hUMA may slow down the GPU's raw execution speed, and contention between CPU and GPU access more tetchy, but it makes interaction between the game logic and presentation much more flexible. Doing this with DDR5 was a painful compromise, I am sure, but DDR3 would have made these systems cost 3 times more than they will for the same memory load-out.
TL:DR: hUMA gives developers a much more flexible and faster way to share resources between the GPU and CPU than the PCI pipes do.
Re:Ethernets great but let's compare to 1981/1982
on
Ethernet Turns 40
·
· Score: 1
I've been running GigE on my LAN for almost 5 years. It is almost as fast as sneaker-netting a FW800 drive over to the other computers on the LAN... with the time lost to juggling power and dealing with finicky crap external drives... screw it. Ethernet FTW!
Re:The name Ethernet is 40 years old...
on
Ethernet Turns 40
·
· Score: 1
That explains why WiFi is still at it's base standard called IEEE 802.... duh! Along with all the other variants of the 802 protocol family. Hint... it is not really about the phys-layer... it is about the data link layer... that is what really defines the 802 protocols.
I actually still have an Ethernet to Appletalk bridge that supports 10BT and 10B2 and it still works... well my local switch will show a link active, and dump packets onto it.
This world has indeed gone mad. When I was a preteen, the suburban neighborhood's post-teen hoodlums hauled a junker car on a flat-bed out to the cleared, but undeveloped edge of the housing tract. Kids were offered an opportunity to beat on it for a while. After everyone got bored of beatting on the car, they detonated a rather large black-poweder PVC pipe-bomb inside of it. As the local model rocketry expert I supplied the electric ignition system complete with igniter, batteries, and a 555 one-shot timer relay circuit to trigger it, including a proper two-stage arming-safty. (no one had long enough cables to risk a regular switch based trigger, typically used for launching model rockets, so we used the wire we had and the relay timer)
There was a junker-shattering kabooom! It echoed across the valley for a long time. Big huge smoke plume. Everyone had a good laugh and explored the damage done to the vehicle. No cops showed up. No one raised a stink. No one was hurt or scared, or gave a rats ass. Everyone who witnessed was asked to do a little KP duty to clean up the chunks blown off the vehicle. The carcass was hauled off to the junkyard. For the local kids like me this was a rather huge event. It wasn't like this happened all the time.
Go get a textbook on discrete mathematics... and then take the time to study. It takes you from your comfy base-10 world and exposes you to a lot of weird stuff that only makes sense when dealing with integers, floating point, alternate bases, and other things that computers do very well.
I think they just wanted a plausible excuse for a holiday. Take a day off when you can get it. Trrists is as good an excuse as a sick dog, or sick kid. YMMV
People who keep their laptops on AC are NOT killing their batteries by keeping the system on AC. The batteries are being killed by the fucking design of the laptop's charging circuit! Bonehead charging systems will keep trickle charging the battery even when they shouldn't! The end result is shortened battery life. In effect the laptop was designed to kill the battery prematurely.
Properly designed charging systems do not do this. If you bought a cheap-ass laptop you can expect it to chew up batteries. The only way to avoid that is if you manage the charge cycle yourself.
Funny... I have a 2001 Apple Ti-book... It is on its second battery pack(the first one was killed by keeping the system in storage for over a year without any charge). I have never taken any special precautions for the battery, other than to make sure it gets FULLY recharged ASAP after a discharge cycle. It spends many months at a time in sleep mode, and is only off AC a few times a year. The battery in it now was purchased new over 7 years ago... It still holds 4 - 5 hours of capacity under conservative loads. This is about 20% less than what it did new.
The reason this is so, is that: 1. The charging system was designed not to abuse the battery in any way. 2. The battery is mounted in a location in the chassis that is not subject to heating from other system components.
Conclusion after 12 years of use? The battery chemistry is not the problem. PEBCAK, and shitty engineering kills batteries.
I recommend that everyone read the transcript of oral arguments.
Myriad claims they deserve patents on BRCA1 and BRCA2 cDNA because they knew exactly where to 'snip' the sequence. This is their innovation.
cDNA (type specification) is little more than a transcription of a DNA sequence from the genome with the 'junk sequences' edited out. The transcription process of the a cell nucleus does this naturally, and it is well known how to extract this naturally edited(de-junked) sequence from a cell culture. After lots of effort Myriad discovered how to isolate BRCA1 and BRCA2 from the 'de-junked' cDNA sequences derived from the subjects DNA. In their relevant patents they claim patentable invention on the grounds that the cDNA sequence of these genes(technically psuedo-genes, or even sub-genes) does not exist in isolation, AND they discovered where to 'snip' the cDNA strand to isolate the sequences at issue.
