'Naturalness', whatever that means too. We've been doing selective breeding on all of our foods for millennia, so I don't consider organic foods to be particularly natural either. Traditional, I'll give them that, but natural? You have to be kidding me.
Don't need 3D performance. Don't need GPGPU performance in 99% of cases.
Doesn't matter, because it's cheap. Also, CAD and Photoshop *do* use GPGPU these days.
Media center PCs
Plenty fast enough already to play video at 1920x1080.
This should handle 4k video decoding.
low-end Steambox
If you want your games to look like crap.
I think you missed the "low end" part of that quote. Also, it will be really, really cheap compared to something with an additional dGPU. You don't even need PCIe on the motherboard. Not everybody can afford to game at 3x 1080p on high. These should handle 1080p on medium just fine.
Integrating the GPU into the CPU gets the BOM cost down and raises the minimum performance standard.
Because lots of people run 3D games on servers.
Certainly we do use GPUs for some floating-point intensive tasks on servers, but this is nowhere near fast enough to be useful.
These have HUMA. GPGPU-CPU interactions will be much faster than on any previous architecture because not only do they share memory space, they are also cache coherent at a hardware level. It suddenly makes having a whole bunch of FPUs on the graphics card useful for regular old FPU applications, because they can be accessed just as quickly as SSE/x87 FPUs. It makes OpenCL suddenly useful for very small kernels, instead of only being useful for massive data-processing chunks where the parallelisation had to be wide and simple enough to make up for memory copying overhead. TL;DR: I want this on my server, even if just for the stuff like generating graphics and accelerating database hashing. Never mind Folding@home and HPC kind of work.
It is Jevon's Paradox, because if gasoline was too expensive for driving cars they would be riding bicycles. However, if the price of gasoline came down enough they would start driving cars instead.
...maybe the car isn't actually the safest car on the planet like Elon claims
A guy went through a cement wall at 100mph, then hit a tree. The car caught fire. News at 11. Oh wait, you mean it was an *electric* car? We're pushing this forward, news at 10. Oh wait, you mean the guy in the car walked away? Damn, slow news day I guess. Hang on, let's just neglect that part. News at 10 it is.
Most gas tanks don't have 1/4" steel plate between them and the road. So far it has a 100% survival rate for passengers in fires, even the guy who went through a cement wall at 100mph. Show me a gas car with those stats.
Basically, fledgling technology isn't quite as safe in its first iteration as 100 year old technology is in it's Nth generation. News at 11.
Maybe the top of the gas tank is high, but the bottom of the gas tank is just a few inches off the road. What's more likely is, as mentioned above, the Tesla's fancy air-suspension lowers the car at cruising speed to improve efficiency, and this makes it more likely to hit something on the road.
Those are all designed to deal with small high-velocity projectiles. The Tesla battery case also needs to deal with larger 'crush'-type impacts. That Chobham armour doesn't work nearly as well as regular-old-steel-plate when somebody drives a truck over you.
Oh, and $$$. Chobham or boron-carbide faced case-hardened steel would cost more than the car.
It isn't comparable, because a town of 20k people isn't going to have the same demographics as a suburb of 20k Tesla owners. These are all incidents which were caused by something flying up from the road and puncturing a reinforced steel plate - in a regular car there would have been injuries in the passenger compartment because they don't have the shielding.
In science, the person setting the requirements is generally the person who is coding. If you set out your problem properly, you can get the entire structure and design right the first time.
As for math, in science, computers are generally used as tools to solve math problems. Scientists aren't using supercomputers for web servers and databases.
Sure there is. Not everything has such a big scope that you can't get it right the first time. Particularly if you do the math before you write the code.
If they'd joined they wouldn't have had any way to prevent the Android handset makers from being sued, it only would have covered them. The offer to join was a trojan horse and they knew it.
In science a lot of code is only used once-off to get results and prove that the idea is valid. Maintenance on something that nobody will ever look at again? Why?
Put another way, often code is the means, and not the end.
I've used FastMail - it was bought by Opera, then sold again to the FastMail employees. They have a free version with 25MB space, no spam filter and ads, or you can pay to get decent space, remove the ads and add spam filters.
Wrong. The SSL private key just allowed the Feds to snoop on any traffic that passed through. The only way the Feds could access the emails was if somebody logged in and unencrypted their emails while the snooping was happening.
Feds asked without a court-sanctioned warrant to insert a backdoor. When LavaBit didn't respond, Feds got a warrant, but this time to hand over the SSL private key. That''s when LavaBit decided that it was useless trying to fight, and quit instead.
Large? Yes. Slow? Not hardly. Think about how the two systems operate for a moment. On a direct-drive multi-rotor, rotor RPM directly controls thrust. If you want to increase thrust, you subtract the power needed to counter drag from whatever the motor is outputting, and whatever is left is available to accelerate the rotor. That means you have rapidly diminishing available power as your thrust increases. I question your claim of being able to meaningfully modulate thrust on a multi-rotor at 200Hz. The only way I could see that possible is if your motors were vastly over-speced for your requirements; a trait increasingly difficult to achieve as you scale up.
