Which means you need a hot emitter, a cool colector, and in between something that will pass electrons, but not too much heat. Basically, a losing proposition, as anything that passes electrons is almost by definition an excellent conductor of heat. Try to think of somethign that conducts electricity but insulates heat. Hard to come up with isnt it?
Well, it does sound as though that's the problem they've solved. Reading the article, it sounds as though they've found some kind of semiconductor that works as you describe.
Yes, except that this is a completely different mechanism, taking what used to be known as a "vacuum thermionic converter" and replacing the vacuum with (what sounds like) a layer of silicon. I don't know how they made that work, but if it does, it'll make manufacturing these things substantially easier...
I think you're referring to the UK, but if you're referring to the U.S., 99.999% of the time insurance follows the vehicle. Yes, policies do exist which follow the Named Insured, but, again, in the U.S. they are rare and more expensive.
In the UK, all insurance is with the driver. On *most* policies, you must give the details of a specific car to be covered. But there are multiple variations. You can get:
* Insurance that will allow any qualified driver to drive your vehicle (usually with the restriction that they're over 21) * Insurance that covers you (or multiple people) for multiple, unspecified vehicles (usually only sold to the motor trade, although you can get it in other circumstances if you ask) * Insurance that covers a specific vehicle plus any other vehicle you are driving (although it is expected that the named vehicle is your "primary" vehicle, which could cause your insurance to be invalidated if it isn't)
All three of these are only slightly more expensive than the single-driver, single-car policy that's most common, and there are plenty of circumstances in which you might find it cheaper to get one of these than multiple policies to cover different vehicles and/or drivers.
If you do not have insurance or a car (therefore no need for insurance) and borrow a friend's car to drive, you are driving under your friend's policy which follows their car.
In the UK, you'd usually need them to make a special arrangement with their insurers for this (i.e., they call the insurer, tell them that you'll be driving the car, give them some basic information about you, and pay a small extra premium to have you covered).
Most of the time, the only time it follows the driver is if you rent a car. Most insurance policies follow the driver in this instance. This does varies by company so you may want to ask what your insurance company's policy is in regards to rental cars.
Funnily enough, in the UK this is the most common situation where the insurance follows the car.:)
That may be true for insurance, but in the UK, every car is required to have "road tax" paid on it. The tax is paid on the car, not the driver. If the registered owner of the car has not paid roadtax and has not declared the car "off the road", then he is committing an offence - even if he has sold the car to someone else (and hasn't registered the transaction with the DVLA).
True, but in this case there's no need to identify the driver, is there? So I don't see why fingerprinting the driver would be useful. Just make a note that the car has been seen on the road, and get a fine sent to the car's registered keeper. Much simpler. Much less hassle for everyone. No need for scary-sounding technology that we have to trust the police are using the way they're supposed to, and not abusing somehow.
Seems like the easier solution, and the far less creepy one, is just to hook the police cars up with terminals that communicate with the drivers-license database, including its photos.
While all new issue drivers licenses in the UK are photocard licenses, this has only been the case for around six or seven years. Licenses issued prior to that don't have an associated photo, and there's no law requiring replacement of licenses over time. These two factors together effectively make such a database (even if one exists, which I'm not sure about) useless.
By registration, what do you mean? If you mean what we call the V5 or "logbook" (i.e., a document that changes hands along with the car as an official ownership record), no it isn't a requirement to carry it. It's required when purchasing tax for the vehicle (which must be prominently displayed in the vehicle whenever it is on the public road), or if the police give you notice requiring you to produce it, but that's about it.
Officers will scan a vehicle's number plates using a special camera that checks if the car is subject to an offence, like being uninsured.
Being insured is a state of the driver, not the vehicle. To imply that "not being on a database of cars that have been named by somebody as their primary vehicle when purchasing insurance" is equivalent to being "subject to an offence" is just wrong. This technique throws up a huge number of both false positives and false negatives.
If the driver does not convince police he is giving them a correct name, they will fingerprint him and verify his identity on the spot, instead of taking him to the police station.
Generally, at the moment (having been subject to such a stop) I can tell you that the police do believe you, in most cases. You give them your name, try to phone an insurance company, and if they can't verify your identity on the spot, you're given 14 days to send proof of being insured to the police. But will they believe you if they have this fingerprinting machine? Is refusing to be fingerprinted enough to make them take you to the station? I suspect so.
I'd personally say that independent verification isn't really necessary here. The kid was using an established design that is well known to work, and all you have to do is crank up the voltage and look at the pretty glow.
