No, they want Ethernet as a transport to contain a lot of the features of TCP so that they can lay their own protocols on top of it while being able to assume it's a reliable transport. That will increase the cost of ethernet by including that intelligence down the stack. Basically the cost of ethernet ports is plummeting compared to things like fiberchannel due to economies of scale and so cash strapped datacenters are trying to use it for everything, but it's not ideally designed to handle those other protocols so the other technology areas are trying to mold ethernet to meet their needs. Basically the way I see it if the industry does what is right there will be no 100Gbit Fiberchannel, there will be 100Gbit FCOE adapters.
It's also measuring the efficiency of the AC setup. Basically this is about reducing infrastructure waste in the datacenter. Getting the most MIPS/Watt from the actual equipment is a different metric which is also important, but better understood as it can be measured at the system level vs the datacenter level.
Since it's Cray it could very well be that physically it's multiple blades but they use an SSI custom setup to present all nodes as one machine. Oh and a bunch of pizza boxes would be way less efficient than blades and more expensive. At least that's the case when comparing HP DL360g5's with BL460c's, especially when you consider the PITA that it would be to run all those Infiniband connections on the pizza boxes vs using the internal fabric in the enclosure. The break even point for us is around 6/16 blades in an enclosure where it makes sense to go with the enclosure vs 1U's.
For general computing tasks x64 definitely offers the best MIPS/Watt/$. IBM has some interesting chips with the POWER line, but they don't sell enough volume or design for cheap enough fabrication to get good MIPS/$.
A shipping container is significantly larger than your average double dorm room so I don't think that's a problem. I would worry about environmental controls, but stacking a layer two deep of containers filled with earth around them should get rid of most of that problem =)
DI water tends to not stay that way in the presence of metals, and as soon as you have re-ionized water you have a big mess on your hands. Not to mention that water+Cap's=bad idea whereas cap's+oil=no biggie.
That explanation is SO stupid. The Quadro line is the same chips just binned and jumpered differently which is why the hard and soft-mods work, it's the same freaking hardware with a different driver. If you have a marginal card you might get some problems when exercising units that games don't normally push, but anything past a first rev of a new chip is likely to be just as good as the Quadro binning of the first rev.
Sorry but no sump is going to handle 5"/hour, that's worse than monsoon rainfall and reasonably priced (and powered!) pumps just aren't sized for that kind of freak event. I mean when Hurricane Ivan made landfall only a handful of places got 2"/hour.
Same here in Ohio, we "pay" ~$.09/KWhr, but the real price for residential customers is closer to $.17/KWhr. At my work it's was $.10 but it's gone up to almost $.11
Later 3.4 releases which exited beta before 4.0 used the new 4.0 ABI so as to move developers towards the new 4.0 ABI as far as I remember. This caused great pains for many Redhat users because they went with a late 3.4 build so almost no existing packages at the time were compatible.
Old Location Bar actually monkeys with the way the algorithm works to make it more like FF2 behavior. It's not completely successful, but good enough for me to go ahead using FF3. Before I found it most of my FF installs were stuck at the latest FF2. The retarded thing is you have to either register for an account or install it from the developers website because the Mozilla guys are so attached to the Awfulbar that they have it stuck at experimental even though it came out in June!
Exactly, when I saw the reports I was like what the f are they complaining about, the new layouts been available forever and most of my colleagues have been using it because it's simply better.
Only a fool would dare switching implementation mid-development,
Huh? All of my apps that run on platforms like JDBC and J2EE support the majority of implementations, ie I can choose to run on Apache/Tomcat or Websphere or Oracle's implementation without any issues with code or support.
None of this is about the physics. It's all plumbing and electrical work.
A LOT of physics went into designing and modeling the most powerful magnets ever made. Heck at full strength they will be about as strong as the strongest fields within the sun.
Ads can (and SHOULD) be turned off for google apps customers, it's an admin preference. Therefore the lack of gears for paying customers is nonsense. Now, the real answer is just use IMAP/POP for offline backup. I prefer IMAP because it retains folders, but both work fine.
Yes. In your organization how many times have your servers went down or had a problem... Compare that to Google Mail... You will probably find that there is a lot less downtime.
Sorry, but the total downtime I've ever caused ALL of my employers over my career has been a LOT less than 28 hours! Heck, even if you add up the downtime for all of the single systems I've admined their collective downtime is probably only close to that. I'm not bragging, I'm pointing out how bad of an outage this is. The only other outages I've personally heard of that were this bad are hosting providers who have critical systems physically damaged and a failed Exchange 2000 pilot at Cisco (They had a corrupted datastore that was so bad that MS and HP and EMC couldn't recover it so they had to fall back to a tape restore which took something similar to this gmail outage)
The Blackberry has done this since at least OS4 and the world hasn't come to an end. For instance if you install Google Maps Mobile by default it cannot access your GPS, your phonebook, make a call, or even access the web (kind of makes it hard for it to provide maps!). You can grant it any or all of these privileges individually.
No, as in your cellphone connects over WiFi and any new calls are completely free. There's no need for an ATA, no need for a different handset, no complicated call forwarding setup, you just connect to WiFi and you cellphone becomes a free VoIP phone with your normal number and cellphone features.
Slashdot Mirror
No, they want Ethernet as a transport to contain a lot of the features of TCP so that they can lay their own protocols on top of it while being able to assume it's a reliable transport. That will increase the cost of ethernet by including that intelligence down the stack. Basically the cost of ethernet ports is plummeting compared to things like fiberchannel due to economies of scale and so cash strapped datacenters are trying to use it for everything, but it's not ideally designed to handle those other protocols so the other technology areas are trying to mold ethernet to meet their needs. Basically the way I see it if the industry does what is right there will be no 100Gbit Fiberchannel, there will be 100Gbit FCOE adapters.
