Because they aren't useless, the huge discrepancy between controllers is one of the biggest issues with really fast SSD's. The fact that a several hundred dollar raid controller can only do half as many IOPS as the builtin ICH controller just shows how poorly many controllers are today. This has always been fine because you couldn't put enough physical disks behind them to show their flaws wrt performance but a single X-25e can perform as well as even dozens of fast (15k FC) HDD's.
They only got 31.7MB/s with the X-25e @4K random writes, that's MUCH slower than I've been able to get out of it. On my HP P400 I get ~75MB/s and on my HP workstation with builtin Intel chipset I get ~150MB/s. I would say it's their testing rig that was seriously holding the drive back and if they redid it with a better controller they would have come to a very different conclusion. Of course I have yet to find an enterprise class RAID controller that can keep up with a 2 drive RAID-10 of X-25e's so the absolute performance may be moot.
Have you run 3.1b3 yet? I'm running it with the portable wrapper right now and it is noticeably faster than 3.0.7 on my old Thinkpad T42. Heck I was even able to run most of the Chrome javascript demo's without much stuttering.
You're on serious crack, NA is a net exporter of food by a LARGE amount. The great plains are the worlds bread basket and California and Mexico produce vast quantities of fruits and vegetables. The only thing we import in any great quantity is fruits and vegetables from central America during the winter months as even the growing areas in Mexico are far enough north to have significant dropoff in productivity then.
ANY type of plant. Like grass. Seems like that if the conversion from cellulose to ethanol was efficient enough, there would be much better plants or trees to use than Jatropha. Maybe Jatropha was the best option when we were chemically converting cellulose or doing it less efficiently?
That's the point, Jatropha isn't a cellulose based method, it's an oil seed and it's about 40% oil by weight. I guess you could use this method on the fiberous parts of the plant but then you would deplete the soil that much faster, probably better to plow it under and have the next generation efficiently convert those nutrients into oil.
Isn't it pretty much a foregone conclusion that cellulose based ethanol makes no sense when compared to algae or Jatropha (or similar oil seed plants that can grow on non-arable land) which can be converted to biodiesel? Even if the yield per acre were similar (they're not) the process sure looks to be much more complicated and the MUCH lower energy density of ethanol means you are going to waste a lot more of the harvested energy in transporting the fuel.
All the ISP's do that and as I have told my friends and family repeatedly over the years, DON'T under any circumstances let the installer near your PC with that thing, it's not needed and can only lead to problems.
Well at least here in Cuyahoga County, Ohio we threw out the massively expensive and even more massively flawed Diebold systems and went with proven, reliable optical scan machines. I haven't heard anything about the board of elections trying to recoup some of the millions we spent on those things but I agree that they definitely should have sought compensation.
$55k for a battery powered NEON?!? You have GOT to be kidding me, that's the stupidest idea ever. I'm sorry but doing a 4 door Volt seems like a much better idea, and in fact the test mule for the Volt is a Malibu so Chevy could actually do it if they wanted to.
An auto plant is a LOT more than just a building, it is a complex system of interlinked parts. Modern plants are designed by machine (the CAM in CAD/CAM) to optimize workflow and fit all the necessary steps into the given footprint. The cost of simply changing an engine plant from one line to another can reach into the hundreds of millions and since Ford built River Rouge I don't believe anyone has done an integrated assembly plant, they are all cogs in the bigger machine that is an auto company with inputs coming from hundreds of suppliers and dozens of plants.
Oh, I guess if you just want to use a pond to do it you can, but that's not a useful way to produce algae oil, to do it in a practical manner you need much better sun exposure and a way to get more CO2 to the algae than simple surface diffusion allows. The answer is large vats of glass with injected CO2, but as I said that's relatively capital intensive.
Nah, hemp is fairly low yield per acre (about a half ton of seeds that contain only 30% oil) and requires much better soil than Jatropha. Cellulose based ethanol from hemp is a non-starter as well.
There is no such thing as a "proper fuel crop" except algae, the only feedstock crop which is not topsoil-based, and thus the only one we should be using.