This is what I want to know. If fusion had occurred (and I doubt it, but still...) then there would have been a burst of neutrons. If he was near enough to the experiment to see a ball of plasma then he's very lucky to be alive.
According to the wikipedia article, "neutron emissions can present a hazard if voltages above 40 kilovolts are used". Sounds like the kid was cutting it fairly close, but should have been reasonably safe. It sounds as though the associated X-ray emissions are actually more problematic.
What we don't have yet is fusion power plants. But then again, that isn't what Bussard is proposing in TFA either [...] Getting a fusion reaction to occur is damn hard; getting a self sufficient reaction to occur is still beyond our reach.
He's saying he'll be there in 5 years. He already claims to have a large number of orders-of-magnitude improvement over previous devices of similar design.
I did not hear him addres the biggest problem with IEC: bremsstrahlung
You're right that he didn't. But if you search on scholar.google.com, you'll find he's published papers on the topic. I don't have access to read them, but hopefully that means he's made progress.
Also, his spiel on what this machine would do if it worked is unnecessary: we all know that a high-efficiency fusion machine would change the world, but we need to be convinced that he can build one.
He's been giving that spiel since the 1970s. When he started, it was important: few people realised how useful almost-free-energy would be. Hopefully, he'll keep giving it until the day it comes true.
Moreover, political bickering (the bane of any multi-national project) has gotten in the way more than once, most recently with the question of where to build the ITER project.
Or even a non-multinational one. Watch the video: Bussard talks about the problems he's had in funding. Most people were too sold on tokamak technology to fund anything alternative. He eventually got funding from the Navy, but they couldn't give him much, because (he says) if they put too much into it, the project would start appearing on top level reports, at which point other departments would start asking why the Navy was funding a project which should have been an AEC project, at which point it would be transferred and killed for political reasons. And then his funding was cancelled in an attempt to pay for Iraq.
Mr Bussard's attack is nothing but mudflinging and blaming unspecified others in the goverment for not funding his research
Well, actually he's quite clear in the video that he blames the AEC for not funding it. And is quite thankful for the DoD for having funded it, although they weren't able to provide all the funding he needed for (quite plausible-sounding) political reasons.
even though he cannot (or will not) actually demonstrate he has something worth funding. (This is, in and of itself, reason to apply the 'kook' label.)
1. I don't believe he has ever refused to demonstrate anything; I certainly see no evidence of this in any of the places people have pointed to here. That he cannot is quite reasonable, given that he no longer has access to the funding that he had when working for the Navy.
At best he deserves a couple of thousand for a few copies for a paper ready to be submitted for peer review. Demanding money, and refusing to supply the data required to determine what that funding is to be used for is ludicrous.
Who'se refusing to supply any data? As far as I can see, he has been quite open. He has explained the principles of operation of his device, and the issues (both physical and engineering related) that he has already overcome in its development. He's open about what the results he achieved were, including specific numbers. He has presented a paper to at least one conference, and is (he claims) working on a more detailed paper now.
The Bussard Ramjet is one of the finest pieces of Pseudo scientific speculation ever dreamed of and integrated into Science Fiction works.
The Bussard Ramjet is not psuedo-scientific. It might not be viable as an engineering proposal given current technological restrictions and actual information about particle densities in our region of space (data collected, BTW, after Bussard made the proposal), but that doesn't mean the science involved is flawed.
I too like Mr. Bussard a great deal, and respect and admire his numerous contributions to our culture and to science fiction. However, it has become clear to me that Mr. Bussard no longer is the man he once was. He, most unfortunately, appears to have become senile, vindictive and single-minded to the point of blindness; read what he says, how he defends his project while attacking all other research constantly.
So he thinks that the current tokomak-oriented research is a dead end unlikely to produce useful results. He may be right; certainly there are quite a few fellow sceptics out there. I'm not qualified to judge his comments about his own project or others', but then I'm pretty certain you aren't either. And I haven't heard anything about this from anyone who is, either.
Mr. Bussard has never been able to reproduce any of his results in front of impartial peers, under controlled conditions. Read his letter on JREF, and see for yourself.
Yes, and he admits as much in this video, which you clearly haven't watched.
Mr. Bussard claims to have tested his device a few times and achieved success, but whenever he tried to test it under controlled conditions, it failed - and he blamed some obscure technical malfunction for this inability to achieve any measurable results.
