That the actual end users explicitly seek out a piece of software, rather than settle for a defacto standard. Having users because they are too lazy to replace what the OEM gave them is, in my opinion, not as impressive.
As to the OEMs, there is the possibility of a kickback from MS from using IE exclusively, just like other 'free software', but I would think that would perhaps be too brazen considering the whole anti-trust thing.
Another possibility is a deep seated fear of distributing open-source software that seems to pervade these companies. Dell at least should be over it since it ships linux pre-installed, but then again, lawyers can insist that though the codebase is the same, they need to be paid to review different uses of it.
And finally, there is the possibility they believe it really not worth their time to bother. Would you *honestly* choose one brand over another *just* because of firefox being pre-installed, even if the firefox one is more expensive? It would take some work to migrate their crapware platform to make sure things work with firefox that would cost more than zero dollars. Meanwhile, the customers probably aren't looking for that explicitly, since they by and large can just download it truly for free themselves. Pre-install of linux is one thing, it gives an assurance about the hardware choices with respect to linux drivers, but Firefox is just a browser.
The testing methodology was flawed to draw any conclusions. The problem is the CPU may have been more active due to less IOWait states. AS a resulte, the drive consumption may be lower, and the benchmark was not throttled to the platter disk performance. The benchmark might have run many more times during the test.
Too many marketing organizations put more weight behind 'partners' who take them golfing or dining than what the customer actually wants. I find marketing as likely to overlook the actual end-users (paying or not) as engineers in favor of their own particular proclivities.
A significant part of the role of project management is externalizing what the engineers *should* have themselves, self-restraint. When you partner very bright, passionate engineers with no restraint and impose PMs on them to rein them in, the engineers morale is going to go down because part of the whole point of the PMs are to be as much of a 'killjoy' as it takes to offset the engineers' unconstrained enthusiasm.
In general, I think good engineers need a large amount of self-restraint to complement what passion they have. I'm personally involved in a PM-absent project in a typically very PM-heavy ecosystem. The engineers in this case understand and interact directly with the customers and have a lot of self-restraint. We have met our stated business dates more consistantly than our peer PM led efforts and are successfully managing maintenance branches without being prodded to do so by someone whose sole job is to prod. We picked a representative sample of engineers from various places to lead and people are mature enough to reach a consensus even if they disagree with the decision.
This may not be realistic for all, but I'd like to think it shows well-rounded, mature engineers are better than mixing all the extremes together and externalizing the balance of various aspects of people in general.
Or they'll know that an obscure gasket on 2003 Ford Taurus fails quite easily (I'm making this example entirely up),
You may make it up, but my experience says if its relatively recent ford, there's probably failing gaskets somewhere...
But I do suppose you are right, but it still seems so simple to double check your part number for oil filter, make sure you have the suggested viscosity, and not checking the dipstick... that's unimaginable to me... I use ramps instead of a lift so that's a tad different.
The thing is not about the 'dealer' generically overcharging. It's about Apple overcharging more than other vendors overcharging. All of them charge more for options for the general philosophy you hold justifying it, but overcharging more than a comparable competitor....
BTW, I did have the dealer change my oil during warranty because they sent me coupons for free oil changes for the duration of my warranty, but in the end, I find it hard to see how an oil change could break anything else, so I do it myself now that it is out of warranty.
I'm not understanding the hand waving about variations in the processor instruction stream. Isn't there an agreed upon measure for super computing and isn't it double floating point operations per second? "Rpeak" which is a useless theoretical number indeed pays no heed to what happens. "Rmax" is the most performace the LINPACK benchmark can extract. This means, for example, some arcthictures can do multiply and add concurrently. If you have floating point operations that happen to have those steps concurrent in the stream 90% of the time, then 90% of your instructions will be executed in half the cycles. However, if the algorithm doesn't lend itself to that, the feature won't be used. It's one example of very many as to why different applications aren't directly comparable.