While I agree in general (not using food as fuel) I also have to point out that Jatropha is another good candidate for fuel production. Jatropha grows in very poor soil with very little water needed and produces seeds which are 1/3rd oil. I'm not sure what huge kind of acerage you would need to supply world energy demands but not every solution has to do it all. Algae is great if you have the water and the infrastructure to support that kind of production, but it's definitely possible that poorer places may need a different form of production which is less capital intensive.
Hmmm, perhaps this will make AT&T's Uverse offering useful after all depending on what kind of throughput it can achieve over the distances to the average remote shelf. I still wish they would roll things out like Verizon did with true fiber to the home.
Gah, their FC switches suck balls, and I have four back to back days with 4 hours or less sleep to show for it. Once we ripped out the Cisco POS's and replaced them with Brocades we are back to months without a peep out of our SAN. I will NEVER again buy Cisco FC switches and have made it my mission to tell everyone I can about them. Cisco massively oversubscribes almost all of the FC switches and even the ones that aren't have insufficient buffer 2 buffer pools in the ASIC to support real world full speed transfers. Heck our 9140's fell over at ~1.5Gbps total throughput, pretty sad for a switch which has 8 'dedicated' 2Gbps ports and another 32 shared ports.
It's WAY faster than the equivalent $ chips today. Two Xeon 5570's get 25,000 SAPS vs FOUR Xeon 7460's (that cost twice as much each) getting 25,830. They really are that much better, and I can't wait till HP starts shipping the DL380 G6 at the end of this month because I have a bunch of projects that can really use that kind of compute density and performance per core (damn Oracle licensing).
Hehe, not likely unless you want to pay Cisco premiums for both the hardware AND the maintenance. I'll stick to my HP C-class blade enclosures TYVM. I get 3 years of 6 hour call to repair service for about 10% of the cost of the blade center, Cisco costs more than 10% per YEAR for SmartNet and it isn't nearly as good of service IMHO.
The old location bar addon is MUCH closer to FF2 style bar wrt functionality as well as looks but since the dev's don't like the idea of people reverting the functionality they have left it in "experimental" status since June of last year forcing you to signup for a (free) account.
The spec is a set of necessary conditions but for many people it will be a sufficient set, they expect a filesystem to be as bulletproof as possible in every situation.
Slashdot Mirror
Because they aren't useless, the huge discrepancy between controllers is one of the biggest issues with really fast SSD's. The fact that a several hundred dollar raid controller can only do half as many IOPS as the builtin ICH controller just shows how poorly many controllers are today. This has always been fine because you couldn't put enough physical disks behind them to show their flaws wrt performance but a single X-25e can perform as well as even dozens of fast (15k FC) HDD's.
100% random writes using IOMeter.
Sorry, HP calls RAID-1 and RAID-10 both RAID-10 in the setup screen for their controllers.
They only got 31.7MB/s with the X-25e @4K random writes, that's MUCH slower than I've been able to get out of it. On my HP P400 I get ~75MB/s and on my HP workstation with builtin Intel chipset I get ~150MB/s. I would say it's their testing rig that was seriously holding the drive back and if they redid it with a better controller they would have come to a very different conclusion. Of course I have yet to find an enterprise class RAID controller that can keep up with a 2 drive RAID-10 of X-25e's so the absolute performance may be moot.
Have you run 3.1b3 yet? I'm running it with the portable wrapper right now and it is noticeably faster than 3.0.7 on my old Thinkpad T42. Heck I was even able to run most of the Chrome javascript demo's without much stuttering.
Accelerators and Web Slices both sound like they are big gaping security holes waiting to be exploited.
Put the SAN in a more controlled environment, you spend a little more on cabling but that's probably peanuts compared to the energy savings.
You're on serious crack, NA is a net exporter of food by a LARGE amount. The great plains are the worlds bread basket and California and Mexico produce vast quantities of fruits and vegetables. The only thing we import in any great quantity is fruits and vegetables from central America during the winter months as even the growing areas in Mexico are far enough north to have significant dropoff in productivity then.
