so a company that does 3d design needs new cards for their systems.. what do you suggest they do?...buy an assortment of cards and try every one themselves then decide which is best then try to return the other cards?
Well, yes. You don't make that kind of investment (you're probably talking a bunch of cards, not just one to outfit a whole shop) without testing how it works in your environment. Hell, if you're going to buy enough of them you can probably get the vendors to loan you a test sample. To buy on benchmarks is just nuts.
Their systems do certain things faster than beowulfs, and frankly, for some govt. agencies/companies faster is more important than cheaper.
Some != viable market. The government can (and does) keep them alive, but it won't make them huge or give them the kind of dominance in the industry they had 15 years ago. Even if they wanted to, cray's competitors (e.g., IBM) would pitch a fit on capitol hill. Consider what well funded agency will give them the kind of money you seem to think is out there. DoE bought something from them, but their big Red Storm deal is basically a commodity cluster. It gets cray cashflow but isn't revolutionary and isn't a special purpose vector super. The three letter agencies have money, but you have to wonder how much demand for vector supers they really have. (Crypto breaking is better on custom hardware, and echelon is a combination of highly parallelizable chunks and big databases.) Big vector supers are great for things like weather forecasting, but there's only so much money in that field (nothing like the national defense apps). So there's definately demand out there, but demand is nothing without funding.:-)
Obviously Tera thinks so, too, since they purchased them in 2000, and profits/sales ARE up
You're reading too much (or the wrong things) into that deal. Neither company had much of a choice in the matter (tera needed revenue and cray needed to get away from sgi) and things were probably decided more by their government customers then by accountants.
You are correct that yes, this IS a very very limited market, but when you sell them for a billion bucks each, you don't need to match Dell's volume to make a profit.
You're dramatically overestimating the size of the market. Cray's own website puts it at about $1.1bn worldwide, and it's not like cray will get 100% market share. The ongoing R&D costs are a staggering percentage of their revenue, to the point that if the NSA wasn't subsidizing them it's unlikely they'd be alive today. The same goes for other pure supercomputing ventures--without huge amounts of government largess they're sunk.
My thinking, however, is that the same is true today and for all of the top 100 supercomputers in the world. That is to say, each one of those machines is a custom hardware installation,
Yes and no. The problem is that a cray box has to cover the whole R&D cost for an entire system. When IBM sells you an SP2 most of the R&D is spread across their much higher volume business lines. Same with an intel based cluster--the technology specific to the HPC market is basically the interconnect, and the rest is subsidized by video game players. There's also the compiler cost (you don't sell many fortran compilers outside the scientific market) but the salaries for a few compiler writers is much lower than the cost of desiging a cutting-edge cpu from scratch.
At the same time, however, any of these applications are fully capable of utilizing as much hardware resources as you have available.
That's always true. The question is whether they can use the resources efficiently, and whether the cost/op is competetive. You're right about the algorithms being the driving force, but I'd argue that it is unusual for an algorithm that's optimized for one architecture to run optimally if you move it a radically different architecture. People can spend years trying to squeeze a couple more percent out of their code, and they don't want to start from scratch unless there's a very good reason. Then there's the problem that researchers tend to not work in a bubble. Even if you can afford to buy the most expensive machine on the block you might end up shooting yourself in the foot if nobody else in your field can collaborate with you.
user interface software is kept at a minimum
You've got that right--most of the examples I've seen are pretty...spartan.
Cray has never sold computers that are anything like a normal company would need. Cray machines are made for heavy number crunching - Vector processors are made for simulation tasks. They're very good at them. However they perform abyssmally at most other tasks - buying one for use as say, a database or application server would be stupid.
I don't recall saying that cray was trying to sell general business machines. But even for scientific applications, the number of customers who need a cray as opposed to being able to use a commodity cluster is much lower then the number who needed a cray instead of an IBM 360. There are businesses out there who use computers for more then spreadsheets and web servers. By "ordinary company" I meant to draw attention to that part of the market whose budget isn't classified.
Last time I checked Cray shipped UNICOS with their machines. It's a fairly BSDish UNIX variant. It's a bit of an oddball, but not all that much more of a PITA than say, IRIX or AIX.