And why would you assume the cell processor in the PS3 is so different from the Cell processor used in computing? From what I understand, the only difference is only 6 cores are available instead of 8. Prior to QS-22, it was the same. QS-22 uses what IBM calls a 'PowerXCell' variant. QS-20 and QS-21 both used the same processor with 8 enabled SPEs, and incidentally never made a Top500 showing. QS-22 is what is used in Roadrunner.
Your point about distributed vs local supercomputers is spot on.
On a side note, it is interesting to see people, without any real data, say something like folding@home can't be as powerful as a "real" supercomputer. Never mind there are a magnitude or two more modern cores being used in folding@home. I'm not implying it isn't a 'real' supercomputer, just that the results can't readily be compared and placed in that particular list. There are plenty of very powerful and very useful supercomputers that can't even approach a top500 score. Folding@home would be one of those, it couldn't possibly achieve a decent score due to the interconnect, which far from makes it worthless, since that is perfectly acceptable to their effort.
They no longer have a trusted third party, so you should validate the SSL certificate as you would an ssh known_hosts entry. One remembered, it affords the same degree and type of protection as known_hosts. A very easy and direct comparison.
Which Wiki article? The problem is *now* there are two varaints of 3.2 ghz cell processors. One used in the Petaflop system, and one in the PS3. The one in the petaflop run I heard *explicitly* sought to make the double-precision number repsectable.
And just because both folding and linpack are "scientific" computing, doesn't mean they have the same computational demands in terms of precision *nor* floating point efficiency. Even if they are the same, which I can't determine still, there can be a wide variation in the processor instruction stream and how a processor copes. It's because of this very complexity of HPC computing that a lot of people justifiably call the top500 benchmark a tad artificial. hpcc tries to cover a greater range, but doesn't have the same marketing clout that xhpl does.
Folding@home is a great endeavor, but it isn't so easy or simple to guess how it stacks up against this list.
"The performance of Windows HPC Server 2008 has yielded efficiencies that are among the highest we've seen for this class of machine," Pennington said. Ok, this is an interesting statement. In the most obvious interpretation, the reality is laid bare on the top500 list itself: linpack efficiency.
They had an Rmax of 68.48 out of an Rpeak of 89.59. That is a Linpack efficiency of 76.4%. With an inifiniband network, that's hardly interesting. It's hard to find a good comparison nearby (must be IB and Intel Core2). The nearest ones I see are #31 (81.5%), #11 (69.9%), and #10 (87%). 76% is about right smack in the midst of them. If comparing it to ethernet based installs, then yes, it is better (ethernet will easily get you down to the 50s), but that isn't particularly a constructive comparison. I guess he did say 'among' the best, meaning it doesn't suck as much as he would have thought, but for a more expensive product, it doesn't get you anywhere.
If speaking about utilization (odd choice of words if so), he isn't comparing linux to windows, but whatever job scheduler (maui, moab, pbs, etc) to MS solution. If that is the case and he is being genuine, I suspect he picked a bad scheduler for that.
The top500 score is a very particular benchmark. Whether it is the best measure of things is a matter of debate, but comparing numbers isn't a good idea. Essentially, if it can't/doesn't submit a top500, it isn't directly comparable to any of those scores in a meaningful fashion.
For example, with the PS3 clients figuring so prominently, my suspicion is that these are 32-bit floating point operations, and top500 only counts 64-bit floating point operations. I couldn't clarify it readily, so I might be wrong, but it seems that way at first glance.
For a lot of the fairly typical stuff, I actually am prepared to admit the base OS overhead may not be that different. A lot of HPC clusters are not set up particularly fundamentally different from a typical linux server randomly set up. This is mainly because it's just easier to understand and set up this way.
However, the ones that do implement something highly efficient or sophisticated at the OS level would have a very very hard time achieving analogous results. The petaflop system, for example a) uses cell processors and b) is mostly ram-resident in terms of OS. Neither one of those is Windows friendly (Windows PE can be run from ramroot, but the base platform is about two orders of magnitude larger than what I would call a 'base' linux platform). And officially, Windows PE is specifically designed to frustrate as a platform (timebombed reboots), because MS fears it being too widely used and cutting into their more traditionally licensed/extorted base.
