Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?

Comment Video drivers (Score 2) 57

So has the problem with getting the latest video drivers been fixed?

Depends on the graphic card.

TL;DR: For Intel and AMD, you really don't give a fuck what you're running.
(Or if you're into gaming, you *absolutely* try to run the latest possible kernel and Mesa combo to get all the latest belss and whistles).

- Intel is exclusively doing opensource drivers (at least for their own cores. For the core that they sub licensed from PowerVR several years back, it's an entirely different story).
The drivers are part of the upstream kernel, so getting a new kernel *IS* how you get a new driver.

- AMD has embraced opensource (beginning almost as far back as when they bought ATI - though the actual implementation has taken much time).
They have a dual offering.

For a few special user (CAD software users, etc. who require opengl profiles) they still provide a closed source driver called AMDGPU-PRO. (As of note, it also shares some code with the current Windows drivers).
For the rest, the opensource source drivers (based around Mesa) are their main target. They have been investing resources (some of the devs of the opensource Mesa drivers are on AMD's payroll). Recently, it has seen tremendous improvement and starts to beat the closed source driver on most use-cases.

Both above driverrs rely on the same opensource kernel module. So if you use the opensource driver, the situation is the same as Intel : getting the latest kernel *IS* how you get the latest kernel module.
Only for the closed source AMDGPU-PRO does AMD spends ressource back-porting the kernel module to older kernel.
(So if you are a CAD user, better stick to major distribution whose kernel are more likely to see such back-ports. Or move to a rolling release distro, but that would be unusual for this class of users).

The current only exception is Vulkan. AMD haven't finished the necessary work to opensource their official Vulkan implementation. So AMDGPU-PRO is the only way to get it.
Meanwhile, opensouce developers have created RADV, their own unofficial Vulkan implementation for AMD cards. Currently, this implementation finally passes the conformancy tests and implements the Vulkan API completely, but isn't optimized yet. So often games run better using their openGL back-end (given that massive effort mentionned above) than using their Vulkan back-end (because as mentioned here, it's a small project, which only recently achieved full conformancy and hasn't even started optimizing).

regarding other API :
AMD has recently opensource a new OpenCL implementation running on top of their ROCm computing platform.
Now that this is opensourced, expect the optimizing to go faster, and eventually reaching the point that OpenGL has reached.

Over all the days of the absolutely aweful fglrx closed-source ATI driver are distant past.

I remember having to find the proper kernel headers (down to ver.a.b-c-ubuntu) when updating to latest nVidia OpenGL driver EVERY FUCKING TIME.

Bad news for you : Nvidia is still doing closed source drivers.

They basically recompile their Windows drivers for Linux, and write their own closed source "shim" kernel driver.
Meaning that you're completely dependent on Nvidia condescending to port their code to your specific kernel.

Also don't expect features that normally work with this hardware on Windows, and that other manufacturer have successfully implemented on Linux to work for you.
(Again, remember : Nvidia is basically recompiling their windows drivers. Some features have an API under Linux, but it's largely different and Nvidia doesn't botter)
(eg.: see who long it took them to properly implement xrandr for output handling)
(eg.: remember the whole fiasco around dual embed+discrete GPU on laptop working badly. Culminating with the public "Fuck you, Nvidia !" by Linus about their lack of collaboration)
(eg.: it's 2017 already, and Nvidia still hasn't got their shit together to have proper kernel-modesetting like everybody else, which has bad implication on modeswitching and/or sleep/wake-cycling).

Best to stick to a big popular distro, and keep that distro's main official kernel, to be sure to get the most acceptable Nvidia closed-source distro.
And hope that it actually work.
(Disclaimer: my laptop is basically unusable with Nvidia's official drivers).

Meanwhile, completely unrelated volunteer are trying hard to reverse engineer the hardware and write independent opensource drivers (nouveau).
None of the developer is in any way related to Nvidia. (unlike Intel and AMD who have open-source devs on their payroll).
They don't receive much help from Nvidia (once every blue moon, an engineer will throw them a bone and give some docs. Usually when it helps further Nvidia's Tegra on Linux agenda).