I guess you didn't do much porting of mainstream applications to a cray. The lack of virtual memory, the funny type sizes in C, and other things that application writers make assumptions about (things that aren't technically guaranteed to work in ANSI C but do work on every other system in the world) could make porting a real problem. Things have gotten a lot better, but I can assure you that a unicos port of, say, perl or gcc was not in the same league as an irix port of the same app. One of the things cray is finally bowing to is the demand for virtual memory. Seymour never wanted it (didn't want the performance hit) but it's real hard to sell that in today's marketplace. The question is how much cray can back off of its old "speed is king" philosophy when their whole business is making fast computers.
Want to port your beowulf apps? No problem! When I spent a summer working on a T3E all of our multi processor apps used MPI.
You've kinda missed the boat. The point of the cutting-edge cray supercomputers isn't to run mpi apps--those do quite nicely on commodity clusters. The T3E is a MPP super--not a vector super. It's where cray was 10+ years ago, not where they want to be tomorrow. The point of cutting edge is to create new paradigms. That definately helps your performance, but it kills your compatibility.
Vectorization of C and FORTRAN apps is largely taken care of by the compiler.
Wow. Let's just say that when you're on the kind of project that can command the state of the art you don't depend on compiler autoparallelization.
So wheres all this programmer investment you're talking about? Most of the kinds of apps that you're going to run on a Cray (Weather models, crash simulations, Gaussian for chemical sims, etc) already run on a Cray,
Please, read up on the tera system, for example, and try to understand how it's different from a T3E.
No, you're a moron. unicos was not available on the first cray systems ("the heyday"). Cray released COS, long before unicos, but it wasn't even used by all of its customers. (Some just wrote their own OS.)
Yeah, your point? You said nothing about the reliability of one system versus another. There's a lot more that goes into designing a reliable system then spouting off some made-up statistics about cpu failures.
Our T3E was having problems well past the point where it was getting long in the tooth. Cray started adding functionality to make it more supportable a few years back, but when it was actually a cutting edge system it was pretty unstable. They probably couldn't widely sell a system today that had the problems of the earlier T3E's (one hardware problem and you need to reboot the whole thing) but that just increases the development costs and time to market in a market where delay means that the peasents will be nipping at your heels. Remember, by the time a super hits maturity, it's obsolete.
Probably not. Cray made some money back when a supercomputer was something that an ordinary company might need. The capabilities of "normal" computers was much more limited then today, so there was a much higher percentage of the buying public likely to want something more. These days the vast majority of users are happy with something mainstream
But, you ask, isn't there a lunatic fringe who wants more power at any price? Well, the lunatic fringe ain't what it used to be. During the heyday of cray you got a damn fine box and nothing else. Cray didn't want to worry about your software--or even an OS. A person who needed the speed would plunk down the money for the box and then pay a couple of guys to code everything from scratch. Those days are gone--software is the driving factor these days, and people are far less willing to buy something that's going to force a total code rewrite. Especially if that thing is only going to buy them a couple of years of edge before they need to recode for the next best thing.
Then there's the question of whether cray can afford to be bigger. The answer is "probably not". If you sell to a lot of customers you need a huge support infrastructure. Cray doesn't have much of one anymore, so they'd need to buy one. (Most of the old support guys left one way or another when SGI came in, or stayed with SGI.) If you have a lot of customers you can spread the costs around, but in the case of a company like cray a support infrastructure means having a people sitting around most of the time in every region you sell a machine. Maybe two to four guys per system (24x7, right?) plus some sorta warehouse facility if you enter a new geographical market. That's expensive. You can bill a lot of that cost back to the customers, but that just makes your systems less competetive.
I think the long term answer is that cray will be a very small niche player, selling to a very select group of (U.S.) government agencies, with the occasional pro forma business customer thrown in so the company can issue press releases. Even most government facilities aren't in a position to buy a cray anymore. (Research money is fairly tight, recoding costs are prohibative, MTBF's are more of an issue then they used to be, etc.)
MTBF: Mean time between failures. Commodity hardware goes kaputt much more often. A cluster capable of teraflop performance of custom hardware tends to need constant and evil levels of care and feeding: ie you better have a grad student on roller blades.