This is no different than asserting that I can patent a specific region of everyone's liver because *I* 'sweat blood and tears' to figure out where to stick a biopsy needle to extract a useful sample of that region. In effect, anyone (except me) who extracts that specific region of a liver BY ANY MEANS would be liable for patent infringement. I now own a piece of EVERYONE'S liver. Anyone who attempts to take that small piece of liver it out of a host OWES ME MONEY!
Oh and BTW, I get the liver pieces too. So I'll be in the market for fava beans and some nice Chianti too.;)
sorry to reply to myself... but a slight qualification:
The GP was concerned about current flowing back into the mother board from a air-spun BLDC cooling fan. While the typical fan will not do this due to the protection circuits that keep the BLDC motor driver electronics from being fried by 'free-wheeling' DON'T COUNT ON IT. It is entirely possible for a poorly designed BLDC drive to push current back to the motherboard if the protection circuits have been damaged.
It is not an inherent feature of BLDC drivers to prevent current from flowing back to the rails when free-wheeling. This must be designed into the driver competently.
PC fans use brushless motors [wikipedia.org], which don't do that.
What a silly thing to say. DC Fan motors (BLDC motors that you linked to) most certainly do act as generators when spun from an external prime-mover. HINT: they have permanent magnets in the rotor. They also have protection circuits to prevent damage to their drive transistors from 'free-wheeling' them... even at high RPM. The only damage you could do to one is to the bearings or causing a fracture of the rotor due to excessive RPM.
What you may be referring to are AC motors. Which have no magnets, and require current flowing in both the rotor and stator to generate torque. ProTip: Alternators work this way as well. The output of the alternator is regulated by the current applied to a 'field coil' which may be either on the rotor or stator. IIRC the rotor is usually the field coil as that makes the alternator cheaper to build.
A little basic electronics math will tell you... might be a bit off on this but I think we can get close with:
For electrical transmission over wire the mass of a bit is approximately* proportional to M =(1/B)*A*V*C*E Where B is bits per second A is average impulse current flow within a bit frame. V is the peak-to-peak voltage of the signal carrying the bits. C is the Coulomb constant: 6.241509x10^18 E is the mass of one electron 9.109382x10^-31 kg
*This will overestimate the mass of a typical bit for a number of reasons. Not all the combinations of bits in a sequence require the same energy to transmit. This calculation assumes a square bit. Circular, or round(ish) bits are more accurate but the maths are uglier(in/.) to show/write.
TL:DR My conclusion is there are some rather noisy and incompetent lawyers that want this guy axed.
After reading through the litany of whinging over this judge, at http://www.therobingroom.com/Judge.aspx?ID=1555 and the few positive ratings, I am wondering if the allegations are really sour grapes from attorneys who got their bullshit handed back to them on a silver platter. This guy is clearly a hardass. However I saw no clear evidence that the negative comments were anything but inarticulate nerd-rage. The few positive comments clearly articulated why the judge deserved positive praise. When someone is howling about being oppressed it is pretty damned important to note how well argued their position is. I did not see a single negative comment that gave any clarity.
Boole, Church, Turing and others in the Pre-war era stood on the shoulders of Wilkins, Huygens, Hooke, Leibniz, Babbage, Lady Ada Lovelace, and others. All of whom helped define Discrete Mathematics, which is the underpinnings of CS, and CIS.
Your average smart phone these days has more raw grunt, RAM and storage than a high end Cray 2 installation. The Cray 2 was the fastest super computer from 1985 until 1989: 1 GB of RAM 1.9GFLOPS. I think many of the monolithic(non-massively parallel) super computers of the early 90's would only be just eking nose and nose with a high-performance smartphone/tablet. Some of the early massively parallel machines would lose to a smartphone when the GPU(s) are added to the comparison. Oh and the power consumption is just hilarious.
Cray 2 ~250KW..(typical. including cooling infrastructure) iPhone CPU 2W (Max. cpu only. passive cooling)
Heck, a 1MHz 6502 drew more than 2W until the 65c02 came out.
Slashdot Mirror
"What Andy giveth, Bill taketh away..."
I see what you did there. Clever.
erm swap DDR5 and DDR3... really... I do know the difference. >_
I'll put it another way.
hUMA may slow down the GPU's raw execution speed, and contention between CPU and GPU access more tetchy, but it makes interaction between the game logic and presentation much more flexible. Doing this with DDR5 was a painful compromise, I am sure, but DDR3 would have made these systems cost 3 times more than they will for the same memory load-out.