On your regular heli, if you have to run at max power and max pitch to hover, you will run into issues. Yes, motors on a quad are over-speced for hover. Just like a regular heli. They don't draw their full rated power the full time, but they are capable of producing bursts much higher. Typically a multi-rotor hovers around half-throttle. Newer firmwares also provide regeneration, so you don't lose energy by slowing the prop, chances are the energy goes into speeding up a different prop, and any excess to the battery.
On a traditional helicopter, the rotor only operates at a single RPM, and you adjust your engine power to maintain that RPM. Your thrust comes from actuation of the swashplate, so your ability to modulate thrust is dependent on the speed of those servos. That's going to be a damn sight better than anything the motors on multi-rotor are capable of. Worst case on being able to modulate proportional thrust in a particular direction is going to be one full blade passing. An RC helicopter of useful size is going to operate at speed of 2000-3000RPM. 3000RPM with a four-blade prop will get you your 200Hz modulation.
1) Nobody uses a 4-bladed prop on model helis for weight lifting. The only reason more-than-2 bladed props are used is for a) scale effect or b) on a real life plane/heli to prevent the blade tips from exceeding mach 1. More blades = less efficient. 2) Worst case is actually worse than that, because of flybars. They put much less strain on the servos, but require an extra half-revolution of the blades for any effect to take place. 3) Because nobody uses more than 2 blades for anything except scale, you have the sinusoidal acceleration for stabilisation. 4) Traditional helis have more blade area, which means more wind disturbance. 5) If you really need the manoeuvrability, and you can handle the increased complexity and chance of failure, variable pitch props can be used for multi-rotor. If anything, they give more manoeuvrability than a standard heli. And you still have the option of high motor counts to get redundancy.
Slashdot Mirror
'Naturalness', whatever that means too. We've been doing selective breeding on all of our foods for millennia, so I don't consider organic foods to be particularly natural either. Traditional, I'll give them that, but natural? You have to be kidding me.
Or if there is more than everybody could possibly consume, and nobody needed to produce it.
This is why I like robots.
Small form factor business PCs,
Don't need 3D performance. Don't need GPGPU performance in 99% of cases.
Doesn't matter, because it's cheap. Also, CAD and Photoshop *do* use GPGPU these days.
Media center PCs
Plenty fast enough already to play video at 1920x1080.
This should handle 4k video decoding.
low-end Steambox
If you want your games to look like crap.
I think you missed the "low end" part of that quote. Also, it will be really, really cheap compared to something with an additional dGPU. You don't even need PCIe on the motherboard. Not everybody can afford to game at 3x 1080p on high. These should handle 1080p on medium just fine.
Integrating the GPU into the CPU gets the BOM cost down and raises the minimum performance standard.
Because lots of people run 3D games on servers.
Certainly we do use GPUs for some floating-point intensive tasks on servers, but this is nowhere near fast enough to be useful.
These have HUMA. GPGPU-CPU interactions will be much faster than on any previous architecture because not only do they share memory space, they are also cache coherent at a hardware level. It suddenly makes having a whole bunch of FPUs on the graphics card useful for regular old FPU applications, because they can be accessed just as quickly as SSE/x87 FPUs. It makes OpenCL suddenly useful for very small kernels, instead of only being useful for massive data-processing chunks where the parallelisation had to be wide and simple enough to make up for memory copying overhead. TL;DR: I want this on my server, even if just for the stuff like generating graphics and accelerating database hashing. Never mind Folding@home and HPC kind of work.
Seriously, stop being such a downer.
To clarify, if the fuel efficiency of cars increased enough to make driving economical, they would start driving cars instead of riding bicycles.
It is Jevon's Paradox, because if gasoline was too expensive for driving cars they would be riding bicycles. However, if the price of gasoline came down enough they would start driving cars instead.
You probably should have accounted for that in your original snarky comment then. /snark
Which is exactly what the stats say. Tesla's are less than 1/4 as likely to have a fire as gasoline cars (a lot of which are old beaters).
For such an overvalued stock this could be a real thing. Even better is if you do it to somebody else's Tesla. Either that or there's insurance money.
...maybe the car isn't actually the safest car on the planet like Elon claims
A guy went through a cement wall at 100mph, then hit a tree. The car caught fire. News at 11. Oh wait, you mean it was an *electric* car? We're pushing this forward, news at 10. Oh wait, you mean the guy in the car walked away? Damn, slow news day I guess. Hang on, let's just neglect that part. News at 10 it is.
Most gas tanks don't have 1/4" steel plate between them and the road. So far it has a 100% survival rate for passengers in fires, even the guy who went through a cement wall at 100mph. Show me a gas car with those stats.
