Google is at least trying to say "Hey, this whole patent troll environment sucks. You should really do something about this problem!"
Hopefully someone will listen to their complaint before they are forced to take matters into their own hands.
And I think everyone also sees the next step, which is retaliation. Google just bought all those Motorola patents, and having them shut down Nokia and Apple with all those 17-year-old cell phone patents would really be a step up in the Mutually-Assured-Destruction conflict, and everyone would suffer for it.
Taking this approach with the nukes in your back pocket seems much more civil than approach taken by the others.
ITER is a hugely expensive project, and won't produce a commercially viable power generation system.
In a lot of areas where research is done on things which don't work yet -- rockets, bridges, transmission systems, etc -- there's a general idea of how things might be able to "scale up" to meet the goals.
Is tokamak fusion really in sight of being commercially viable source of energy? If we need unobtanium to make a commercially viable reactor, wouldn't it make sense to wait until the materials are viable before making even larger tokamaks? What do we learn from making these new, bigger, more expensive reactors?
Or are we trying to build ever-bigger spark gap transmitters as a way to make radio better? Maybe we should look at other schemes?
Or, alternatively, we know of a nice, large, gravity-fed fusion reactor fairly nearby, is the engineering simpler to harness energy from that on a large scale?
Sure. Most DSPs don't have floating point. Looks like the AVR has an 8x8 multiply, so that really helps. Most DSPs use 16x16 multiplies, accumulating with a 32 bit or larger accumulator. That's going to take several instructions on an 8 bit micro, but it's certainly feasible for things like audio with low data rates and small FFTs. On the other hand, if you're using an Arduino to do audio FFT for things like a spectrum analyzer, this technique won't help, since you're not interested in picking up a few signals, but all the frequency bands.
Oh, I see. The problem is not that pancake flipping itself is hard, that determining the optimal pancake flipping algorithm for a stack is hard. That's believable.
I had something like this was an interview question once. My solution:
Assume we are stacking pancakes with largest at the bottom.
Find the largest unsorted pancake
Flip that to the top
Flip from the bottom-most unsorted pancake. (One additional pancake is now sorted)
Repeat until sorted
To me, assuming that you consider "Find the largest unsorted pancake" to be O(N), the algorithm is O(N^2). Number of flips is 2N. Where's my turing award?
So I must be missing something... Is one not able to find the largest unsorted pancake easily? Perhaps you are only able to look at the size of the topmost pancake. The article was unclear.
The WiMAX network is pretty bad. Coverage is virtually nonexistant, even in cities "with WiMAX coverage" In Austin, there are very few places where WiMAX works... and seemingly never in places like the airport where you actually want it. If you ever happen to get it working, speeds are marginally better than EVDO.
LTE should work much better, and it will align with the rest of the industry.
Engineers manipulate the environment to make it better (more valuable).
Got a desert? Irrigate it and it's much more valuable.
Got a large land mass? Add transportation infrastructure and people can get around easier.
Computing is a new environment that is an extraordinarily awesome environment to be engineered in.
Finance is another complex environment that can be made more valuable with engineering. If the tech companies are really hurting for the smart folks, they'll start paying competitively.
Friends don't tell friends not to go into finance if they think it's in their best interest. It seems like interesting, challenging, profitable work. I don't see why that's wrong.
Do you think that Google has no patents that might apply to you? Do you think that they might have anything that reads on, oh, Bing or your Cloud services? Do you really want to start this battle?
News flash: ARM designs low-performance processors that are also lower power!
Seriously, Slashdot? This is news?
And now ARM is going after high clock rates with deep pipelines. They'll end up with microarchitectures that are are more or less equivalent to x86 ones. Oh, and they're well behind the game when it comes to important architecture features like 64 bit. A 32 bit "server" architecture is a laughable concept.
The real thing that ARM has that x86 doesn't? You can license their core and put it in your SoC, where all the important stuff actually lives.
We see this same ARM article every few weeks. It's the same bull every time.
I'm starting to expect that tomorrow I'll see slashdot / slashvertisement articles that "nature's harmonic simultaneous 4-day time cube" has been proven.
