They generally are, look up 'attractive nuisance' laws in your jurisdiction. Some specifically ban signs that are not externally-illuminated (making electronic signs illegal.)
"Kinda hard to thoroughly run millions of units through a QA department."
No, not really, even with an actual staff of QA engineers vs automated tools.
Source: I did the QA on every LED light I've sold, Hydroponics system I've built, and very building I've constructed, plus I continuously do QA on every mine I currently own as I continue to prospect further into mountains. People that know what they're doing will find the flaws very early on.
Only because people fail at understanding basic power systems and basic fucking math. I have no problems using a wall wart with exposed wires to charge individual Li-ion cells.
" Gratuitously attributing failure to the battery when the fires could be attributed to other system components"
Nope, most other system components can't get that hot if you paid attention to Ohm's Law and did the math yourself.
"No. Because li-ion battery design is supposed to prevent fires in the case of other system components/software being faulty"
Humans that think they can beat physics suffer the consequences of physics. In this case, fire is the consequence.
"This is because that battery chemistry is inherently fire-prone"
Do the math on Nickel-Zinc and try saying that again with a straight face, assuming you even know how the physics and math for specific types of power storage systems works. Norman castle doesn't apply, here.
" Oh and before you claim it won't work for you my day job is running a HPC system with thousands of cores and we have users doing exactly what you are doing so try again."
Sounds like your programmers need an education update - we've done similar for ten years just simulating light projections from LEDs.
Come back when you've actually got a grasp of the basics and can do it on 386-class hardware, you n00b.
I built that emulation environment like three years ago. It can run on the cloud and then some.
Sadly, my CFO was stupid enough to convince the CEO to sell this tech to another company. Well, what can you expect when your CFO faces jail time and needs money?
That's HP laptops. (I worked for them at Solectron Global) specifically the DV9000 models. No other model in that year range from any other maker had that problem. (I was a former HP lead repair tech on business and consumer laptops.)
"Somehow however there is not enough market for some entrepeneurial company to produce either of those at a price competitive with current 'budget' x86 offerings (under 100-300 dollars for a mobo+cpu) "
You fail. Pricewatch.com - even Newegg and TigerDirect advertise there and get their prices routinely beaten down.
"Focussing on vastly improving the performance of a single core in order to offset the bottlenecking, while other cores work more like auxiliary processors? Is something like that feasible?"
Quit using languages like Python and Java which have very little optimization for multi-core/multi-thread performance..
"During the Pentium 4 era they were expecting P4's design to scale upwards of 10Ghz and predicting stuff like 30Ghz processors, we are nowhere near that 30Ghz though 10 years later and that was in 2006"
We can do 30 GHz easily on analog-based chips (think VGA-era electronics,) the problem lies in keeping things like capacitive and inductive forces down so that such rapid switching speeds are possible with low chance of bit-level errors and bad branching being introduced in 'digital' (aka discrete analog signal) ICs.
"The P4's 'bad' design was because it was rushed into mass production on an existing and nearly obsolete die scale"
Wrong, the long instruction pipeline is what caused most of the energy and performance losses, proven on a smaller die scale later on with the Northwood 64-bit P4 procs (and 'fixed' with microcode and cache updates to bring performance levels around the sub-GHz Tualiatin PIII.) Do you ever bother studying the underlying architecture?
"The reality is CPU stagnation because AMD pulled a Pentium-4, and Intel decided to get in touch with its Green side, instead of pushing further in the Hz-race"
Spoken like a true n00b that knows nothing about how CPUs are optimized. The Hertz wars were over in the early 2000s once multiple instruction multiple data became the norm. From there, it became cache+core wars. Now it's energy efficiency (which leads to faster clock speeds and smaller die sizes when everything is designed right.)
"The reality is CPU stagnation because CPU's hit a frequency wall and multi-core isn't good enough for the future of computing"
No, the reality is that people aren't taught how to perform parallelization of algorithms that can be split and run more efficiently in smaller timed chunks.
"basically we need a technology that re-enables single threaded performance"
That's not happening with newer games, scientific simulations, and etc. demanding more power and more cores.
"That's probably a good 50 years to a century away though."
It's likely to never happen without a serious fundamental change in computing ideals and a fundamental shift in the understanding of facts regarding multi-core parallel processing.