Err... no. He doesn't claim that in either this video, or the post in the comment thread you appear to be talking about. He claims he had a working device, but that it was structurally damaged during testing and he no longer has the funding to recreate that experiment.
Then he says that only by having 200M$ can he show that his techniques work - he will not rebuild his original demonstration machine, nor allow anybody to do so.
1. In the Google video, he clearly states that he *will* rebuild the original demonstration machine, at a cost of $2M. 2. Can you point me to any evidence that he has attempted to prevent somebody else from doing so?
According to Mr. Bussard, it is easy to test for the proper operation of his machine, hence confirming that scaling the machine up in a 200M$ version would produce lots and lots of energy. However, he refuses to construct such a workable prototype and have it tested by independent experts.
Again, this isn't true. If you watch the video, you'll see him explain that a small part of the $200M is for creating an initial test device, a relatively minor improvement on the device he has already built, for the purposes of demonstration. He argues that building a 1/2 scale model would be inappropriate, because it would prove nothing that the 1/10th and 1/8th scale models he has already built and is proposing to recreate hasn't already shown. He plans to recreate a 1/8th model and then move straight to a full scale one. Reading between the lines, I suspect he feels this is a waste of time when not doing it could save a substantial amount of time (if not money) but accepts that he will need to do it to secure the funding.
Bussard himself is relatively famous as is, btw. His 'Bussard ramscoops' were adopted into science fiction with Star Trek: TNG (And abckwards adapted into the original series).
You need to break out your science fiction enclopedia. Bussard ramjets were used much earlier than that. I believe Larry Niven may have been the first to use them, back in the 70s.
Of course it was later discovered that Bussard Ramscoops wouldn't work in our area of space because the density of stray particles is to low near us to make them useful, but the idea still has merit once we get away from our current region of space where higher particle densities exist.
I'm not sure this is an entirely accurate summary of the issues. Obviously, their usefulness varies with the amount of power available to the magnetic field. Current calculations are based on fairly small nuclear fission power sources which don't allow for large scoops and a fairly slow electrically-accelerated exhaust. Switch that to a small fusion source with a fusion-accelerated exhaust, though, and you're likely to have substantially more space to catch ions and more acceleration from each ion captured. Also, at worst case, the scoop performs somewhat better than an equivalent-sized solar sail.
Because Google is in the information-retrieval business, and not the power-generating business.
My understanding is that Google already have power generation capacity; I believe they own a hydroelectric generator in the vicinity of one of their datacentres, and are of course installing a large solar array at their head office.
OK, so fusion power would be a bit of a step up, but...
Well then, if you really believe this *cough*load of tripe*cough* then it won't be a big deal if we store the inevitable waste products, say, in your living room?
You'll struggle. Helium has an annoying tendency to escape from non-hermetically sealed areas.
Instead of spending hundred of millions of dollars, they should first try to invest just a few millions in the Focus Fusion Society. If they idea works it would be a really neat solution. It is really sad that they have so much problems to get funds.
That said, the fact that Bussard has struggled to fund his program is saddening. Unlike Lerner, he's emminently qualified for the research he's undertaking.
Mod parent up. The current bandwidth in processor design is getting data from the memory to the CPU fast enough, and has been for *ages* -- when was the last time you had a machine where the RAM fast enough for there to be no delay cycles, ever? The latest I know of was the BBC Model B: a machine that ran on a 2MHz 6502. While moving RAM to the processor would probably not *solve* this issue, it might be a significant improvement over the current state of affairs.
A fairly large proportion of the power requirement of a general-purpose CPU goes into running the L2 cache. I'm not familiar with GPU design techniques, but I'd hazard a guess that the same is true there. If this is the case, a design with a unified cache would not only be more efficient time-wise (because data produced by the CPU is already in the cache when the GPU comes to render it) but energy-wise as well (because a smaller overall cache is necessary).
I don't think that's a common setup, though. Most servers I've worked with are built from commodity components, because it's cheaper. Yeah, if things go really wrong you need to connect a monitor, but really, how often will you need to do that? Usually, first time you install an OS on it, and then again if you completely reinstall due to disk failure (or whatever).
Slashdot Mirror
Which means you need a hot emitter, a cool colector, and in between something that will pass electrons, but not too much heat. Basically, a losing proposition, as anything that passes electrons is almost by definition an excellent conductor of heat. Try to think of somethign that conducts electricity but insulates heat. Hard to come up with isnt it?