Basically, fledgling technology isn't quite as safe in its first iteration as 100 year old technology is in it's Nth generation. News at 11.
Your "town of 20k" don't all drive high-performance saloons. Most of them drive old beaters.
Maybe the top of the gas tank is high, but the bottom of the gas tank is just a few inches off the road. What's more likely is, as mentioned above, the Tesla's fancy air-suspension lowers the car at cruising speed to improve efficiency, and this makes it more likely to hit something on the road.
Those are all designed to deal with small high-velocity projectiles. The Tesla battery case also needs to deal with larger 'crush'-type impacts. That Chobham armour doesn't work nearly as well as regular-old-steel-plate when somebody drives a truck over you.
Oh, and $$$. Chobham or boron-carbide faced case-hardened steel would cost more than the car.
And they were all fires after a reinforced steel plate was punctured in an accident. Big whoop.
It isn't comparable, because a town of 20k people isn't going to have the same demographics as a suburb of 20k Tesla owners. These are all incidents which were caused by something flying up from the road and puncturing a reinforced steel plate - in a regular car there would have been injuries in the passenger compartment because they don't have the shielding.
How fast can your M5 go in 1st gear though?
In science, the person setting the requirements is generally the person who is coding. If you set out your problem properly, you can get the entire structure and design right the first time.
As for math, in science, computers are generally used as tools to solve math problems. Scientists aren't using supercomputers for web servers and databases.
Sure there is. Not everything has such a big scope that you can't get it right the first time. Particularly if you do the math before you write the code.
Perhaps they tried to buy them. Or perhaps they just took a calculated risk bidding up on them so that the others would have to pay more.
If they'd joined they wouldn't have had any way to prevent the Android handset makers from being sued, it only would have covered them. The offer to join was a trojan horse and they knew it.
In science a lot of code is only used once-off to get results and prove that the idea is valid. Maintenance on something that nobody will ever look at again? Why?
Put another way, often code is the means, and not the end.
I've used FastMail - it was bought by Opera, then sold again to the FastMail employees. They have a free version with 25MB space, no spam filter and ads, or you can pay to get decent space, remove the ads and add spam filters.
Wrong. The SSL private key just allowed the Feds to snoop on any traffic that passed through. The only way the Feds could access the emails was if somebody logged in and unencrypted their emails while the snooping was happening.
Feds asked without a court-sanctioned warrant to insert a backdoor. When LavaBit didn't respond, Feds got a warrant, but this time to hand over the SSL private key. That''s when LavaBit decided that it was useless trying to fight, and quit instead.
Large? Yes. Slow? Not hardly. Think about how the two systems operate for a moment. On a direct-drive multi-rotor, rotor RPM directly controls thrust. If you want to increase thrust, you subtract the power needed to counter drag from whatever the motor is outputting, and whatever is left is available to accelerate the rotor. That means you have rapidly diminishing available power as your thrust increases. I question your claim of being able to meaningfully modulate thrust on a multi-rotor at 200Hz. The only way I could see that possible is if your motors were vastly over-speced for your requirements; a trait increasingly difficult to achieve as you scale up.
On your regular heli, if you have to run at max power and max pitch to hover, you will run into issues. Yes, motors on a quad are over-speced for hover. Just like a regular heli. They don't draw their full rated power the full time, but they are capable of producing bursts much higher. Typically a multi-rotor hovers around half-throttle. Newer firmwares also provide regeneration, so you don't lose energy by slowing the prop, chances are the energy goes into speeding up a different prop, and any excess to the battery.
On a traditional helicopter, the rotor only operates at a single RPM, and you adjust your engine power to maintain that RPM. Your thrust comes from actuation of the swashplate, so your ability to modulate thrust is dependent on the speed of those servos. That's going to be a damn sight better than anything the motors on multi-rotor are capable of. Worst case on being able to modulate proportional thrust in a particular direction is going to be one full blade passing. An RC helicopter of useful size is going to operate at speed of 2000-3000RPM. 3000RPM with a four-blade prop will get you your 200Hz modulation.
1) Nobody uses a 4-bladed prop on model helis for weight lifting. The only reason more-than-2 bladed props are used is for a) scale effect or b) on a real life plane/heli to prevent the blade tips from exceeding mach 1. More blades = less efficient.
2) Worst case is actually worse than that, because of flybars. They put much less strain on the servos, but require an extra half-revolution of the blades for any effect to take place.
3) Because nobody uses more than 2 blades for anything except scale, you have the sinusoidal acceleration for stabilisation.
4) Traditional helis have more blade area, which means more wind disturbance.
5) If you really need the manoeuvrability, and you can handle the increased complexity and chance of failure, variable pitch props can be used for multi-rotor. If anything, they give more manoeuvrability than a standard heli. And you still have the option of high motor counts to get redundancy.