Well, yea. After all, nobody attacked us badly enough to make us want to fight until Pearl Harbor, then we just popped in and ended it in a relative flash. We may show up late, but we can pretty much guarantee a win no matter what.
"And as I've mentioned, we've all been quite content to demean government, drop civics and in general conspire to produce an unaware and compliant citizenry."
If you were an EE you'd know thermal cycling is an issue for any component from resistors to simple low-power LEDs. NOTHING IS IMMUNE TO THE LAWS OF PHYSICS.
"If the reports about the Yahoo order are accurate -- including requiring the company to custom build new software to accomplish the scanning"
I do believe there is a law out there that forbids the federal government from requiring companies to build things a specific way or design specific things specifically for the US Gov't. Slashdot has discussed this many times in the comments regarding the iPhone. The Gov't has no legal power to compel Yahoo or any other entity to do their bidding in this manner and some/.er has pointed this law out several times.
Uh, yea, they usually do. I can tell you've never released the magic smoke from a capacitor or resistor, or blown up a transformer, let alone done any EE design.
"Hell no. We are not talking about a 100W CPU/GPU, we are talking about a touch controller IC that uses almost no power. Thermal cycling due to regular use is not an issue"
Thermal cycling is an issue for ANY electronic component; even quarter-watt LEDs get hot enough to melt themselves if not given proper thermal dissipation.
Slashdot Mirror
They generally are, look up 'attractive nuisance' laws in your jurisdiction. Some specifically ban signs that are not externally-illuminated (making electronic signs illegal.)
"Kinda hard to thoroughly run millions of units through a QA department."
No, not really, even with an actual staff of QA engineers vs automated tools.
Source: I did the QA on every LED light I've sold, Hydroponics system I've built, and very building I've constructed, plus I continuously do QA on every mine I currently own as I continue to prospect further into mountains. People that know what they're doing will find the flaws very early on.
No, because frozen things tend to be less explodey, thus this is smart.
"Chargers are complicated"
Only because people fail at understanding basic power systems and basic fucking math. I have no problems using a wall wart with exposed wires to charge individual Li-ion cells.
" Gratuitously attributing failure to the battery when the fires could be attributed to other system components"
Nope, most other system components can't get that hot if you paid attention to Ohm's Law and did the math yourself.
"No. Because li-ion battery design is supposed to prevent fires in the case of other system components/software being faulty"
Humans that think they can beat physics suffer the consequences of physics. In this case, fire is the consequence.
"This is because that battery chemistry is inherently fire-prone"
Do the math on Nickel-Zinc and try saying that again with a straight face, assuming you even know how the physics and math for specific types of power storage systems works. Norman castle doesn't apply, here.
No, it doesn't, until you move to Windows server 2012 (aka Windows 8 Server.)
Before that, RemoteFX had some serious performance. Then came 2012, which cut that performance literally in half for 'security measures.'
For roughly 3X the cost I could get almost 30X the performance.
Try again when you design systems for specific purposes.
(BTW, note this system has a huge memory bottleneck architecturally [if you bothered to read the spec sheet and understand it.])
He's not an engineer - his explanation would be about 4x more verbose.
" Oh and before you claim it won't work for you my day job is running a HPC system with thousands of cores and we have users doing exactly what you are doing so try again."
Sounds like your programmers need an education update - we've done similar for ten years just simulating light projections from LEDs.
Come back when you've actually got a grasp of the basics and can do it on 386-class hardware, you n00b.
I built that emulation environment like three years ago. It can run on the cloud and then some.
Sadly, my CFO was stupid enough to convince the CEO to sell this tech to another company. Well, what can you expect when your CFO faces jail time and needs money?
" In any case, either is needed to address 512 GB."
Funny, I can address that much in Windows Sever 2K3 64-bit mode plus I can do it in Server 2K3 32-bit with PAE. It still runs as my primary gaming OS.
Do you not know how to efficiently utilize your system resources?
That's HP laptops. (I worked for them at Solectron Global) specifically the DV9000 models. No other model in that year range from any other maker had that problem. (I was a former HP lead repair tech on business and consumer laptops.)
"Yes, multi-threaded is overrated"
Not when you're running multiple virtual worlds that need coherence between instances, like I'm doing with my 2D Second Life alternative project.
"Hyperthreading is a joke."
Only because you're incompetent at making such things work.