Hahahaha. Have you ever actually run a supercomputer? They tend to have much higher failure rates then normal servers. Couple of reasons: first, they push the envelope of a given technology. The sweet spot for stability is not the leading edge. Second, they're not nearly as well tested as mainstream hardware. On a platform with thousands of installations you're much less likely to run into a problem nobody has seen before than you are on a platform with only dozens of installations.
The dual Xeon 2.4Ghz you speak of, what are its other features? Firewire? USB2? Serial ATA? What video card? Apple sells a package, so you can't really compare it to that server setup.
Sure you can. The problem with "packages" is that the package is only good if it contains exactly the set of features that you need. It might well be that a machine used for computations doesn't need firewire, usb2, serial ata, or a video card. It probably does need networking (and both the apple and many xeons include gigabit, but you'd need to buy an add-in card for either if you wanted quad or fiber gigabit).
3 gigabit interfaces (2 in redundant failover mode, one for backups), 6 fiber channel disks on 2 diskplanes, and 2 qlogic HBA's to our EMC array. Each one of those cards has quite the ability to saturate a single PCI bus.
Actually, not one of those has the ability to saturate a single PCI bus, if you're using a current high-end PCI bus. A new 2 Gbit FC HBA should come with a 133MHz 64bit PCI interface, which can do more than 2Gbits throughput. Dual FC HBA's are starting to push things, but as long as they have their own bus should be mostly ok. 10Gbit ethernet is currently a challenge, but you didn't specify that.:)
Go out and get a firewire controller.
Right there you have almost 3 times the bandwidth of a SCSI controller, and 4 times the bandwidth of an ata100 IDE controller.
Moderators on crack. This is just plain wrong. Firewire is 400 or 800 Mbits/s, while SCSI is up to 320 Mbyte/s, IDE is up to 133 MByte/s, and Fibre Channel is up to 250 Mbyte/s. These numbers are directly comparable, because different buses have different amounts of overhead, but for sure firewire is a slow also-ran when talking only about performance. (When talking about cost, flexibility, etc., firewire looks better, of course.) As far as PCI goes, the top end is over 1 Gbyte/s, which is a bottleneck for some applications, but not firewire. Also, in high-end servers you'll have a number of pci buses to improve performance.
Re:110VAC outlets available today
on
42-Volt Autos
·
· Score: 1
Yeah, you can convert 12VDC to 120VAC, but you need one heck of a lot of amps running through your car's electrical system in order to do it. Switching to 42VDC means you have less chance of melting the wires in your car.
Why is it hypocritical to note that government agencies as a whole move slowly into new technologies, while individuals at goverment agencies sometimes introduce unauthorized elements? This is a very big problem because the cost of an access point is so low that it doesn't need special high-level approval (so it's hard for central authority to restrict such purchases)--and the security vulnerability introduced by such a cheap access point is very hard to mitigate.
I'm afraid you have it backwards. Bulk mail, even at its reduced rate, is what allows you to send a letter at 39 cents. Bulk mail is presorted so as to make processing time for the post office almost nothing. Your letter with sloppily written address actually takes time to be read and sorted.
You hear that a lot, but I don't know if it's completely true. If the mail volume were suddenly cut by 90% due to the elimination of junk mail, and the postal service could make an equivalent reduction in staff, trucks, air freight expenses, etc., would it really cost a lot more money for regular mail? What if we only got mail every couple of days instead of daily? The reality is that important/time critical stuff doesn't go via USPS first class mail anyway--it gets fedex or is otherwise overnighted. Personally, I only get a magazine and a couple of bills in a normal week, and the other 3 or 4 mails every day are junk. So it may be true that the bulk mail subsidizes the current postal service but it's an open question whether the current service is the best model.
Fair enough, but you can't blame Sun for being expensive, then criticise when low cost machines come in.
Sure I can. The new machines are attractive, but we've already jettisoned most of our suns--so the question isn't whether the new machines are nice, but whether they're nice enough to justify another platform change. The answer to that question is, "probably not". I still would like to get some and give them a chance, but it will be a hard sell to get me to tie myself back to sun at this point.
Also, the 280R's also never been 10 times the price of an Intel equivalent. Equally, there's no Intel box with the equivalent internal bandwidth or 64 bit capability which'll run your enterprise app quite as well.