But at least nouveau is part of the upsteam linux kernel, so when on a rolling distro (I'm on Suse Tumbleweed) you're sure to get the latest driver in your latest kernel.

Expect the occasional quirk (it's reverse engineered driver. not designed using official specs).
And far from stellar performance (usually newer hardware completely lack re-clocking).

(But at least it makes my laptop work better).

Comment LTS alternative (Score 1) 57

Might be wise to get yourself acquainted with how RH manages kernel. {...} Since then it received numerous patches {...} mostly security and bug fixes

So, a sort of LTS kernel (though the concept wasn't as formally organised back in the 2.6.xx era.

Which brings us to another solution which wasn't mentioned in the summary :
If you don't want to move *forward* to kernel 4.11 (I don't know maybe there's a regression or a dropped support that affects you),
you can move *backward* to kernel 4.9 which is a LTS version and is going to be supported for quite a few years.

Comment Emacs. (Score 1) 57

systemd has everything I need.

A yeah, I think I've heard about that one.
It's a module that runs inside emacs, isn't it ?
(Though not as popular as the kernel module inside emacs that FSF decided to use instead of Hurd).

Comment LTS kernel (Score 1) 32

No, 4.9.25 is less than a month old, and will continue to receive updates until January 2019.

4.10 came after 4.9.0, but 4.10 isn't an LTS version, so it's supported only until the next "unstable" version becomes stable

And that was even mentioned in the summary of the article about 4.9 that the above poster speaks about.
That summary begins with :

As it's not an LTS (Long Term Support) branch, the Linux 4.10 kernel series was doomed to reach end of life sooner or later,

the explanation is literally the first few words of the summary.
But I guess that "didn't read the summary" is the new "didn't read the article".

Comment Depends on the sector (Score 1) 115

Are these posters active in the workforce? Every relevant office in the world uses windows.
{...} But out here in the functional world, windows is everywhere.

Depends of the field you work.
Academic research ?
Specially in fields like computational biology ?
It's going to be exclusively UNIX.
With Mac OS X being a bit more popular on the laptops and workstations of the researchers,
and Linux having monopoly on the servers and compute nodes.

Comment Currently ? Not working. (Score 5, Interesting) 580

Suppose you assigned an AI to observe which user stories go in, and what code comes out as a result. How many programs would you have to complete before the AI is able to take over a majority of the work involved in building an application? {...} I'd honestly be surprised if they aren't already doing something like this.

Yes it's done. Not by google, but by others.
The short answer is that the deep neural nets produce texts that looks like code on the first glance, but doesn't even compile.
e.g.: The variables aren't even properly declared. it can write a formula (like "a = b + c")
but isn't even able to realise the link with the declaration of the variable (that the "int a;" 10 lines above is linked to the "a").

The problem is the size and complexity of modern AI.
The size of the context they can consider,
the amount of abstract models hidden behind the code, etc.

Currently what AI has managed to recreate with deep neural nets, is on the level of WW2's Pigeon guided bombs.
i.e.: leverage some image recognition net and similar basic tasks, and string a few together.

The complexity required to write actual code is several orders of magnitude bigger.
Even some humans can't do it reliably, and you hope to do it with what currently is the equivalent of the visual cortex sub-part of bird's brain.
Good luck with that.

Before achieving that we need :
- more raw processing power (you'll need way much more neurons that currently used in today's deep neural nets)
- advances in science to better understand how to combine together tons of such "function specific nets" to build a higher level of AI.
(the same way a brain is a sum of lot of small specific region, each linked to a higher level/more abstract associative layer).

Comment Consensus government (Score 5, Interesting) 468

Stop treating government as a sporting contest where you cheer for your team to crush the other team.

Some countries like Switzerland have not a single head of executive, but a council of 7 people.

