the Celtic Isles our lights are typically modulated by pressure sensors embedded in the road, to lengthen or shorten green cycles according to whether there are cars going through a car or not and whether there are cars waiting at red on other sides. These typically sensors need at least 150 kg to trigger, according to a technician I spoke to a while ago, while waiting at lights (this was Ireland, DCC).

This is _infuriating_ if you're on a bicycle. And it means I end up ignoring red and just using common sense if there's no car behind me.

According to that technician, they can use sensors that trigger at a much lower pressure and detect cyclists, but they generally never use them.

What about the gas guzzlers?

LtWorf — Tue, 04 Jun 2024 07:51:00 +0000

Consider that:

1. garages and underground parking exists
2. city streets, especially where narrow (typical in southern europe) get sunlight only 4 hours a day at most
3. I think you're considering the energy created by optimally oriented solar panels. Not badly oriented ones that sit in the shade for hours
4. As the owner of a 5 years old hybrid vehicle. At this point the batteries are essentially dead weight that I pay more fuel to move around for absolutely no advantage.

I take it you've never spent a summer in southern europe. But as a british with no experience you feel nonetheless compelled to teach the sicilian how it's not a problem to park in the sun in the summer. I guess you can tolerate 70° and up better than me!

Also overheated batteries can explode.

What about the gas guzzlers?

mathstuf — Tue, 04 Jun 2024 01:52:28 +0000

> leave the sun roof slightly open. It dumps a lot of heat :-)

Though, being on the English Isles, I'd expect rain to be a constant threat to turn that into a "rain roof" (it's a threat here too as northeast US weather can be quite…indecisive). Not to mention pollen season. There are guards for side windows to allow for a slight opening while protecting from everything short of actual-horizontal rain. It even allows for a cross-breeze.

Anyways, cars are already growing gadgets at an alarming rate. Even more things to go wrong doesn't sound great to me. With all the sensors around cars these days, even a fender bender can result in a "total it" conclusion :( . With extension to parts that are inert metal even in Teslas being "safety critical", what's to be done with comprehensive insurance coverage premiums if a bird strike can result in thousands of dollars of repairs and a potential "total it" result?