I'm not sure I used the figure 10x, but the 280R is certainly a lot more than the entry level price for an x86 dual-processor system. Also remember that the V100/120 have a IIi processor, which made the 280R the real entry point for any sort of cpu-intensive application in sun's rackmount lineup. So for a cpu app you had a choice between a $1k or $2k intel rackmount or a 280R...
And the intel box would probably be a 1u, which is much nicer from a rack density standpoint compared to the 4u 280R. (The 280R takes a lot of space for what it's doing.) Can the 280R do some things a 1u intel box can't do? Sure--but the number of apps that need those feature is fairly limited. (If there were more demand for those features I'd have more 280R's, sun would be making more money than dell, and sun wouldn't have introduced the much smaller and better priced 210's and 240's.) Even from a capability standpoint there were some nutty things about the 280R that made it somewhat less attractive than it might have been. (Single 100Mbps ethernet, only 40MB/s SCSI included, only one pci slot running at more than 33MHZ, and that only at 66MHz, only 2 internal disk drives--in a 4u box!)
The 210s and 240s have great features - each actually has 4 built in Gb ethernet ports!
I stand corrected. I could have sworn the last datasheet only had 10/100 listed, but I might have been thinking of a different line.
Sorry, Slashdotters, but this perception that the equivalent Intel box is 10-15% of the price is utterly ludicrous. Have you seen the prices of the 210s, 240s and up?
Sorry, but how long have those boxes been around? Prior to the arrival of the 210's & 240's (less then a month ago, IIRC) the entry point for sun was the 100/120, which was an underpowered, overpriced heap. The next step up was a 280R, which was way too expensive for someone who just wanted a dual processor rackmount machine (or any machine not based on an obsolete chip like the IIi). The new 210/240's might be a nice system, but I haven't gotten any in-house to look at them yet. I am a little surprised at how weak the included features are (no built-in GBE!) but at least there's now a reasonable price point for an entry-level sun server that's competitive with an intel box. Probably too little too late, though. Maybe if they'd sold these two years ago...
But what if you configure the b & g with two different channels and two different ssids? You can create a seperate infrastructure for each type of network and in that case you should be able to use both at full speed.
It's plumbing and the corners will be to sharp for fiber. If you could even manage to get a fish threw it you'll never manage to get the fiber pulled into it with the fish. But strait sections no problem it's the bends that will kill you.
It's not indoor plumbing, it's a distribution system for a municiple steam system. The pipes are probably huge, not some little tiny things like you'd buy in a hardware store. The bend radius from outer wall to outer wall of the pipe, touching the inner curve of a bend, is probably not that tight.
Ok. Your fan can cool 1 Athlon, no problem. The question is how you'd cool, say, a fairly densly packaged system of 1500 Athlons in 10 racks. (8 procs per 2u.) That's 75kW. Liquid cooling is simply more practical once you start packing enough hot processors into a small space. (This sort of density isn't implausible, it's only a short evolution from what cray was achieving in the T3E 8 years ago--and they did it with liquid cooling.)
The real problem with stow is it doesn't do dependencies. It's fine for programs that use static binding
It's also fine for a network environment when you have control over the machine configs. As I mentioned elsewhere, think of it as a way of managing/usr/local hell. If you are responsible for admining a bunch of machines you probably want a standard config. ("All machines will have the following libraries installed:...") So the stuff you're stow managing doesn't have to worry about that. Where stow shines is keeping all those little local packages organized.
a developer of a program can have their ftp site added to the sources list and when the person runs software update it will install any new versions you release
Not only is that scary in a business environment, it's completely unrelated to the problems stow is trying to solve.
how long is it going to take for something on linux to do what installshield does on windows?
And what would that be? Guess what things are installed and try to remove them properly? This is a tool for a system manager to keep track of what's installed on a system--windows has no advantage whatsoever for that. (Multiple programs overwriting the same dll, registry entries spreading everywhere, lack of seperation of users in many programs, etc., mean that there's a lot of work to do in windows.)
I admit not knowing about linux, and am doing little to change that.
If you don't know and don't care, wouldn't it be easier to just not post?
Slashdot Mirror
Well, yes. You don't make that kind of investment (you're probably talking a bunch of cards, not just one to outfit a whole shop) without testing how it works in your environment. Hell, if you're going to buy enough of them you can probably get the vendors to loan you a test sample. To buy on benchmarks is just nuts.