TL:DR: hUMA gives developers a much more flexible and faster way to share resources between the GPU and CPU than the PCI pipes do.
I've been running GigE on my LAN for almost 5 years. It is almost as fast as sneaker-netting a FW800 drive over to the other computers on the LAN... with the time lost to juggling power and dealing with finicky crap external drives... screw it. Ethernet FTW!
That explains why WiFi is still at it's base standard called IEEE 802.... duh! Along with all the other variants of the 802 protocol family. Hint... it is not really about the phys-layer... it is about the data link layer... that is what really defines the 802 protocols.
Yeah... well. The industry has called it RJ45 as long as it has been available. While you are technically correct....
It doesn't matter any more.
I actually still have an Ethernet to Appletalk bridge that supports 10BT and 10B2 and it still works... well my local switch will show a link active, and dump packets onto it.
This world has indeed gone mad. When I was a preteen, the suburban neighborhood's post-teen hoodlums hauled a junker car on a flat-bed out to the cleared, but undeveloped edge of the housing tract. Kids were offered an opportunity to beat on it for a while. After everyone got bored of beatting on the car, they detonated a rather large black-poweder PVC pipe-bomb inside of it. As the local model rocketry expert I supplied the electric ignition system complete with igniter, batteries, and a 555 one-shot timer relay circuit to trigger it, including a proper two-stage arming-safty. (no one had long enough cables to risk a regular switch based trigger, typically used for launching model rockets, so we used the wire we had and the relay timer)
There was a junker-shattering kabooom! It echoed across the valley for a long time. Big huge smoke plume. Everyone had a good laugh and explored the damage done to the vehicle. No cops showed up. No one raised a stink. No one was hurt or scared, or gave a rats ass. Everyone who witnessed was asked to do a little KP duty to clean up the chunks blown off the vehicle. The carcass was hauled off to the junkyard. For the local kids like me this was a rather huge event. It wasn't like this happened all the time.
Go get a textbook on discrete mathematics... and then take the time to study. It takes you from your comfy base-10 world and exposes you to a lot of weird stuff that only makes sense when dealing with integers, floating point, alternate bases, and other things that computers do very well.
I think they just wanted a plausible excuse for a holiday.
Take a day off when you can get it.
Trrists is as good an excuse as a sick dog, or sick kid. YMMV
People who keep their laptops on AC are NOT killing their batteries by keeping the system on AC. The batteries are being killed by the fucking design of the laptop's charging circuit! Bonehead charging systems will keep trickle charging the battery even when they shouldn't! The end result is shortened battery life. In effect the laptop was designed to kill the battery prematurely.
Properly designed charging systems do not do this. If you bought a cheap-ass laptop you can expect it to chew up batteries. The only way to avoid that is if you manage the charge cycle yourself.
Funny... I have a 2001 Apple Ti-book... It is on its second battery pack(the first one was killed by keeping the system in storage for over a year without any charge). I have never taken any special precautions for the battery, other than to make sure it gets FULLY recharged ASAP after a discharge cycle. It spends many months at a time in sleep mode, and is only off AC a few times a year. The battery in it now was purchased new over 7 years ago... It still holds 4 - 5 hours of capacity under conservative loads. This is about 20% less than what it did new.
The reason this is so, is that:
1. The charging system was designed not to abuse the battery in any way.
2. The battery is mounted in a location in the chassis that is not subject to heating from other system components.
Conclusion after 12 years of use? The battery chemistry is not the problem. PEBCAK, and shitty engineering kills batteries.
I recommend that everyone read the transcript of oral arguments.
Myriad claims they deserve patents on BRCA1 and BRCA2 cDNA because they knew exactly where to 'snip' the sequence. This is their innovation.
cDNA (type specification) is little more than a transcription of a DNA sequence from the genome with the 'junk sequences' edited out. The transcription process of the a cell nucleus does this naturally, and it is well known how to extract this naturally edited(de-junked) sequence from a cell culture. After lots of effort Myriad discovered how to isolate BRCA1 and BRCA2 from the 'de-junked' cDNA sequences derived from the subjects DNA. In their relevant patents they claim patentable invention on the grounds that the cDNA sequence of these genes(technically psuedo-genes, or even sub-genes) does not exist in isolation, AND they discovered where to 'snip' the cDNA strand to isolate the sequences at issue.
This is no different than asserting that I can patent a specific region of everyone's liver because *I* 'sweat blood and tears' to figure out where to stick a biopsy needle to extract a useful sample of that region. In effect, anyone (except me) who extracts that specific region of a liver BY ANY MEANS would be liable for patent infringement. I now own a piece of EVERYONE'S liver. Anyone who attempts to take that small piece of liver it out of a host OWES ME MONEY!