On top of that, the general business scenario is mind-numbing. HPC is a market dominated by linux, for which there is a nicely competitive market of vendors. Red Hat, Novell, Canonical, and others all can offer a platform supporting linux apps and all have power to help. A cluster going for a monopolistic company providing a platform that isn't open to replacing the vendor, and a platform not particularly interesting in and of itself from a technical standpoint... Well, I just don't get it.
I agree with the bandwidth assessment, but there really are two usage experiences:
rdesktop: 'screen' like functionality. You have a self-contained workspace that can be disconnected and resumed. Useful, but keeps the apps very separate and provides the most jarring distinction between Windows and native apps.
-X11 style forwarding. Interleave Windows with other native applications. Notably, this includes system tray icons going to the right place and a single window list. For a cohesive experience, this is great.
Two problems with that: -No virtualization seems to be in shape to provide a good quality hardware 3d acceleration functionality. Among the most desired software, games are high on the list. Wine can play many of them accelerated, Virtualization can not.
-You are still required to buy a product from a monopolist. Well, to be legal at least. Part of the goal is to have a competing implementation of MS APIs, so that users are given a viable choice. Virtualization does not address this.
The best thing I think would be for the industry to move toward cross-platform toolkits (DirectX->SDL, Direct3D->OpenGL, MFC (or whatever it is)->wx (or any number of toolkits). They could still target Windows first, but be left with a portable codebase. Jumping on.Net because of mono is a braindead approach when you have so many other similar sorts of implementations that aren't driven by one platform vendor.
The assumption inherent in all these complaints is 2H2+O2=2H2O being the reaction. And while cheating, others have said the missing element in the scheme of things is Aluminum. Therefore, there is another 'fuel' that is not consumed as obviously quickly, so they probably slant that as a 'maintenance' action.
Burning isn't the best choice of words in retrospect. If they keep adding water, then it probably isn't 4H+O2=H2O, since then it could recycle the 'exhaust'. It's obviously something else, and others have done a good job of explaining better than I.
While misleading, the key is that water must be added to the system. *IF* the reaction took the hydrogen and consumed it in *any* sort of combustion or whatever, the product would be H2O. If the products are the same as the 'fuel', then why would it keep needing more.
As others have pointed out, there is another component here, a different reaction. The reactant is probably a solid, and probably (hopefully) the output is a solid. As such, 'water' is perceived as the analog to the lay man as gasoline (the stuff you have to frequently add, that causes nasty gaseous emissions), and the other reactant is a 'maintenance' thing, that is consumed at a fairly slow pace compare to the volume carried, and 'replaced' when it has been consumed.
As I said, this isn't a closed system. Less water comes out than goes in (why else would they need to add water so much?). Obvious, the hydrogen and oxygen are being used for something else than being made back into water. Obviously, there is something else going on, and some other byproduct. Other commenters have mentioned an Aluminum reaction that probably requires the Aluminum to be replaced ever so often.
Sorry, 'burning' was a poor choice of words, I should have said 'using'. As others have pointed out, this isn't any sort of reaction at the end that is hydrogen being combusted with oxygen or otherwise put back together with oxygen to form water. It's something else that undoubtedly requires more maintenance than 'pour water in and go', but I'm suspecting the non-water component needs replacing far less than you have to put in Gasoline.
But your logic I think is flawed. Hypothetically, they would use some process to start it, and then feed back in as it goes. Any typical car acts at a high level the same. To start extracting energy from gasoline, an electric motor starts the work, and then the fuel is consumed, mostly gone to heat, some used to move the car, and some reclaimed to recharge the battery.
In this case, it's describing sort of 'mining' hydrogen from the water. So it's not claiming a closed system is self sustaining, but that they burn hydrogen somehow in a way that yields more energy than goes into extracting it from the most stable source of it, water.
I'm not sure how this will actually pan out. As far as I know, separating hydrogen from oxygen has been considered expensive energy wise. But I don't think laws of thermodynamics are necessarily being violated here...