It the same was practiced in your country, that would lead to totally different campaigns.
It would be very difficult for the candidate of one or the other of your bipartite system, to spend time arguing that the other is "an incompetent idiot", because with such system, they are guaranteed to then later have to work together reaching a consensus.

Hard to crush a team, when all the team *must in practice* work together.

The only big suspense would be who out of the minor parties are going to get the last of the coucil seats.

Comment Details. (Score 1) 482

Yeah they save 700lbs by using aluminium instead of steel, but then they go and add about 1000 lbs of crap you don't need

TFA mentions the *total vehicle weight* has gone down from 4000lbs (~2000kg) to 3400 (~1700kg), extra crap included.

like heated cupwarmers,

what the fuck is a "heated cupwarmer" and how does it weights several kg ?
(or maybe it is only "a thing" in the US and that's why I've never heard about it).

56 airbags

pressurized gaz canisters and firing mecanisms don't nearly weight as much as you think.
so it doesn't count much on the total vehicle weight.

Also something that can reduce risks of critical wounds in case of collision (as long as you also wear the necessary seatbelt) doesn't count as "useless" in my book.
Better add a few hundreds gram and avoid ending up with a broken skull or rib-cage with pneumo-/haemo-thorax.

and and a shitload pointless tech that keeps phoning home

big news for you : modern day electronics are suprisingly compact and lightweight.
Your smartphone actually contains much more stuff (screen, big lithium battery) than needed for tech that can phone home.
Also the power envelope of such electronics ( and basically means you can drive like a clueless distracted retard and the car will just sort it out for you.

(btw: adaptive cruise control, lane departure alert, and other such tech doesn't actually need to phone home to work.
it requires a very fast feedback loop, and a cloud-processing and unreliable cell uplink aren't acceptable.
the image processing actually run on chips and computers *inside* the car).

For the weight and power budget, see above.

And again (similar to airbags), a technology which helps avoiding that collision happens in the first place isn't pointless in my book.
it's a very nice additional security feature, both for the safety of the above distracted retards AND for the safety of all the innocent by standers who might find themselves on their collision course (this could be even you or me).

I find the extra couple of watts and hundreds of grams an acceptable compromise for the safety of everybody on the road.

Also :
The things which *do* phone home are usually thing in the category of car alarms/anti-theft features, and black-boxes/other insurance shit.
That I would be much more likely to call "useless".
But even if find the justification of presence is dubious, they don't count much in the total weight of a vehicle.

Comment More acceleration (Score 1) 482

You can't push as far as you can on the accelerator because the wheels will spin with regular street tires.

A. I was under the impression that "flooring the gas pedal" and similar expression were a figure of speech for what I'm talking about (accelerating as fast as possible), not literally as in making sure that the pedal is level with the floor of the car. But maybe, I'm not expressing myself correctly (I am not a ntaive speaker, as you might have deduced already)

B. Actually with a modern car : Yes *YOU CAN*. Try it next time when entering the highway.
It's been ages since the gas pedal isn't actually directly actuating the valve controlling the gaz admission anymore (or actuating anything else directly, for that matters).
On modern cars, the gaz pedal is just an input that give the car information about how fast you want to go. The car reads it to get your intent, and then the on-board electronics try whatever they can to ful-fill it, within the set of parameters they have to comply to, including anti-skidding.
So, when you litteraly push on the accelerator, you're not directly causing the tires to spin to fast and start skidding.
You're just signaling the car that you want to go as fast as possible, and the electronics do the necessary step to help you achieve the above mentioned 0mph to 60mph (~90km/h) in the modern much shorter time.
The engine will start revving up, but only within what the car can reasonably achieve. The "There is an art to it" part is handled by the car it self.

By the way, it's also closer to the behavior of modern cars when left on adaptive cruise control : when the way is clear they tend to accelerate rather fast, in order to achieve quickly their coasting speed in shorter period of time, spending less time in the least-efficiency "accelerating" mode, and more time in the efficient mode. (If your car's adaptive cruise control is designed to also work in city - aka "City Safe" by some constructors like Volvo - try using it and observe the car's acceleration behaviour. Also pay attention at the drop of you "average consumption per 100km" - or rise of "mileage per gallon" depending on how the computer counts it) (To go back to the beginning of this discussion: modern cars will self balance on the scale of efficiency vs. power. Modern car aren't build for either one or the other as the top post implied, they try to compromise given the specific needs at a time).