Some != viable market. The government can (and does) keep them alive, but it won't make them huge or give them the kind of dominance in the industry they had 15 years ago. Even if they wanted to, cray's competitors (e.g., IBM) would pitch a fit on capitol hill. Consider what well funded agency will give them the kind of money you seem to think is out there. DoE bought something from them, but their big Red Storm deal is basically a commodity cluster. It gets cray cashflow but isn't revolutionary and isn't a special purpose vector super. The three letter agencies have money, but you have to wonder how much demand for vector supers they really have. (Crypto breaking is better on custom hardware, and echelon is a combination of highly parallelizable chunks and big databases.) Big vector supers are great for things like weather forecasting, but there's only so much money in that field (nothing like the national defense apps). So there's definately demand out there, but demand is nothing without funding.
You're reading too much (or the wrong things) into that deal. Neither company had much of a choice in the matter (tera needed revenue and cray needed to get away from sgi) and things were probably decided more by their government customers then by accountants.
You're dramatically overestimating the size of the market. Cray's own website puts it at about $1.1bn worldwide, and it's not like cray will get 100% market share. The ongoing R&D costs are a staggering percentage of their revenue, to the point that if the NSA wasn't subsidizing them it's unlikely they'd be alive today. The same goes for other pure supercomputing ventures--without huge amounts of government largess they're sunk.
Yes and no. The problem is that a cray box has to cover the whole R&D cost for an entire system. When IBM sells you an SP2 most of the R&D is spread across their much higher volume business lines. Same with an intel based cluster--the technology specific to the HPC market is basically the interconnect, and the rest is subsidized by video game players. There's also the compiler cost (you don't sell many fortran compilers outside the scientific market) but the salaries for a few compiler writers is much lower than the cost of desiging a cutting-edge cpu from scratch.
That's always true. The question is whether they can use the resources efficiently, and whether the cost/op is competetive. You're right about the algorithms being the driving force, but I'd argue that it is unusual for an algorithm that's optimized for one architecture to run optimally if you move it a radically different architecture. People can spend years trying to squeeze a couple more percent out of their code, and they don't want to start from scratch unless there's a very good reason. Then there's the problem that researchers tend to not work in a bubble. Even if you can afford to buy the most expensive machine on the block you might end up shooting yourself in the foot if nobody else in your field can collaborate with you.
You've got that right--most of the examples I've seen are pretty...spartan.
I don't recall saying that cray was trying to sell general business machines. But even for scientific applications, the number of customers who need a cray as opposed to being able to use a commodity cluster is much lower then the number who needed a cray instead of an IBM 360. There are businesses out there who use computers for more then spreadsheets and web servers. By "ordinary company" I meant to draw attention to that part of the market whose budget isn't classified.
I guess you didn't do much porting of mainstream applications to a cray. The lack of virtual memory, the funny type sizes in C, and other things that application writers make assumptions about (things that aren't technically guaranteed to work in ANSI C but do work on every other system in the world) could make porting a real problem. Things have gotten a lot better, but I can assure you that a unicos port of, say, perl or gcc was not in the same league as an irix port of the same app. One of the things cray is finally bowing to is the demand for virtual memory. Seymour never wanted it (didn't want the performance hit) but it's real hard to sell that in today's marketplace. The question is how much cray can back off of its old "speed is king" philosophy when their whole business is making fast computers.
You've kinda missed the boat. The point of the cutting-edge cray supercomputers isn't to run mpi apps--those do quite nicely on commodity clusters. The T3E is a MPP super--not a vector super. It's where cray was 10+ years ago, not where they want to be tomorrow. The point of cutting edge is to create new paradigms. That definately helps your performance, but it kills your compatibility.
Wow. Let's just say that when you're on the kind of project that can command the state of the art you don't depend on compiler autoparallelization.
Please, read up on the tera system, for example, and try to understand how it's different from a T3E.
No, you're a moron. unicos was not available on the first cray systems ("the heyday"). Cray released COS, long before unicos, but it wasn't even used by all of its customers. (Some just wrote their own OS.)
Yeah, your point? You said nothing about the reliability of one system versus another. There's a lot more that goes into designing a reliable system then spouting off some made-up statistics about cpu failures.