I have a harder time seeing red against earthy tones. Blue stands out like a sore thumb. I'm color blind, so that's why I have trouble quickly seeing a red player. That is something to keep in mind.
Slashdot Mirror
That the actual end users explicitly seek out a piece of software, rather than settle for a defacto standard. Having users because they are too lazy to replace what the OEM gave them is, in my opinion, not as impressive.
As to the OEMs, there is the possibility of a kickback from MS from using IE exclusively, just like other 'free software', but I would think that would perhaps be too brazen considering the whole anti-trust thing.
Another possibility is a deep seated fear of distributing open-source software that seems to pervade these companies. Dell at least should be over it since it ships linux pre-installed, but then again, lawyers can insist that though the codebase is the same, they need to be paid to review different uses of it.
And finally, there is the possibility they believe it really not worth their time to bother. Would you *honestly* choose one brand over another *just* because of firefox being pre-installed, even if the firefox one is more expensive? It would take some work to migrate their crapware platform to make sure things work with firefox that would cost more than zero dollars. Meanwhile, the customers probably aren't looking for that explicitly, since they by and large can just download it truly for free themselves. Pre-install of linux is one thing, it gives an assurance about the hardware choices with respect to linux drivers, but Firefox is just a browser.
The testing methodology was flawed to draw any conclusions. The problem is the CPU may have been more active due to less IOWait states. AS a resulte, the drive consumption may be lower, and the benchmark was not throttled to the platter disk performance. The benchmark might have run many more times during the test.
Too many marketing organizations put more weight behind 'partners' who take them golfing or dining than what the customer actually wants. I find marketing as likely to overlook the actual end-users (paying or not) as engineers in favor of their own particular proclivities.
A significant part of the role of project management is externalizing what the engineers *should* have themselves, self-restraint. When you partner very bright, passionate engineers with no restraint and impose PMs on them to rein them in, the engineers morale is going to go down because part of the whole point of the PMs are to be as much of a 'killjoy' as it takes to offset the engineers' unconstrained enthusiasm.
In general, I think good engineers need a large amount of self-restraint to complement what passion they have. I'm personally involved in a PM-absent project in a typically very PM-heavy ecosystem. The engineers in this case understand and interact directly with the customers and have a lot of self-restraint. We have met our stated business dates more consistantly than our peer PM led efforts and are successfully managing maintenance branches without being prodded to do so by someone whose sole job is to prod. We picked a representative sample of engineers from various places to lead and people are mature enough to reach a consensus even if they disagree with the decision.
This may not be realistic for all, but I'd like to think it shows well-rounded, mature engineers are better than mixing all the extremes together and externalizing the balance of various aspects of people in general.
Or they'll know that an obscure gasket on 2003 Ford Taurus fails quite easily (I'm making this example entirely up),
You may make it up, but my experience says if its relatively recent ford, there's probably failing gaskets somewhere...
But I do suppose you are right, but it still seems so simple to double check your part number for oil filter, make sure you have the suggested viscosity, and not checking the dipstick... that's unimaginable to me... I use ramps instead of a lift so that's a tad different.
Wow we can pull things offtopic.
The thing is not about the 'dealer' generically overcharging. It's about Apple overcharging more than other vendors overcharging. All of them charge more for options for the general philosophy you hold justifying it, but overcharging more than a comparable competitor....
BTW, I did have the dealer change my oil during warranty because they sent me coupons for free oil changes for the duration of my warranty, but in the end, I find it hard to see how an oil change could break anything else, so I do it myself now that it is out of warranty.
And, the key bit from that article on the one used in QS-22 vs. PS3:
dramatically improving double-precision floating-point performance on the SPEs from a peak of about 14 GFLOPS to 102 GFLOPS total for 8 SPEsSo yes, the PS3s cannot run linpack well at all, but the QS22 varaint can.
Your point about distributed vs local supercomputers is spot on.