Which brings us to...

And yes the lights are usually very close together.

But unless the lights are *literally* 20 meters appart, over all you won't spend a high percentage of the time accelerating.
You'll be spending most of the time at the normal street driving speed, and only accelerate after the red turn green, until you reach the before mentioned speed.
So indeed modern cars aren't all "power to the detriment of efficiency".

Which brings us back to TFA : it shows number that, yes, overall, even if power is going up, the end result is that efficiency is getting better too, no matter what the top power of cars has become.
Efficiency hasn't been neglected at the alter of power.

As far as I understand, the laws of physics dictate that many fast accelerations wear down the battery than casual driving. Correct me if I am wrong, but the laws of inertia also apply to a Tesla do they not?

Yes law of physics apply. To the car as a whole.
If you want to accerelate a car, you need to give an energy delta of kinetic energy (Ekin = 0.5 m * v^2).
The thing is, in reality, you'll never ever going to get away *with that little energy* in real life.
There's tons of other things going on to consider where laws of physics will apply too, and will cause losses, meaning at the end, you'll need to spend a lot more joules than just the difference in Ekin.

On an electric motor, and on a pure ICE these extra losses are different.

(e.g.: an ICE will usually have a best efficiency near a very narrow RPM. You need to constantly shift gears (or have the automatic transmission do it for you) to keep the motor spinning as close as possible to this.
An electric AC motor can spin any where between 0 rpm and as high as you want it (providing the driving electronics support it), and still have the optimal power efficiency. Electric motor only use a fixed gear ratio.
The specific physics of ICE and electric motor are different, and it does have an impact on the *extra* energy that needs to be spent in addition to the inertia affecting the total mass of the vehicle
Please google for additional information. - Cue in Dr Bones "I'm a doctor, not a car mechanics")

(also, nearly any modern electric drive performs "regenerative braking". Meaning unless you need to stomp on the brakes to perform an emergency stops (or the adaptive cruise control does if for you...), when slowing down and stopping the car won't actually uses its brakes but uses the electric motor as a glorified dynamo the recharge the battery. Progressively stopping not only spares your brakes as in a pure-ICE, it also gives you back extra mileage).

But you're right, my above statement isn't clear enough. I should have said :
"Electric motors are a little bit *less catastrophically bad* at efficiency during acceleration/stopping,
so you won't be emptying the battery at an *as much alarming rate* as an ICE's gas tank under the same condition".
i.e.: both drives are inefficient, but electric drive is a bit closer to the ideal efficiency than ICE".

Practical experience :
- a somewhat powerful pure ICE sedan car will get you anything between 500 and 800 km of range on highway, depending on how aggressively you drive.
in city, due to constant accelerations and stopping, this will dramatically drop to a 450 range.

- a old-generation Zoe, with its puny little 22kWh battery will get you anywhere between 100km to 150km of range on the highway depending on how much you drive like an asshole vs. conservatively.
in city, ... it will also get you anywhere between 100km and 150km of range, depending on how much aggressively you drive.
(Which is probably the reason why it is rated at 120km range officially).
it's less affected by constant stopping accelerations than ICE (though it seems to me in my experience that it's a bit more affected by top speed)

If it is a gas vehicle then I drive into the one of twenty gas stations along my cruise route and I'm on my way in ten minutes. If it is an EV, I don't think there are any charging stations along my route. So I have to spend 2 hours driving home and charging, and by that time everything is over.

(You were complaining about getting less mileage out of an electric car, not about the availability of charging stations in your neighborhood).