"Extra memory doesn't pay off: It makes programmers lazier"
Yet my 2D SL-alternative runs in under 4 megs of RAM minus the loading for music and sprites.
"Windows 10 = A Bore"
Thank goodness my stuff runs on 2000/XP without fail.
"What you've got is good enough: Browse the web. Watch Netflix. Ummm... type? Unless you're a hardcore gamer, PCs that are years old are good enough."
You obviously don't make programs.
"Somehow however there is not enough market for some entrepeneurial company to produce either of those at a price competitive with current 'budget' x86 offerings (under 100-300 dollars for a mobo+cpu) "
You fail. Pricewatch.com - even Newegg and TigerDirect advertise there and get their prices routinely beaten down.
"Focussing on vastly improving the performance of a single core in order to offset the bottlenecking, while other cores work more like auxiliary processors? Is something like that feasible?"
Quit using languages like Python and Java which have very little optimization for multi-core/multi-thread performance..
"During the Pentium 4 era they were expecting P4's design to scale upwards of 10Ghz and predicting stuff like 30Ghz processors, we are nowhere near that 30Ghz though 10 years later and that was in 2006"
We can do 30 GHz easily on analog-based chips (think VGA-era electronics,) the problem lies in keeping things like capacitive and inductive forces down so that such rapid switching speeds are possible with low chance of bit-level errors and bad branching being introduced in 'digital' (aka discrete analog signal) ICs.
"The P4's 'bad' design was because it was rushed into mass production on an existing and nearly obsolete die scale"
Wrong, the long instruction pipeline is what caused most of the energy and performance losses, proven on a smaller die scale later on with the Northwood 64-bit P4 procs (and 'fixed' with microcode and cache updates to bring performance levels around the sub-GHz Tualiatin PIII.) Do you ever bother studying the underlying architecture?
"The reality is CPU stagnation because AMD pulled a Pentium-4, and Intel decided to get in touch with its Green side, instead of pushing further in the Hz-race"
Spoken like a true n00b that knows nothing about how CPUs are optimized. The Hertz wars were over in the early 2000s once multiple instruction multiple data became the norm. From there, it became cache+core wars. Now it's energy efficiency (which leads to faster clock speeds and smaller die sizes when everything is designed right.)
"The reality is CPU stagnation because CPU's hit a frequency wall and multi-core isn't good enough for the future of computing"
No, the reality is that people aren't taught how to perform parallelization of algorithms that can be split and run more efficiently in smaller timed chunks.
"basically we need a technology that re-enables single threaded performance"
That's not happening with newer games, scientific simulations, and etc. demanding more power and more cores.
"That's probably a good 50 years to a century away though."
It's likely to never happen without a serious fundamental change in computing ideals and a fundamental shift in the understanding of facts regarding multi-core parallel processing.
Well, yea. After all, nobody attacked us badly enough to make us want to fight until Pearl Harbor, then we just popped in and ended it in a relative flash. We may show up late, but we can pretty much guarantee a win no matter what.
https://wikileaks.org/podesta-...
Oh fucking really?
"And as I've mentioned, we've all been quite content to demean government, drop civics and in general conspire to produce an unaware and compliant citizenry."
Go sit down you ill-educated twit.
If you were an EE you'd know thermal cycling is an issue for any component from resistors to simple low-power LEDs. NOTHING IS IMMUNE TO THE LAWS OF PHYSICS.
"If the reports about the Yahoo order are accurate -- including requiring the company to custom build new software to accomplish the scanning"
I do believe there is a law out there that forbids the federal government from requiring companies to build things a specific way or design specific things specifically for the US Gov't. Slashdot has discussed this many times in the comments regarding the iPhone. The Gov't has no legal power to compel Yahoo or any other entity to do their bidding in this manner and some /.er has pointed this law out several times.
"Erm no, electronics don't magically burn up."
Uh, yea, they usually do. I can tell you've never released the magic smoke from a capacitor or resistor, or blown up a transformer, let alone done any EE design.
"Hell no. We are not talking about a 100W CPU/GPU, we are talking about a touch controller IC that uses almost no power. Thermal cycling due to regular use is not an issue"
Thermal cycling is an issue for ANY electronic component; even quarter-watt LEDs get hot enough to melt themselves if not given proper thermal dissipation.