Our T3E was having problems well past the point where it was getting long in the tooth. Cray started adding functionality to make it more supportable a few years back, but when it was actually a cutting edge system it was pretty unstable. They probably couldn't widely sell a system today that had the problems of the earlier T3E's (one hardware problem and you need to reboot the whole thing) but that just increases the development costs and time to market in a market where delay means that the peasents will be nipping at your heels. Remember, by the time a super hits maturity, it's obsolete.
Probably not. Cray made some money back when a supercomputer was something that an ordinary company might need. The capabilities of "normal" computers was much more limited then today, so there was a much higher percentage of the buying public likely to want something more. These days the vast majority of users are happy with something mainstream
But, you ask, isn't there a lunatic fringe who wants more power at any price? Well, the lunatic fringe ain't what it used to be. During the heyday of cray you got a damn fine box and nothing else. Cray didn't want to worry about your software--or even an OS. A person who needed the speed would plunk down the money for the box and then pay a couple of guys to code everything from scratch. Those days are gone--software is the driving factor these days, and people are far less willing to buy something that's going to force a total code rewrite. Especially if that thing is only going to buy them a couple of years of edge before they need to recode for the next best thing.
Then there's the question of whether cray can afford to be bigger. The answer is "probably not". If you sell to a lot of customers you need a huge support infrastructure. Cray doesn't have much of one anymore, so they'd need to buy one. (Most of the old support guys left one way or another when SGI came in, or stayed with SGI.) If you have a lot of customers you can spread the costs around, but in the case of a company like cray a support infrastructure means having a people sitting around most of the time in every region you sell a machine. Maybe two to four guys per system (24x7, right?) plus some sorta warehouse facility if you enter a new geographical market. That's expensive. You can bill a lot of that cost back to the customers, but that just makes your systems less competetive.
I think the long term answer is that cray will be a very small niche player, selling to a very select group of (U.S.) government agencies, with the occasional pro forma business customer thrown in so the company can issue press releases. Even most government facilities aren't in a position to buy a cray anymore. (Research money is fairly tight, recoding costs are prohibative, MTBF's are more of an issue then they used to be, etc.)
Hahahaha. Have you ever actually run a supercomputer? They tend to have much higher failure rates then normal servers. Couple of reasons: first, they push the envelope of a given technology. The sweet spot for stability is not the leading edge. Second, they're not nearly as well tested as mainstream hardware. On a platform with thousands of installations you're much less likely to run into a problem nobody has seen before than you are on a platform with only dozens of installations.
I believe the term for SCO's actions is "barratry".
Sure you can. The problem with "packages" is that the package is only good if it contains exactly the set of features that you need. It might well be that a machine used for computations doesn't need firewire, usb2, serial ata, or a video card. It probably does need networking (and both the apple and many xeons include gigabit, but you'd need to buy an add-in card for either if you wanted quad or fiber gigabit).
Actually, not one of those has the ability to saturate a single PCI bus, if you're using a current high-end PCI bus. A new 2 Gbit FC HBA should come with a 133MHz 64bit PCI interface, which can do more than 2Gbits throughput. Dual FC HBA's are starting to push things, but as long as they have their own bus should be mostly ok. 10Gbit ethernet is currently a challenge, but you didn't specify that.
Moderators on crack. This is just plain wrong. Firewire is 400 or 800 Mbits/s, while SCSI is up to 320 Mbyte/s, IDE is up to 133 MByte/s, and Fibre Channel is up to 250 Mbyte/s. These numbers are directly comparable, because different buses have different amounts of overhead, but for sure firewire is a slow also-ran when talking only about performance. (When talking about cost, flexibility, etc., firewire looks better, of course.) As far as PCI goes, the top end is over 1 Gbyte/s, which is a bottleneck for some applications, but not firewire. Also, in high-end servers you'll have a number of pci buses to improve performance.
Yeah, you can convert 12VDC to 120VAC, but you need one heck of a lot of amps running through your car's electrical system in order to do it. Switching to 42VDC means you have less chance of melting the wires in your car.
Dude! The 20th century called and it wants its ATM back!