Is this the classic head switch with feedback to adjust the output voltage? This kind of voltage regulator has been around for a long time, and is extremely common in embedded devices. You have the head switch there anyway for power collapse, just add some control to the gate voltage. Not terribly efficient, but you get increased R and so decreased V squared over R. Better than no regulation for small increase in area over what you had already (A big head switch).
Perhaps it's yet another case of Academia "discovering" what someone in industry figured out a long time ago...
Now if it's a easy to fabricate buck converter, it might be interesting... we have to have those off-die. But I think fabricating those capacitors and inductors isn't easy.
FM Radio could interfere with television broadcast channels 5 and 6, or aircraft navigation, since they're right next to each other!
AM Radio could interfere with aircraft beacons, since they're right next to each other!
Please. We've been allocating spectrum for things for a long time. Interference can be monitored and controlled. Do you really think that mobile telephone companies would put up with broadcasters puking all over their spectrum? Or vice versa? Or either putting up with amateur radio interference?
Or, perhaps worst of all, do you think the Hams would put up with someone interfering with their spectrum? They can triangulate secret government projects accidentally using their shortwave spectrum.
Yes, interference happens from all sorts of places. You'll likely find that devices in your adjacent spectrum are less likely to interfere than other sources of interference.
Untrue. A9's are still lower performance per clock than Atom on things like Spec Int (and real-world apps in my experience). And you can't get them at 1.6GHz+. Are you still believing dhrystone numbers are representative?
No 64 bit architecture now... I'll bet anyone they won't have any 64 bit products by 2011.
I guess you could stick a Coretex A9 in a box and call it a server, just like you can stick an Atom in a box and call it a server. At least the atom would support large address spaces.
If they've reduced their ideas to practice, then hooray for them. More ideas that must be licensed on Fair, Reasonable, and Non-Discriminatory grounds.
If they haven't, then their patents may not be valid.
The GSM side of 3G standards has many different upgrades to the basic WCDMA air interface:
HSDPA: 7.2 MBit/sec downlink
HSUPA: 5MBit/sec uplink
HSPA+: 21 MBit/sec - 48 MBit/sec downlink
The most interesting thing is that HSPA+ is getting close to the same efficiency (bits/Hz) as LTE; 21MBit/sec in a 5MHz channel vs. 100MBit/sec in a 20MHz channel.
Haven't disk manufacturers been doing this forever, using faster memories to cache disk? I guess the difference now is that the memory is slower than DRAM and non-volatile so it isn't lost in the event of power failure? Or maybe you can get more flash storage at a low price point?
C. S. Lewis wrote a series of books about this exact premise. In his stories, there is indeed life on other planets, but the life there is not fallen as it is on Earth. The first one is entitled "Out of the Silent Planet," and the collection is known as the "Space Trilogy".
Look, I'm not saying x86 isn't crazy. It doesn't have just shifting addressing modes, but ones with multiplies. That really forces you to have (A) an architecture that uses multicycle instructions, (B) a really horrid pipeline, or (C) splitting up instructions into multiple components that flow through a normal pipeline.
Having shifts in the address calculation is fine for ARM7 where you're trying to squeeze every possible functionality out of a tiny number of gates, and don't really care about performance. But for even a reasonably high-performance design, you need to have a consistent pipeline.
Probably the most important pipeline is the Decode->RegRead->AddressFormation->Dcache->Writeback pipeline. The latency of this pipeline is critical for performance. ARM has some advantages here: uniform (or, somewhat less so, semi-uniform, a la Thumb2) is easier to decode than variable-length at the byte level x86. Most architectures have an adder in the AddressFormation part (though notably not ia64). If you add two registers (which you can't in MIPS) you probably want to be able to shift by the access size because you're doing something like indexing into an array. So a small left shifter before the adder isn't uncommon, and it's usually about a 4:1 mux in terms of delay.
But ARM allows you to do full rotations in front of the adder. This means you need more levels of logic in front of the address calculation adder, which hurts your memory latency. You can make it a multicycle instruction or split it up into multiple instructions (and many implementations do), but that of course adds significant complexity.
The page table formats are kind of kooky. Most 32 bit architectures choose 4K pages as the minimal page size. 4K L1 translation and 4K L2 translation translates all 20 bits you need. The page tables are a multiple of the page size, which is handy. It's so clean, it's pretty obviously the "right" thing to do.