Let's see the numbers :
- old renault zoe with puny 22kWh battery gets you ~120km range, which is more than enough here around in Europe (with all its dense population). For getting around city, or driving between close cities.
- modern version has a 45kWh battery, getting you ~250km, which is more than enough in most local scenarios in europe, unless you're getting on a long distance road trip.
But okay, that's us european with our strange urban organisation (several dense cities not far from each other) (which also make public transportation like train possible, and actually very popular and successful in northern/central europe. See Germany, Switzerland, etc.)

- on your side of the Atlantic pond, your Teslas have batteries with anywhere between 65kWh and 100kWh bringing you range somewhere between 300 and 500km.
That's the point at which the car can drive for a much longer distance than you can actually drive before getting dangerously tired/inattentive and needing a rest. After 2 hours driving, you should definitely rest, and the 30minutes it take at the supercharging station to put some significant charge back into the car won't be that problematic.
So unless you're driving as a rotating team (1 driver, 1 watcher/copilot, 1 sleeping on the back) and trying to put as many km in a single night, you won't even be affected by the mileage of a Tesla.

Also the whole logic behind battery-powered car is a bit different.
ICE is about opportunistic refilling, whenever the gas tank get low. It relies on a network of *gaz stations on the way* for daily operations. You can start driving, notive the red light signaling you only have 20km of range left in the gaz tank, and decide to pull to a gaz station.

A battery-powered also relies on the car being plugged in your home, and eventually at work. So whenever you start driving, your car is full (or nearly full if your workplace's parking lot are cheap scoorges who only provide domestic low-amp sockets and no Mennekes) and you'll always have close to full range.
You normally never rely on stations when you drive around.
It's only when you start making longer road trips that you'll need to pull to charging stations (e.g.: Super chargers, for Tesla. Charging socket for EV in public parkings. Even highway gaz stations have started to add charging stations).
At that point, you're planning a trip anyway. Planning where to recharge isn't much an added effort.

So EV-charging station is only a concern for long distance road trips. Not for typical in-city activity, or between city commute.

But year, probably we're a bit spoiled here in Europe with some significant efforts put toward providing charging opportunites.
(Though the situation varies with countries. You're better driving electric raod-trips through germany or switzerland, rather than italy or spain).

Comment Evolution (Score 1) 321

Phage therapy {...} bacterias will also evolve resistance against it.

Luckily for us, bacteriophages themselves - as a type of primitive parasitic semi-life form, similar to viruses - are also subject of evolution and are also under evolutive pressure to adapt to keep having access to hosts in order to be able to replicate.
They also replicate on a much higher speed than bacteria, meaning that they are much faster affected by evolution.
(i.e.: it is realistically possible to imagine keeping culture of semi-resistant bacteria and trying to grow phages on them and get those to evolve.
You might end up with mutated phages who'll be better at attacking these specific bacteria).

This is unlike the pharmacochemical industry which needs to come up with new formula of their own.

This is also less like the various yeasts which through evolution have came up with solutions against their resources competitors, and have historically been where antibiotics have been discovered. (Yeast does evolve, but at a much slower pace. A culture of yeast and semi-resistant bacteria is most likely to end up with the yeast starving).

Comment Money investement (Score 1) 321

If you invest $1B+ bringing a new antibiotic to market, but only manage to make $100M before it goes off patent (or $0 if it fails in phase IV clinical trials), you can't sustain that.

It all depends on who's doing the investement.

- Public sector - i.e.: countries investing in their universities - do *NOT* need to make a return on investment (i.e.: sell a profitable drug), they only want to make the science progress.
The problem, is that the budget necessary is beyond the financial means they can invest into a project.
i.e.: your government would gladly pay you to try new anti-biotics but doesn't have the money to it.

- So, nowadays, these investment are handled by pharmaceutical companies to whom this is really within budget - not to say small change (these sums looks probably like rounding errors next to their marketing budgets). But, as companies, they have to think about profitability.
i.e.: they would have the money but don't want to throw it away on something that will never make profit.

- Thankfully we start to see whole continent-level academic collaboration (e.g.: universities accross whole europe), and now financing projects all the way to clinical use is within the financial reach of publicly funded projects.