It's also measuring the efficiency of the AC setup. Basically this is about reducing infrastructure waste in the datacenter. Getting the most MIPS/Watt from the actual equipment is a different metric which is also important, but better understood as it can be measured at the system level vs the datacenter level.
Since it's Cray it could very well be that physically it's multiple blades but they use an SSI custom setup to present all nodes as one machine. Oh and a bunch of pizza boxes would be way less efficient than blades and more expensive. At least that's the case when comparing HP DL360g5's with BL460c's, especially when you consider the PITA that it would be to run all those Infiniband connections on the pizza boxes vs using the internal fabric in the enclosure. The break even point for us is around 6/16 blades in an enclosure where it makes sense to go with the enclosure vs 1U's.
For general computing tasks x64 definitely offers the best MIPS/Watt/$. IBM has some interesting chips with the POWER line, but they don't sell enough volume or design for cheap enough fabrication to get good MIPS/$.
A shipping container is significantly larger than your average double dorm room so I don't think that's a problem. I would worry about environmental controls, but stacking a layer two deep of containers filled with earth around them should get rid of most of that problem =)
HAHAHA, you're right their webmaster is an idiot, probably used IE6 to design the site and never tested in IE7 or FF.
DI water tends to not stay that way in the presence of metals, and as soon as you have re-ionized water you have a big mess on your hands. Not to mention that water+Cap's=bad idea whereas cap's+oil=no biggie.
That explanation is SO stupid. The Quadro line is the same chips just binned and jumpered differently which is why the hard and soft-mods work, it's the same freaking hardware with a different driver. If you have a marginal card you might get some problems when exercising units that games don't normally push, but anything past a first rev of a new chip is likely to be just as good as the Quadro binning of the first rev.
Be careful with the compressed air nozzle, I've seen shop air blow IC's off boards and vacuums suck them off.
Sorry but no sump is going to handle 5"/hour, that's worse than monsoon rainfall and reasonably priced (and powered!) pumps just aren't sized for that kind of freak event. I mean when Hurricane Ivan made landfall only a handful of places got 2"/hour.
They probably mean bullet resistant aka Lexan.
Same here in Ohio, we "pay" ~$.09/KWhr, but the real price for residential customers is closer to $.17/KWhr. At my work it's was $.10 but it's gone up to almost $.11
Later 3.4 releases which exited beta before 4.0 used the new 4.0 ABI so as to move developers towards the new 4.0 ABI as far as I remember. This caused great pains for many Redhat users because they went with a late 3.4 build so almost no existing packages at the time were compatible.
Old Location Bar actually monkeys with the way the algorithm works to make it more like FF2 behavior. It's not completely successful, but good enough for me to go ahead using FF3. Before I found it most of my FF installs were stuck at the latest FF2. The retarded thing is you have to either register for an account or install it from the developers website because the Mozilla guys are so attached to the Awfulbar that they have it stuck at experimental even though it came out in June!
I'm sure someone could write a little greasemonkey script that turns the sidebar into tabs, shouldn't be too difficult.
Exactly, when I saw the reports I was like what the f are they complaining about, the new layouts been available forever and most of my colleagues have been using it because it's simply better.
Only a fool would dare switching implementation mid-development,
Huh? All of my apps that run on platforms like JDBC and J2EE support the majority of implementations, ie I can choose to run on Apache/Tomcat or Websphere or Oracle's implementation without any issues with code or support.
Have you tried opera mini? On the Blackberry it's definitely better than the OS4 browser.
Um, you do realize that while it's a big number the US debt as a percentage of GDP is among the lowest in the industrialized world right?
None of this is about the physics. It's all plumbing and electrical work.
A LOT of physics went into designing and modeling the most powerful magnets ever made. Heck at full strength they will be about as strong as the strongest fields within the sun.
About a rounding error in the interest on the E2,000B the EU countries are spending bailing out their over-leveraged banks?
Ads can (and SHOULD) be turned off for google apps customers, it's an admin preference. Therefore the lack of gears for paying customers is nonsense. Now, the real answer is just use IMAP/POP for offline backup. I prefer IMAP because it retains folders, but both work fine.
Yes. In your organization how many times have your servers went down or had a problem... Compare that to Google Mail... You will probably find that there is a lot less downtime.
Sorry, but the total downtime I've ever caused ALL of my employers over my career has been a LOT less than 28 hours! Heck, even if you add up the downtime for all of the single systems I've admined their collective downtime is probably only close to that. I'm not bragging, I'm pointing out how bad of an outage this is. The only other outages I've personally heard of that were this bad are hosting providers who have critical systems physically damaged and a failed Exchange 2000 pilot at Cisco (They had a corrupted datastore that was so bad that MS and HP and EMC couldn't recover it so they had to fall back to a tape restore which took something similar to this gmail outage)
The Blackberry has done this since at least OS4 and the world hasn't come to an end. For instance if you install Google Maps Mobile by default it cannot access your GPS, your phonebook, make a call, or even access the web (kind of makes it hard for it to provide maps!). You can grant it any or all of these privileges individually.
No, as in your cellphone connects over WiFi and any new calls are completely free. There's no need for an ATA, no need for a different handset, no complicated call forwarding setup, you just connect to WiFi and you cellphone becomes a free VoIP phone with your normal number and cellphone features.