ARM has a 16KB l1 translation, because they used to support 1KB pages, but no longer do. They have strange attributes that move around the format, which makes it more difficult to manipulate the page table entries. They also have no free bits, which makes it a pain in Linux to keep information like how new or clean the page is.
I will say that the page tables are getting cleaner as they deprecate things like 1KB pages, but they're still pretty painful compared with other architectures.
The Alpha Architecture Handbook is a good read, and Alpha is my very favorite RISC. Not that it's magical, either, but it's a lot cleaner than ARM. And it's less than half the length of the ARM Architecture Reference Manual (ARM ARM, which I must admit is a clever acronym).
Slashdot Mirror
Hopefully someone will listen to their complaint before they are forced to take matters into their own hands.
And I think everyone also sees the next step, which is retaliation. Google just bought all those Motorola patents, and having them shut down Nokia and Apple with all those 17-year-old cell phone patents would really be a step up in the Mutually-Assured-Destruction conflict, and everyone would suffer for it.
Taking this approach with the nukes in your back pocket seems much more civil than approach taken by the others.
In a lot of areas where research is done on things which don't work yet -- rockets, bridges, transmission systems, etc -- there's a general idea of how things might be able to "scale up" to meet the goals.
Is tokamak fusion really in sight of being commercially viable source of energy? If we need unobtanium to make a commercially viable reactor, wouldn't it make sense to wait until the materials are viable before making even larger tokamaks? What do we learn from making these new, bigger, more expensive reactors?
Or are we trying to build ever-bigger spark gap transmitters as a way to make radio better? Maybe we should look at other schemes?
Or, alternatively, we know of a nice, large, gravity-fed fusion reactor fairly nearby, is the engineering simpler to harness energy from that on a large scale?
Sure. Most DSPs don't have floating point. Looks like the AVR has an 8x8 multiply, so that really helps. Most DSPs use 16x16 multiplies, accumulating with a 32 bit or larger accumulator. That's going to take several instructions on an 8 bit micro, but it's certainly feasible for things like audio with low data rates and small FFTs. On the other hand, if you're using an Arduino to do audio FFT for things like a spectrum analyzer, this technique won't help, since you're not interested in picking up a few signals, but all the frequency bands.
Not very usable for things that we need super-fast FFTs for, a gazillion times a second, like LTE.
I wonder if this is just re-discovering compressed sensing.
Oh, I see. The problem is not that pancake flipping itself is hard, that determining the optimal pancake flipping algorithm for a stack is hard. That's believable.
Assume we are stacking pancakes with largest at the bottom.
To me, assuming that you consider "Find the largest unsorted pancake" to be O(N), the algorithm is O(N^2). Number of flips is 2N. Where's my turing award?
So I must be missing something... Is one not able to find the largest unsorted pancake easily? Perhaps you are only able to look at the size of the topmost pancake. The article was unclear.
Smashing all the nuclear sites around the world in the 1880s would not be very disastrous.
LTE should work much better, and it will align with the rest of the industry.
We read when we're bored, but what's the point in commenting among the noise?
Got a desert? Irrigate it and it's much more valuable.
Got a large land mass? Add transportation infrastructure and people can get around easier.
Computing is a new environment that is an extraordinarily awesome environment to be engineered in.
Finance is another complex environment that can be made more valuable with engineering. If the tech companies are really hurting for the smart folks, they'll start paying competitively.
Friends don't tell friends not to go into finance if they think it's in their best interest. It seems like interesting, challenging, profitable work. I don't see why that's wrong.
Do you think that Google has no patents that might apply to you? Do you think that they might have anything that reads on, oh, Bing or your Cloud services? Do you really want to start this battle?
Seriously, Slashdot? This is news?
And now ARM is going after high clock rates with deep pipelines. They'll end up with microarchitectures that are are more or less equivalent to x86 ones. Oh, and they're well behind the game when it comes to important architecture features like 64 bit. A 32 bit "server" architecture is a laughable concept.
The real thing that ARM has that x86 doesn't? You can license their core and put it in your SoC, where all the important stuff actually lives.