On a side note, it is interesting to see people, without any real data, say something like folding@home can't be as powerful as a "real" supercomputer. Never mind there are a magnitude or two more modern cores being used in folding@home. I'm not implying it isn't a 'real' supercomputer, just that the results can't readily be compared and placed in that particular list. There are plenty of very powerful and very useful supercomputers that can't even approach a top500 score. Folding@home would be one of those, it couldn't possibly achieve a decent score due to the interconnect, which far from makes it worthless, since that is perfectly acceptable to their effort.I would help my Folding@Home score.
And at the same time, have it drive my desktop screensaver.
This is actually exactly what I did with a cluster I had for a few weeks with nothing better to do on it during the interval.
They no longer have a trusted third party, so you should validate the SSL certificate as you would an ssh known_hosts entry. One remembered, it affords the same degree and type of protection as known_hosts. A very easy and direct comparison.
Which Wiki article? The problem is *now* there are two varaints of 3.2 ghz cell processors. One used in the Petaflop system, and one in the PS3. The one in the petaflop run I heard *explicitly* sought to make the double-precision number repsectable.
And just because both folding and linpack are "scientific" computing, doesn't mean they have the same computational demands in terms of precision *nor* floating point efficiency. Even if they are the same, which I can't determine still, there can be a wide variation in the processor instruction stream and how a processor copes. It's because of this very complexity of HPC computing that a lot of people justifiably call the top500 benchmark a tad artificial. hpcc tries to cover a greater range, but doesn't have the same marketing clout that xhpl does.
Folding@home is a great endeavor, but it isn't so easy or simple to guess how it stacks up against this list.
They had an Rmax of 68.48 out of an Rpeak of 89.59. That is a Linpack efficiency of 76.4%. With an inifiniband network, that's hardly interesting. It's hard to find a good comparison nearby (must be IB and Intel Core2). The nearest ones I see are #31 (81.5%), #11 (69.9%), and #10 (87%). 76% is about right smack in the midst of them. If comparing it to ethernet based installs, then yes, it is better (ethernet will easily get you down to the 50s), but that isn't particularly a constructive comparison. I guess he did say 'among' the best, meaning it doesn't suck as much as he would have thought, but for a more expensive product, it doesn't get you anywhere.
If speaking about utilization (odd choice of words if so), he isn't comparing linux to windows, but whatever job scheduler (maui, moab, pbs, etc) to MS solution. If that is the case and he is being genuine, I suspect he picked a bad scheduler for that.
The top500 score is a very particular benchmark. Whether it is the best measure of things is a matter of debate, but comparing numbers isn't a good idea. Essentially, if it can't/doesn't submit a top500, it isn't directly comparable to any of those scores in a meaningful fashion.
For example, with the PS3 clients figuring so prominently, my suspicion is that these are 32-bit floating point operations, and top500 only counts 64-bit floating point operations. I couldn't clarify it readily, so I might be wrong, but it seems that way at first glance.
For a lot of the fairly typical stuff, I actually am prepared to admit the base OS overhead may not be that different. A lot of HPC clusters are not set up particularly fundamentally different from a typical linux server randomly set up. This is mainly because it's just easier to understand and set up this way.
However, the ones that do implement something highly efficient or sophisticated at the OS level would have a very very hard time achieving analogous results. The petaflop system, for example a) uses cell processors and b) is mostly ram-resident in terms of OS. Neither one of those is Windows friendly (Windows PE can be run from ramroot, but the base platform is about two orders of magnitude larger than what I would call a 'base' linux platform). And officially, Windows PE is specifically designed to frustrate as a platform (timebombed reboots), because MS fears it being too widely used and cutting into their more traditionally licensed/extorted base.
On top of that, the general business scenario is mind-numbing. HPC is a market dominated by linux, for which there is a nicely competitive market of vendors. Red Hat, Novell, Canonical, and others all can offer a platform supporting linux apps and all have power to help. A cluster going for a monopolistic company providing a platform that isn't open to replacing the vendor, and a platform not particularly interesting in and of itself from a technical standpoint... Well, I just don't get it.
Yes, just use remote desktop...