Well, it does sound as though that's the problem they've solved. Reading the article, it sounds as though they've found some kind of semiconductor that works as you describe.
Yes, except that this is a completely different mechanism, taking what used to be known as a "vacuum thermionic converter" and replacing the vacuum with (what sounds like) a layer of silicon. I don't know how they made that work, but if it does, it'll make manufacturing these things substantially easier...
'Cause a cache of a 404 page is really useful.
I think you're referring to the UK, but if you're referring to the U.S., 99.999% of the time insurance follows the vehicle. Yes, policies do exist which follow the Named Insured, but, again, in the U.S. they are rare and more expensive.
:)
In the UK, all insurance is with the driver. On *most* policies, you must give the details of a specific car to be covered. But there are multiple variations. You can get:
* Insurance that will allow any qualified driver to drive your vehicle (usually with the restriction that they're over 21)
* Insurance that covers you (or multiple people) for multiple, unspecified vehicles (usually only sold to the motor trade, although you can get it in other circumstances if you ask)
* Insurance that covers a specific vehicle plus any other vehicle you are driving (although it is expected that the named vehicle is your "primary" vehicle, which could cause your insurance to be invalidated if it isn't)
All three of these are only slightly more expensive than the single-driver, single-car policy that's most common, and there are plenty of circumstances in which you might find it cheaper to get one of these than multiple policies to cover different vehicles and/or drivers.
If you do not have insurance or a car (therefore no need for insurance) and borrow a friend's car to drive, you are driving under your friend's policy which follows their car.
In the UK, you'd usually need them to make a special arrangement with their insurers for this (i.e., they call the insurer, tell them that you'll be driving the car, give them some basic information about you, and pay a small extra premium to have you covered).
Most of the time, the only time it follows the driver is if you rent a car. Most insurance policies follow the driver in this instance. This does varies by company so you may want to ask what your insurance company's policy is in regards to rental cars.
Funnily enough, in the UK this is the most common situation where the insurance follows the car.
That may be true for insurance, but in the UK, every car is required to have "road tax" paid on it. The tax is paid on the car, not the driver. If the registered owner of the car has not paid roadtax and has not declared the car "off the road", then he is committing an offence - even if he has sold the car to someone else (and hasn't registered the transaction with the DVLA).
True, but in this case there's no need to identify the driver, is there? So I don't see why fingerprinting the driver would be useful. Just make a note that the car has been seen on the road, and get a fine sent to the car's registered keeper. Much simpler. Much less hassle for everyone. No need for scary-sounding technology that we have to trust the police are using the way they're supposed to, and not abusing somehow.
Seems like the easier solution, and the far less creepy one, is just to hook the police cars up with terminals that communicate with the drivers-license database, including its photos.
While all new issue drivers licenses in the UK are photocard licenses, this has only been the case for around six or seven years. Licenses issued prior to that don't have an associated photo, and there's no law requiring replacement of licenses over time. These two factors together effectively make such a database (even if one exists, which I'm not sure about) useless.
By registration, what do you mean? If you mean what we call the V5 or "logbook" (i.e., a document that changes hands along with the car as an official ownership record), no it isn't a requirement to carry it. It's required when purchasing tax for the vehicle (which must be prominently displayed in the vehicle whenever it is on the public road), or if the police give you notice requiring you to produce it, but that's about it.
Officers will scan a vehicle's number plates using a special camera that checks if the car is subject to an offence, like being uninsured.
Being insured is a state of the driver, not the vehicle. To imply that "not being on a database of cars that have been named by somebody as their primary vehicle when purchasing insurance" is equivalent to being "subject to an offence" is just wrong. This technique throws up a huge number of both false positives and false negatives.
If the driver does not convince police he is giving them a correct name, they will fingerprint him and verify his identity on the spot, instead of taking him to the police station.
Generally, at the moment (having been subject to such a stop) I can tell you that the police do believe you, in most cases. You give them your name, try to phone an insurance company, and if they can't verify your identity on the spot, you're given 14 days to send proof of being insured to the police. But will they believe you if they have this fingerprinting machine? Is refusing to be fingerprinted enough to make them take you to the station? I suspect so.
I'd personally say that independent verification isn't really necessary here. The kid was using an established design that is well known to work, and all you have to do is crank up the voltage and look at the pretty glow.
This is what I want to know. If fusion had occurred (and I doubt it, but still...) then there would have been a burst of neutrons. If he was near enough to the experiment to see a ball of plasma then he's very lucky to be alive.
