This reeks of Smarr, the man who ran an NCSA that did virtually nothing except some flashy demos and Mosaic. Mosaic being a project that was actively discouraged by upper management until it became too successful, whereopon they took credit. Smarr is a master politician, but lacks an eye for people and projects that accomplish something, rather than just looking superficially cool.
These are single precision FLOPS on some apple fractal program optimized for Altivec and undoubtedly embarassingly parallel.
The top500 list is based on double-precision linpack scores. This cluster would not score anywhere near that level on the top500 test because Altivec doesn't do double precision, so you use the regular scalar FPU. Furthermore, you need a fairly fast interconnect to get a good fraction of theoretical peak on linpack, so I would estimate that this cluster wouldn't get more than 40 gflops or so in the top500 test.
P4s can do a double precision vector, and as a result, they get much better linpack scores in a similarly equipped cluster, and for far less money. This is why you don't see big clusters being built out of macs.
The French also have quite a powerful SSBN force. They currently have 2 Triomphant and 2 Roudoutable class boats, each with 96 warheads. Carriers look pretty academic when your 384 biggest cities are radioactive cinders.
Please do tell me where I can buy a dynamically steerable multi-beam phased array antenna for my wireless network. There's a huge difference between a directional antenna, and one that tracks multiple wireless cards as they move with directional beams.
You're obviously unfamiliar with what a phased-array is, so the fact that their product sounds like a cell-phone sticker to you is only a symptom of your own ignorance.
You are incorrect. There is no port number in an IP header. TCP and UDP both have their source and destination ports as the first 4 bytes after the IP header, but ICMP, for example, does not have ports at all. Thus, the concept of a port is defined by the upper layer protocol and has no meaning at the IP level.
In my market, 18/21 phones from Sprint, and 12/13 phones offered from Verizon support Analog. Clearly manufacturers believe that analog is an important feature.
Looking at T-mobile's coverage map: http://www.t-mobile.com/coverage/national_popup.as p, one can see that less than 20% of the landmass of the U.S. is covered. You can't even drive coast to coast on any highway and maintain coverage the whole time.
Compare to this: http://www.mobile911alarms.com/911map.htm, listing total AMPS coverage in North America. Over 90% of the landmass is covered, including almost every mile of Interstate highway.
Sprint may charge 30 cents a minute or so extra for a roaming call, but if I'm broken down on the highway, I'll gladly pay it. You can be in lots of places in the U.S. with a GSM phone, and simply not be able to make a call for any ammount of money in an emergency.
T-Mobile has NO coverage in Vermont. On the other hand, Vermont is fully covered by AMPS. Even if you are in the Northeast, there are major holes in GSM coverage.
It is true that AMPS service will be optional in 2007, but that is quite a while off, and when it happens, you will see analog coverage removed from big cities first where the spectrum is tight. In those areas, if you have a CDMA/AMPS phone, you will have no trouble getting coverage.
Re: AT&T making a business decision that is "anti-consumer", AT&T did not have the option to switch to 1900. The licenses are owned by other operational carriers. What would you have carriers do with their existing 800Mhz licenses? In this age of multi-band phones, why should anyone stop using 50Mhz of prime spectrum just because it's not in the GSM standard?
The vast majority of people do not travel outside the U.S. with any regularity. AMPS works in Canada and Mexico. Having a phone which works in Europe and Asia is simply not an issue for most people. On the other hand, most people do travel with some regularity around the U.S., where GSM coverage is very poor.
You really don't understand the economics of this situation. AT&T owns the licenses to a bunch of 800Mhz spectrum, as well as all the infrastructure to go along with it. Even if they wanted to, they couldn't just switch to 1900 because all of the licenses are owned by other companies. Their decision to use GSM is predicated on using their existing spectrum. AT&T is a big enough customer to warrant making new phones. Furthermore, if you care about roaming, you aren't going to get a GSM phone anyhow from any of the existing providers, because AFAIK, they don't support analog, which is the only service which is fairly universally available in the U.S. There aren't that many people who care about roaming in Europe but not in smaller towns in the U.S.
Europe has the same situation, albeit in different bands. The original GSM system operated at 900Mhz, and the new band is 1800Mhz. It's too bad that there wasn't cooperation between Europe and North America when this was decided, but it's water under the bridge, and people just have to deal with it.
Perhaps AT&T are not idiots, instead, they happen to own a lot of 800mhz spectrum, and aren't going to switch to 1900 (with all its disadvantages) just because you think they are idiots.
Yes the relationship between global warming and the ozone hole is small, I don't know anyone who says otherwise. The problem is that more UV light comes in and gives everyone skin cancer. There is a big lag between emissions of CFCs and the size of the hole because it takes a long time for the CFCs to get out of the ozone layer -- they are a catalyst for the O3-O2 reaction, and they stick around a long time.
Most scientists agree that the reason the ozone hole is starting to reduce in size is the 1976 ban on CFCs in aerosol cans. This was by far the largest source of CFCs ending up in the atmosphere.