(had to be done for Luyseyal's benefit)
Clustering in the sense I think you are discussing is the HA-clustering stuff. HPC clustering is a tad different.
No, it really cannot. Since it isn't x86 based, it doesn't meet the minimum requirements for Vista.
I agree with the bandwidth assessment, but there really are two usage experiences:
rdesktop: 'screen' like functionality. You have a self-contained workspace that can be disconnected and resumed. Useful, but keeps the apps very separate and provides the most jarring distinction between Windows and native apps.
-X11 style forwarding. Interleave Windows with other native applications. Notably, this includes system tray icons going to the right place and a single window list. For a cohesive experience, this is great.
Two problems with that:
.Net because of mono is a braindead approach when you have so many other similar sorts of implementations that aren't driven by one platform vendor.
-No virtualization seems to be in shape to provide a good quality hardware 3d acceleration functionality. Among the most desired software, games are high on the list. Wine can play many of them accelerated, Virtualization can not.
-You are still required to buy a product from a monopolist. Well, to be legal at least. Part of the goal is to have a competing implementation of MS APIs, so that users are given a viable choice. Virtualization does not address this.
The best thing I think would be for the industry to move toward cross-platform toolkits (DirectX->SDL, Direct3D->OpenGL, MFC (or whatever it is)->wx (or any number of toolkits). They could still target Windows first, but be left with a portable codebase. Jumping on
The assumption inherent in all these complaints is 2H2+O2=2H2O being the reaction. And while cheating, others have said the missing element in the scheme of things is Aluminum. Therefore, there is another 'fuel' that is not consumed as obviously quickly, so they probably slant that as a 'maintenance' action.
Burning isn't the best choice of words in retrospect. If they keep adding water, then it probably isn't 4H+O2=H2O, since then it could recycle the 'exhaust'. It's obviously something else, and others have done a good job of explaining better than I.
While misleading, the key is that water must be added to the system. *IF* the reaction took the hydrogen and consumed it in *any* sort of combustion or whatever, the product would be H2O. If the products are the same as the 'fuel', then why would it keep needing more.
As others have pointed out, there is another component here, a different reaction. The reactant is probably a solid, and probably (hopefully) the output is a solid. As such, 'water' is perceived as the analog to the lay man as gasoline (the stuff you have to frequently add, that causes nasty gaseous emissions), and the other reactant is a 'maintenance' thing, that is consumed at a fairly slow pace compare to the volume carried, and 'replaced' when it has been consumed.
As I said, this isn't a closed system. Less water comes out than goes in (why else would they need to add water so much?). Obvious, the hydrogen and oxygen are being used for something else than being made back into water. Obviously, there is something else going on, and some other byproduct. Other commenters have mentioned an Aluminum reaction that probably requires the Aluminum to be replaced ever so often.
Sorry, 'burning' was a poor choice of words, I should have said 'using'. As others have pointed out, this isn't any sort of reaction at the end that is hydrogen being combusted with oxygen or otherwise put back together with oxygen to form water. It's something else that undoubtedly requires more maintenance than 'pour water in and go', but I'm suspecting the non-water component needs replacing far less than you have to put in Gasoline.
But your logic I think is flawed. Hypothetically, they would use some process to start it, and then feed back in as it goes. Any typical car acts at a high level the same. To start extracting energy from gasoline, an electric motor starts the work, and then the fuel is consumed, mostly gone to heat, some used to move the car, and some reclaimed to recharge the battery.
In this case, it's describing sort of 'mining' hydrogen from the water. So it's not claiming a closed system is self sustaining, but that they burn hydrogen somehow in a way that yields more energy than goes into extracting it from the most stable source of it, water.
I'm not sure how this will actually pan out. As far as I know, separating hydrogen from oxygen has been considered expensive energy wise. But I don't think laws of thermodynamics are necessarily being violated here...
I have a harder time seeing red against earthy tones. Blue stands out like a sore thumb. I'm color blind, so that's why I have trouble quickly seeing a red player. That is something to keep in mind.