According to the wikipedia article, "neutron emissions can present a hazard if voltages above 40 kilovolts are used". Sounds like the kid was cutting it fairly close, but should have been reasonably safe. It sounds as though the associated X-ray emissions are actually more problematic.
What we don't have yet is fusion power plants. But then again, that isn't what Bussard is proposing in TFA either [...] Getting a fusion reaction to occur is damn hard; getting a self sufficient reaction to occur is still beyond our reach.
He's saying he'll be there in 5 years. He already claims to have a large number of orders-of-magnitude improvement over previous devices of similar design.
I did not hear him addres the biggest problem with IEC: bremsstrahlung
You're right that he didn't. But if you search on scholar.google.com, you'll find he's published papers on the topic. I don't have access to read them, but hopefully that means he's made progress.
Also, his spiel on what this machine would do if it worked is unnecessary: we all know that a high-efficiency fusion machine would change the world, but we need to be convinced that he can build one.
He's been giving that spiel since the 1970s. When he started, it was important: few people realised how useful almost-free-energy would be. Hopefully, he'll keep giving it until the day it comes true.
Moreover, political bickering (the bane of any multi-national project) has gotten in the way more than once, most recently with the question of where to build the ITER project.
Or even a non-multinational one. Watch the video: Bussard talks about the problems he's had in funding. Most people were too sold on tokamak technology to fund anything alternative. He eventually got funding from the Navy, but they couldn't give him much, because (he says) if they put too much into it, the project would start appearing on top level reports, at which point other departments would start asking why the Navy was funding a project which should have been an AEC project, at which point it would be transferred and killed for political reasons. And then his funding was cancelled in an attempt to pay for Iraq.
Yes, politics gets in the way.
Mr Bussard's attack is nothing but mudflinging and blaming unspecified others in the goverment for not funding his research
Well, actually he's quite clear in the video that he blames the AEC for not funding it. And is quite thankful for the DoD for having funded it, although they weren't able to provide all the funding he needed for (quite plausible-sounding) political reasons.
even though he cannot (or will not) actually demonstrate he has something worth funding. (This is, in and of itself, reason to apply the 'kook' label.)
1. I don't believe he has ever refused to demonstrate anything; I certainly see no evidence of this in any of the places people have pointed to here. That he cannot is quite reasonable, given that he no longer has access to the funding that he had when working for the Navy.
At best he deserves a couple of thousand for a few copies for a paper ready to be submitted for peer review. Demanding money, and refusing to supply the data required to determine what that funding is to be used for is ludicrous.
Who'se refusing to supply any data? As far as I can see, he has been quite open. He has explained the principles of operation of his device, and the issues (both physical and engineering related) that he has already overcome in its development. He's open about what the results he achieved were, including specific numbers. He has presented a paper to at least one conference, and is (he claims) working on a more detailed paper now.
The Bussard Ramjet is one of the finest pieces of Pseudo scientific speculation ever dreamed of and integrated into Science Fiction works.
The Bussard Ramjet is not psuedo-scientific. It might not be viable as an engineering proposal given current technological restrictions and actual information about particle densities in our region of space (data collected, BTW, after Bussard made the proposal), but that doesn't mean the science involved is flawed.
I too like Mr. Bussard a great deal, and respect and admire his numerous contributions to our culture and to science fiction. However, it has become clear to me that Mr. Bussard no longer is the man he once was. He, most unfortunately, appears to have become senile, vindictive and single-minded to the point of blindness; read what he says, how he defends his project while attacking all other research constantly.
So he thinks that the current tokomak-oriented research is a dead end unlikely to produce useful results. He may be right; certainly there are quite a few fellow sceptics out there. I'm not qualified to judge his comments about his own project or others', but then I'm pretty certain you aren't either. And I haven't heard anything about this from anyone who is, either.
Mr. Bussard has never been able to reproduce any of his results in front of impartial peers, under controlled conditions. Read his letter on JREF, and see for yourself.
Yes, and he admits as much in this video, which you clearly haven't watched.
Mr. Bussard claims to have tested his device a few times and achieved success, but whenever he tried to test it under controlled conditions, it failed - and he blamed some obscure technical malfunction for this inability to achieve any measurable results.
Err... no. He doesn't claim that in either this video, or the post in the comment thread you appear to be talking about. He claims he had a working device, but that it was structurally damaged during testing and he no longer has the funding to recreate that experiment.