Oh and BTW, I get the liver pieces too. So I'll be in the market for fava beans and some nice Chianti too. ;)
Opportunity?
Curiosity!
Stop.
Stop feeding APK.
Ignore APK; it will go away.
replying to the blasts only serves to feed its dependence.
Apply discipline and resist the urge to respond...
end-of-line
Are you sure that isn't a malapropism crossed with a portmantaeu?
Dare I say it, the dreaded malamantaeu?
I was kinda thinking it is a propmanteauism.
sorry to reply to myself... but a slight qualification:
The GP was concerned about current flowing back into the mother board from a air-spun BLDC cooling fan. While the typical fan will not do this due to the protection circuits that keep the BLDC motor driver electronics from being fried by 'free-wheeling' DON'T COUNT ON IT. It is entirely possible for a poorly designed BLDC drive to push current back to the motherboard if the protection circuits have been damaged.
It is not an inherent feature of BLDC drivers to prevent current from flowing back to the rails when free-wheeling. This must be designed into the driver competently.
PC fans use brushless motors [wikipedia.org], which don't do that.
What a silly thing to say. DC Fan motors (BLDC motors that you linked to) most certainly do act as generators when spun from an external prime-mover. HINT: they have permanent magnets in the rotor. They also have protection circuits to prevent damage to their drive transistors from 'free-wheeling' them... even at high RPM. The only damage you could do to one is to the bearings or causing a fracture of the rotor due to excessive RPM.
What you may be referring to are AC motors. Which have no magnets, and require current flowing in both the rotor and stator to generate torque. ProTip: Alternators work this way as well. The output of the alternator is regulated by the current applied to a 'field coil' which may be either on the rotor or stator. IIRC the rotor is usually the field coil as that makes the alternator cheaper to build.
Tell me, how much mass is in a bit?
A little basic electronics math will tell you... might be a bit off on this but I think we can get close with:
For electrical transmission over wire the mass of a bit is approximately* proportional to
M =(1/B)*A*V*C*E
Where B is bits per second
A is average impulse current flow within a bit frame.
V is the peak-to-peak voltage of the signal carrying the bits.
C is the Coulomb constant: 6.241509x10^18
E is the mass of one electron 9.109382x10^-31 kg
*This will overestimate the mass of a typical bit for a number of reasons. Not all the combinations of bits in a sequence require the same energy to transmit. This calculation assumes a square bit. Circular, or round(ish) bits are more accurate but the maths are uglier(in /.) to show/write.
Do they teach proper capitalization in your son's kindergarten?
Aww, come on now. They are obviously on NYC time. Trying to apply proper capitalization on an iPhone takes too long.
I guess you missed the 'Foxtrot Alpha Kilo Echo' under the title....
Not very subtle, really.
TL:DR My conclusion is there are some rather noisy and incompetent lawyers that want this guy axed.
After reading through the litany of whinging over this judge, at http://www.therobingroom.com/Judge.aspx?ID=1555
and the few positive ratings, I am wondering if the allegations are really sour grapes from attorneys who
got their bullshit handed back to them on a silver platter. This guy is clearly a hardass. However I saw no clear evidence
that the negative comments were anything but inarticulate nerd-rage. The few positive comments clearly articulated why
the judge deserved positive praise. When someone is howling about being oppressed it is pretty damned important to
note how well argued their position is. I did not see a single negative comment that gave any clarity.
Boole, Church, Turing and others in the Pre-war era stood on the shoulders of Wilkins, Huygens, Hooke, Leibniz, Babbage, Lady Ada Lovelace, and others. All of whom helped define Discrete Mathematics, which is the underpinnings of CS, and CIS.
Your average smart phone these days has more raw grunt, RAM and storage than a high end Cray 2 installation. The Cray 2 was the fastest super computer from 1985 until 1989: 1 GB of RAM 1.9GFLOPS. I think many of the monolithic(non-massively parallel) super computers of the early 90's would only be just eking nose and nose with a high-performance smartphone/tablet. Some of the early massively parallel machines would lose to a smartphone when the GPU(s) are added to the comparison. Oh and the power consumption is just hilarious.
Cray 2 ~250KW..(typical. including cooling infrastructure) iPhone CPU 2W (Max. cpu only. passive cooling)
Heck, a 1MHz 6502 drew more than 2W until the 65c02 came out.
+1 informative.... WTF are my mod points!
I learned to surf that wave in the 70's. When the going gets tough, the tough go surfing...
When the waves die down... you go back into business repairing the new shit.
Wet, apply, lather, rinse, repeat.