Why is it hypocritical to note that government agencies as a whole move slowly into new technologies, while individuals at goverment agencies sometimes introduce unauthorized elements? This is a very big problem because the cost of an access point is so low that it doesn't need special high-level approval (so it's hard for central authority to restrict such purchases)--and the security vulnerability introduced by such a cheap access point is very hard to mitigate.
You hear that a lot, but I don't know if it's completely true. If the mail volume were suddenly cut by 90% due to the elimination of junk mail, and the postal service could make an equivalent reduction in staff, trucks, air freight expenses, etc., would it really cost a lot more money for regular mail? What if we only got mail every couple of days instead of daily? The reality is that important/time critical stuff doesn't go via USPS first class mail anyway--it gets fedex or is otherwise overnighted. Personally, I only get a magazine and a couple of bills in a normal week, and the other 3 or 4 mails every day are junk. So it may be true that the bulk mail subsidizes the current postal service but it's an open question whether the current service is the best model.
Sure I can. The new machines are attractive, but we've already jettisoned most of our suns--so the question isn't whether the new machines are nice, but whether they're nice enough to justify another platform change. The answer to that question is, "probably not". I still would like to get some and give them a chance, but it will be a hard sell to get me to tie myself back to sun at this point.
I'm not sure I used the figure 10x, but the 280R is certainly a lot more than the entry level price for an x86 dual-processor system. Also remember that the V100/120 have a IIi processor, which made the 280R the real entry point for any sort of cpu-intensive application in sun's rackmount lineup. So for a cpu app you had a choice between a $1k or $2k intel rackmount or a 280R...
And the intel box would probably be a 1u, which is much nicer from a rack density standpoint compared to the 4u 280R. (The 280R takes a lot of space for what it's doing.) Can the 280R do some things a 1u intel box can't do? Sure--but the number of apps that need those feature is fairly limited. (If there were more demand for those features I'd have more 280R's, sun would be making more money than dell, and sun wouldn't have introduced the much smaller and better priced 210's and 240's.) Even from a capability standpoint there were some nutty things about the 280R that made it somewhat less attractive than it might have been. (Single 100Mbps ethernet, only 40MB/s SCSI included, only one pci slot running at more than 33MHZ, and that only at 66MHz, only 2 internal disk drives--in a 4u box!)
I stand corrected. I could have sworn the last datasheet only had 10/100 listed, but I might have been thinking of a different line.
Sorry, but how long have those boxes been around? Prior to the arrival of the 210's & 240's (less then a month ago, IIRC) the entry point for sun was the 100/120, which was an underpowered, overpriced heap. The next step up was a 280R, which was way too expensive for someone who just wanted a dual processor rackmount machine (or any machine not based on an obsolete chip like the IIi). The new 210/240's might be a nice system, but I haven't gotten any in-house to look at them yet. I am a little surprised at how weak the included features are (no built-in GBE!) but at least there's now a reasonable price point for an entry-level sun server that's competitive with an intel box. Probably too little too late, though. Maybe if they'd sold these two years ago...
But what if you configure the b & g with two different channels and two different ssids? You can create a seperate infrastructure for each type of network and in that case you should be able to use both at full speed.
It's not indoor plumbing, it's a distribution system for a municiple steam system. The pipes are probably huge, not some little tiny things like you'd buy in a hardware store. The bend radius from outer wall to outer wall of the pipe, touching the inner curve of a bend, is probably not that tight.
Ok. Your fan can cool 1 Athlon, no problem. The question is how you'd cool, say, a fairly densly packaged system of 1500 Athlons in 10 racks. (8 procs per 2u.) That's 75kW. Liquid cooling is simply more practical once you start packing enough hot processors into a small space. (This sort of density isn't implausible, it's only a short evolution from what cray was achieving in the T3E 8 years ago--and they did it with liquid cooling.)
It's also fine for a network environment when you have control over the machine configs. As I mentioned elsewhere, think of it as a way of managing
Not only is that scary in a business environment, it's completely unrelated to the problems stow is trying to solve.
And what would that be? Guess what things are installed and try to remove them properly? This is a tool for a system manager to keep track of what's installed on a system--windows has no advantage whatsoever for that. (Multiple programs overwriting the same dll, registry entries spreading everywhere, lack of seperation of users in many programs, etc., mean that there's a lot of work to do in windows.)
If you don't know and don't care, wouldn't it be easier to just not post?