The sign probably has vertical V-shaped grooves in it at 90 degree angles to reflect light back to the source if it is at roughly the same height. The manufacturing and reflectivity isn't perfect, so you still see it from other angles.
When you fire the one electron or photon, it's still a wave. This is unexpected bit, as with only a single electron, it has to go through both slits in order to create interference.
CPB funding for PBS is way down, I believe under 20% now. PBS could probably get by without it. Corporate sponsorships are effectively ads. No tax money is lost if this money would have been spent on other forms of advertising, as they are both writeoffs as a business expense. Individual contributions are deductable, however, so are contributions to any other charitable org. There are a number of shows on PBS that simply are social goods, and should be funded in whatever way is practical. Frontline for example is simply outstanding, and serves a serious public interest.
Your free-market dogma simply doesn't work on IP. Standard economics does not produce optimal results when goods have near zero incremental cost, and hence new types of markets or government intervention are required. If you have a solution to this, I'd like to hear it.
Charlie Rose, Nova, Frontline? You don't see shows like this on cable channels. The problem is that the demand for shows is not necessarily the same is the demand for advertising on those shows. Advertisers want the 18-24 demo because they are easy to influence with ads, even though older people have more disposable income. PBS largely survives because people are willing to pay for the programming. Since the people willing to pay for programming are different than the people easily influenced by ads, there is a different set of programming, unavailable even on specialized cable channels.
They are offering $3.00 per share in the buyout, less than the cash holdings of the company. They are effectively offering to buy a pile of cash for less than 90 cents on the dollar. The investors are saying "We'll just take the full $3.41, thanks." The management supports the buyout perhaps because of a sentimental attachment to the company, or perhaps because of golden parachutes they may (disclamer: I do not know this) be getting out of the 41 cents.
The U.S. Army is switching to wheeled armored cars instead of APCs for the new force. You don't need to worry about RPGs killing the tracks, they'll just shoot the tires out with an AK-47.
Padilla was arrested in the U.S., a U.S. citizen, and is currently being held incommunicado in a military prison. According to the justice department, they have no intention of even bringing him before a military tribunal, they're just going to indefinitely imprison him.
Slashdot Mirror
This reeks of Smarr, the man who ran an NCSA that did virtually nothing except some flashy demos and Mosaic. Mosaic being a project that was actively discouraged by upper management until it became too successful, whereopon they took credit. Smarr is a master politician, but lacks an eye for people and projects that accomplish something, rather than just looking superficially cool.
These are single precision FLOPS on some apple fractal program optimized for Altivec and undoubtedly embarassingly parallel.
The top500 list is based on double-precision linpack scores. This cluster would not score anywhere near that level on the top500 test because Altivec doesn't do double precision, so you use the regular scalar FPU. Furthermore, you need a fairly fast interconnect to get a good fraction of theoretical peak on linpack, so I would estimate that this cluster wouldn't get more than 40 gflops or so in the top500 test.
P4s can do a double precision vector, and as a result, they get much better linpack scores in a similarly equipped cluster, and for far less money. This is why you don't see big clusters being built out of macs.
They dug the canal. Panama was chosen, as a canal across the United States was deemed impractical using methods then available.
The French also have quite a powerful SSBN force. They currently have 2 Triomphant and 2 Roudoutable class boats, each with 96 warheads. Carriers look pretty academic when your 384 biggest cities are radioactive cinders.
Please do tell me where I can buy a dynamically steerable multi-beam phased array antenna for my wireless network. There's a huge difference between a directional antenna, and one that tracks multiple wireless cards as they move with directional beams.
You're obviously unfamiliar with what a phased-array is, so the fact that their product sounds like a cell-phone sticker to you is only a symptom of your own ignorance.
You are incorrect. There is no port number in an IP header. TCP and UDP both have their source and destination ports as the first 4 bytes after the IP header, but ICMP, for example, does not have ports at all. Thus, the concept of a port is defined by the upper layer protocol and has no meaning at the IP level.
In my market, 18/21 phones from Sprint, and 12/13 phones offered from Verizon support Analog. Clearly manufacturers believe that analog is an important feature.
s p, one can see that less than 20% of the landmass of the U.S. is covered. You can't even drive coast to coast on any highway and maintain coverage the whole time.
Looking at T-mobile's coverage map: http://www.t-mobile.com/coverage/national_popup.a
Compare to this:
http://www.mobile911alarms.com/911map.htm, listing total AMPS coverage in North America. Over 90% of the landmass is covered, including almost every mile of Interstate highway.
Sprint may charge 30 cents a minute or so extra for a roaming call, but if I'm broken down on the highway, I'll gladly pay it. You can be in lots of places in the U.S. with a GSM phone, and simply not be able to make a call for any ammount of money in an emergency.
T-Mobile has NO coverage in Vermont. On the other hand, Vermont is fully covered by AMPS. Even if you are in the Northeast, there are major holes in GSM coverage.