Then he says that only by having 200M$ can he show that his techniques work - he will not rebuild his original demonstration machine, nor allow anybody to do so.
1. In the Google video, he clearly states that he *will* rebuild the original demonstration machine, at a cost of $2M.
2. Can you point me to any evidence that he has attempted to prevent somebody else from doing so?
According to Mr. Bussard, it is easy to test for the proper operation of his machine, hence confirming that scaling the machine up in a 200M$ version would produce lots and lots of energy. However, he refuses to construct such a workable prototype and have it tested by independent experts.
Again, this isn't true. If you watch the video, you'll see him explain that a small part of the $200M is for creating an initial test device, a relatively minor improvement on the device he has already built, for the purposes of demonstration. He argues that building a 1/2 scale model would be inappropriate, because it would prove nothing that the 1/10th and 1/8th scale models he has already built and is proposing to recreate hasn't already shown. He plans to recreate a 1/8th model and then move straight to a full scale one. Reading between the lines, I suspect he feels this is a waste of time when not doing it could save a substantial amount of time (if not money) but accepts that he will need to do it to secure the funding.
Bussard himself is relatively famous as is, btw. His 'Bussard ramscoops' were adopted into science fiction with Star Trek: TNG (And abckwards adapted into the original series).
You need to break out your science fiction enclopedia. Bussard ramjets were used much earlier than that. I believe Larry Niven may have been the first to use them, back in the 70s.
Of course it was later discovered that Bussard Ramscoops wouldn't work in our area of space because the density of stray particles is to low near us to make them useful, but the idea still has merit once we get away from our current region of space where higher particle densities exist.
I'm not sure this is an entirely accurate summary of the issues. Obviously, their usefulness varies with the amount of power available to the magnetic field. Current calculations are based on fairly small nuclear fission power sources which don't allow for large scoops and a fairly slow electrically-accelerated exhaust. Switch that to a small fusion source with a fusion-accelerated exhaust, though, and you're likely to have substantially more space to catch ions and more acceleration from each ion captured. Also, at worst case, the scoop performs somewhat better than an equivalent-sized solar sail.
Because Google is in the information-retrieval business, and not the power-generating business.
My understanding is that Google already have power generation capacity; I believe they own a hydroelectric generator in the vicinity of one of their datacentres, and are of course installing a large solar array at their head office.
OK, so fusion power would be a bit of a step up, but...
the power output of these scales as something like the 7th or 9th power of the radius of the device (don't quote me on these numbers)
In TFV, Bussard said 7th power.
The theoretical possibility of aneutronic fusion deserves a mention
Particularly seeing as it's what TFA is talking about. Bussard's proposal is for a cheap mechanism of producing proton/Boron-11 fusion.
Well then, if you really believe this *cough*load of tripe*cough* then it won't be a big deal if we store the inevitable waste products, say, in your living room?
You'll struggle. Helium has an annoying tendency to escape from non-hermetically sealed areas.
Instead of spending hundred of millions of dollars, they should first try to invest just a few millions in the Focus Fusion Society. If they idea works it would be a really neat solution. It is really sad that they have so much problems to get funds.
It's hard to get funds for scientific research when you don't even have a doctorate, and your theory postulates that almost all of modern cosmology is wrong. And, frankly, I'm glad that this is the case -- it focusses the money where it's most likely to be useful.
That said, the fact that Bussard has struggled to fund his program is saddening. Unlike Lerner, he's emminently qualified for the research he's undertaking.
Mod parent up. The current bandwidth in processor design is getting data from the memory to the CPU fast enough, and has been for *ages* -- when was the last time you had a machine where the RAM fast enough for there to be no delay cycles, ever? The latest I know of was the BBC Model B: a machine that ran on a 2MHz 6502. While moving RAM to the processor would probably not *solve* this issue, it might be a significant improvement over the current state of affairs.
A fairly large proportion of the power requirement of a general-purpose CPU goes into running the L2 cache. I'm not familiar with GPU design techniques, but I'd hazard a guess that the same is true there. If this is the case, a design with a unified cache would not only be more efficient time-wise (because data produced by the CPU is already in the cache when the GPU comes to render it) but energy-wise as well (because a smaller overall cache is necessary).
I don't think that's a common setup, though. Most servers I've worked with are built from commodity components, because it's cheaper. Yeah, if things go really wrong you need to connect a monitor, but really, how often will you need to do that? Usually, first time you install an OS on it, and then again if you completely reinstall due to disk failure (or whatever).