We see this same ARM article every few weeks. It's the same bull every time. I'm starting to expect that tomorrow I'll see slashdot / slashvertisement articles that "nature's harmonic simultaneous 4-day time cube" has been proven.
Perhaps it's yet another case of Academia "discovering" what someone in industry figured out a long time ago...
Now if it's a easy to fabricate buck converter, it might be interesting... we have to have those off-die. But I think fabricating those capacitors and inductors isn't easy.
Are there more people on Itanium VMS than Alpha yet? I hadn't heard that was the case...
AM Radio could interfere with aircraft beacons, since they're right next to each other!
Please. We've been allocating spectrum for things for a long time. Interference can be monitored and controlled. Do you really think that mobile telephone companies would put up with broadcasters puking all over their spectrum? Or vice versa? Or either putting up with amateur radio interference?
Or, perhaps worst of all, do you think the Hams would put up with someone interfering with their spectrum? They can triangulate secret government projects accidentally using their shortwave spectrum.
Yes, interference happens from all sorts of places. You'll likely find that devices in your adjacent spectrum are less likely to interfere than other sources of interference.
Untrue. A9's are still lower performance per clock than Atom on things like Spec Int (and real-world apps in my experience). And you can't get them at 1.6GHz+. Are you still believing dhrystone numbers are representative?
They should charge them with manufacturing child pornography.
You could send them an invoice for the use of your work.
I guess you could stick a Coretex A9 in a box and call it a server, just like you can stick an Atom in a box and call it a server. At least the atom would support large address spaces.
If they haven't, then their patents may not be valid.
The most interesting thing is that HSPA+ is getting close to the same efficiency (bits/Hz) as LTE; 21MBit/sec in a 5MHz channel vs. 100MBit/sec in a 20MHz channel.
Haven't disk manufacturers been doing this forever, using faster memories to cache disk? I guess the difference now is that the memory is slower than DRAM and non-volatile so it isn't lost in the event of power failure? Or maybe you can get more flash storage at a low price point?
So even if Google doesn't index, say, the Wall Street Journal, can't Google still get the same news contributions form the AP newswire?
Or is there something special about AP license terms or something?
C. S. Lewis wrote a series of books about this exact premise. In his stories, there is indeed life on other planets, but the life there is not fallen as it is on Earth. The first one is entitled "Out of the Silent Planet," and the collection is known as the "Space Trilogy".
Having shifts in the address calculation is fine for ARM7 where you're trying to squeeze every possible functionality out of a tiny number of gates, and don't really care about performance. But for even a reasonably high-performance design, you need to have a consistent pipeline.
Probably the most important pipeline is the Decode->RegRead->AddressFormation->Dcache->Writeback pipeline. The latency of this pipeline is critical for performance. ARM has some advantages here: uniform (or, somewhat less so, semi-uniform, a la Thumb2) is easier to decode than variable-length at the byte level x86. Most architectures have an adder in the AddressFormation part (though notably not ia64). If you add two registers (which you can't in MIPS) you probably want to be able to shift by the access size because you're doing something like indexing into an array. So a small left shifter before the adder isn't uncommon, and it's usually about a 4:1 mux in terms of delay.
But ARM allows you to do full rotations in front of the adder. This means you need more levels of logic in front of the address calculation adder, which hurts your memory latency. You can make it a multicycle instruction or split it up into multiple instructions (and many implementations do), but that of course adds significant complexity.
The page table formats are kind of kooky. Most 32 bit architectures choose 4K pages as the minimal page size. 4K L1 translation and 4K L2 translation translates all 20 bits you need. The page tables are a multiple of the page size, which is handy. It's so clean, it's pretty obviously the "right" thing to do.
ARM has a 16KB l1 translation, because they used to support 1KB pages, but no longer do. They have strange attributes that move around the format, which makes it more difficult to manipulate the page table entries. They also have no free bits, which makes it a pain in Linux to keep information like how new or clean the page is.
I will say that the page tables are getting cleaner as they deprecate things like 1KB pages, but they're still pretty painful compared with other architectures.
The Alpha Architecture Handbook is a good read, and Alpha is my very favorite RISC. Not that it's magical, either, but it's a lot cleaner than ARM. And it's less than half the length of the ARM Architecture Reference Manual (ARM ARM, which I must admit is a clever acronym).