It is true that AMPS service will be optional in 2007, but that is quite a while off, and when it happens, you will see analog coverage removed from big cities first where the spectrum is tight. In those areas, if you have a CDMA/AMPS phone, you will have no trouble getting coverage.
Re: AT&T making a business decision that is "anti-consumer", AT&T did not have the option to switch to 1900. The licenses are owned by other operational carriers. What would you have carriers do with their existing 800Mhz licenses? In this age of multi-band phones, why should anyone stop using 50Mhz of prime spectrum just because it's not in the GSM standard?
The vast majority of people do not travel outside the U.S. with any regularity. AMPS works in Canada and Mexico. Having a phone which works in Europe and Asia is simply not an issue for most people. On the other hand, most people do travel with some regularity around the U.S., where GSM coverage is very poor.
You really don't understand the economics of this situation. AT&T owns the licenses to a bunch of 800Mhz spectrum, as well as all the infrastructure to go along with it. Even if they wanted to, they couldn't just switch to 1900 because all of the licenses are owned by other companies. Their decision to use GSM is predicated on using their existing spectrum. AT&T is a big enough customer to warrant making new phones. Furthermore, if you care about roaming, you aren't going to get a GSM phone anyhow from any of the existing providers, because AFAIK, they don't support analog, which is the only service which is fairly universally available in the U.S. There aren't that many people who care about roaming in Europe but not in smaller towns in the U.S.
Europe has the same situation, albeit in different bands. The original GSM system operated at 900Mhz, and the new band is 1800Mhz. It's too bad that there wasn't cooperation between Europe and North America when this was decided, but it's water under the bridge, and people just have to deal with it.
Perhaps AT&T are not idiots, instead, they happen to own a lot of 800mhz spectrum, and aren't going to switch to 1900 (with all its disadvantages) just because you think they are idiots.
They didn't put the NEC Earth Simulator on their graph, which at 40 tflops peak is above the line.
Yes the relationship between global warming and the ozone hole is small, I don't know anyone who says otherwise. The problem is that more UV light comes in and gives everyone skin cancer. There is a big lag between emissions of CFCs and the size of the hole because it takes a long time for the CFCs to get out of the ozone layer -- they are a catalyst for the O3-O2 reaction, and they stick around a long time.
Most scientists agree that the reason the ozone hole is starting to reduce in size is the 1976 ban on CFCs in aerosol cans. This was by far the largest source of CFCs ending up in the atmosphere.
If a bullet or shell casing is recovered from a crime scene, wouldn't a database, even if imperfect, narrow down the possible suspects?
You obviously don't know much about LANL.
The sign probably has vertical V-shaped grooves in it at 90 degree angles to reflect light back to the source if it is at roughly the same height. The manufacturing and reflectivity isn't perfect, so you still see it from other angles.
When you fire the one electron or photon, it's still a wave. This is unexpected bit, as with only a single electron, it has to go through both slits in order to create interference.
If you notice, the max is $50,000, so that's all a big company would pay.
CPB funding for PBS is way down, I believe under 20% now. PBS could probably get by without it. Corporate sponsorships are effectively ads. No tax money is lost if this money would have been spent on other forms of advertising, as they are both writeoffs as a business expense. Individual contributions are deductable, however, so are contributions to any other charitable org. There are a number of shows on PBS that simply are social goods, and should be funded in whatever way is practical. Frontline for example is simply outstanding, and serves a serious public interest.
Your free-market dogma simply doesn't work on IP. Standard economics does not produce optimal results when goods have near zero incremental cost, and hence new types of markets or government intervention are required. If you have a solution to this, I'd like to hear it.
Charlie Rose, Nova, Frontline? You don't see shows like this on cable channels. The problem is that the demand for shows is not necessarily the same is the demand for advertising on those shows. Advertisers want the 18-24 demo because they are easy to influence with ads, even though older people have more disposable income. PBS largely survives because people are willing to pay for the programming. Since the people willing to pay for programming are different than the people easily influenced by ads, there is a different set of programming, unavailable even on specialized cable channels.
They are offering $3.00 per share in the buyout, less than the cash holdings of the company. They are effectively offering to buy a pile of cash for less than 90 cents on the dollar. The investors are saying "We'll just take the full $3.41, thanks." The management supports the buyout perhaps because of a sentimental attachment to the company, or perhaps because of golden parachutes they may (disclamer: I do not know this) be getting out of the 41 cents.
All three of them?
The U.S. Army is switching to wheeled armored cars instead of APCs for the new force. You don't need to worry about RPGs killing the tracks, they'll just shoot the tires out with an AK-47.
Padilla was arrested in the U.S., a U.S. citizen, and is currently being held incommunicado in a military prison. According to the justice department, they have no intention of even bringing him before a military tribunal, they're just going to indefinitely imprison him.
O(1) means constant time no matter what size input. O(n) is linear with input size.