Comment Specificity (Score 1) 321

Not all bacteria are harmful. This phage therapy needs to not wipe out the symbiotes that make it possible for us to stay alive.

Actually it's going the other way around :

- Antibiotics are rather indiscriminate and can kill large swaths of bacterial population, including commensal flora (= "the bacteria which normally live here", i.e.: non-dangerous).
That's one of the reasons (the non-ecological/resistance one) why doctors try to avoid over-prescribing. (Just ask any girl who got yeast infection - e.g.: candida - because her flora got disturbed by a wide-spectum antibiotics)
That's also a reason why antibiotics can be prescribed with micro-flora supplements (the antibiotics will kill the commensal flora in addition to the bacteria causing the disease you're trying to cure, so you need to import new microorganisms to compensate - usually Saccharomyces, a type of benign yeast)
(Disclaimer: IAAD)

- Phage are the bacteria equivalent of viruses. They target *specific* surface receptors. It's like viruses and eukaryote (you might catch flu from a swine because surface cell receptors are close enough for a virus targeting one to be able to bind the other - ie. we're closely enough related. You'll probably never catch a virus usually affecting plants.)
A phage might be able to recognize and bind a few related bacteria, but will never affect other completely different prokaryotes).

Comment Acceleration (Score 1) 482

Lol when I'm driving my 500 hp vehicle I st{e}p on the gas at every light.

Yeah, once the traffic light turns back to green, you push as far as possible on the accelerator to have the car jump forward as fast as possible.
But once you've reach the normal driving speed (which on modern car you'll reach in only half the time than before, according to the summary),
what do you do ? Do you keep pushing the pedal all the way up to 250 km/h and/or until you collide something ?

Nope, at some speed, you'll coast, and at that point the emission of the car will drop dramatically.
The car will automatically go for a different compromise point on the efficiency vs. power scale.
(e.g.: it might even shut some cylinders down).

So, unless the road you drive is not only empty, but consist entirely of red lights spaced each only 20m appart (with red-light cameras to force you to stop and re-accelerate at each), you'll spend most time coasting, and only sporadically accelerating.
So even if the acceleration are optimised for power and don't emit less, the rest of the time the car will be in efficient mode, and over the whole trip, they tend to emit less.

If I had a Tesla Model S I'd probably get frustrated that the battery couldn't even take me on a single night's cruise.

Actually, electric motors are much more power-efficient at accelerating.
You won't be killing your battery as fast as you would be emptying the gas tank.

If you need to constantly stop and accelerate, electric motors are actually much better. That is the reason why electrical vehicle have been used by public services for quite some time. (e.g.: in France, Citroen have been making electric vans for the post office delivery service for nearly 4 decade. This thing even *predates lithium batteries*, the first vans used nickel) (e.g.: unlike in the US with Tesla which concentrate on passenger cars, in Europe most electrical vehicle company have also been producing electrical utility vehicle for quite some time. See also Renault's Z.E. Kangoo)

(This is also where part of the better mileage of hybrid vehicle comes from: if they accelerate on electric, they are more efficient).

Comment other therapies (Score 5, Insightful) 321

The worst part is, if a new antibiotics is discovered, it might help you right now, but after a couple of year, because of over use(*), the bacteria will eventually evolve some resistance against it. So the next patient with the same kind of infection will be again in the same situation...

Maybe time to dust off alternative therapies, like phage therapy ? (**)
Cue in citation of your favorite strategist (Churchill, Sun-Tzu, Machiavelli, etc.) commenting about the millennia-old proverb that the enemy of your enemy is your friend.


(*) : over-prescription, industrial/agricultural use, etc.

(**) : phage are like viruses but specialize in infecting bacteria. So phage therapy is basically curing your sickness, by making your sickness itself sick, with its own sickness, in a kind of pathogen-ception.

Slashdot Top Deals

If Machiavelli were a hacker, he'd have worked for the CSSG. -- Phil Lapsley