ANY type of plant. Like grass. Seems like that if the conversion from cellulose to ethanol was efficient enough, there would be much better plants or trees to use than Jatropha. Maybe Jatropha was the best option when we were chemically converting cellulose or doing it less efficiently?
That's the point, Jatropha isn't a cellulose based method, it's an oil seed and it's about 40% oil by weight. I guess you could use this method on the fiberous parts of the plant but then you would deplete the soil that much faster, probably better to plow it under and have the next generation efficiently convert those nutrients into oil.
Isn't it pretty much a foregone conclusion that cellulose based ethanol makes no sense when compared to algae or Jatropha (or similar oil seed plants that can grow on non-arable land) which can be converted to biodiesel? Even if the yield per acre were similar (they're not) the process sure looks to be much more complicated and the MUCH lower energy density of ethanol means you are going to waste a lot more of the harvested energy in transporting the fuel.
All the ISP's do that and as I have told my friends and family repeatedly over the years, DON'T under any circumstances let the installer near your PC with that thing, it's not needed and can only lead to problems.
Well at least here in Cuyahoga County, Ohio we threw out the massively expensive and even more massively flawed Diebold systems and went with proven, reliable optical scan machines. I haven't heard anything about the board of elections trying to recoup some of the millions we spent on those things but I agree that they definitely should have sought compensation.
$55k for a battery powered NEON?!? You have GOT to be kidding me, that's the stupidest idea ever. I'm sorry but doing a 4 door Volt seems like a much better idea, and in fact the test mule for the Volt is a Malibu so Chevy could actually do it if they wanted to.
An auto plant is a LOT more than just a building, it is a complex system of interlinked parts. Modern plants are designed by machine (the CAM in CAD/CAM) to optimize workflow and fit all the necessary steps into the given footprint. The cost of simply changing an engine plant from one line to another can reach into the hundreds of millions and since Ford built River Rouge I don't believe anyone has done an integrated assembly plant, they are all cogs in the bigger machine that is an auto company with inputs coming from hundreds of suppliers and dozens of plants.
Oh, I guess if you just want to use a pond to do it you can, but that's not a useful way to produce algae oil, to do it in a practical manner you need much better sun exposure and a way to get more CO2 to the algae than simple surface diffusion allows. The answer is large vats of glass with injected CO2, but as I said that's relatively capital intensive.
Nah, hemp is fairly low yield per acre (about a half ton of seeds that contain only 30% oil) and requires much better soil than Jatropha. Cellulose based ethanol from hemp is a non-starter as well.
Haven't seen the Notes 8 client have you? Yeah it's built on the Eclipse framework so Notes and Java are already quite friendly.
There is no such thing as a "proper fuel crop" except algae, the only feedstock crop which is not topsoil-based, and thus the only one we should be using.
While I agree in general (not using food as fuel) I also have to point out that Jatropha is another good candidate for fuel production. Jatropha grows in very poor soil with very little water needed and produces seeds which are 1/3rd oil. I'm not sure what huge kind of acerage you would need to supply world energy demands but not every solution has to do it all. Algae is great if you have the water and the infrastructure to support that kind of production, but it's definitely possible that poorer places may need a different form of production which is less capital intensive.
Hmmm, perhaps this will make AT&T's Uverse offering useful after all depending on what kind of throughput it can achieve over the distances to the average remote shelf. I still wish they would roll things out like Verizon did with true fiber to the home.
The HP blade solution is superior because even if you want Cisco switches they are available for the C-class enclosures.
Gah, their FC switches suck balls, and I have four back to back days with 4 hours or less sleep to show for it. Once we ripped out the Cisco POS's and replaced them with Brocades we are back to months without a peep out of our SAN. I will NEVER again buy Cisco FC switches and have made it my mission to tell everyone I can about them. Cisco massively oversubscribes almost all of the FC switches and even the ones that aren't have insufficient buffer 2 buffer pools in the ASIC to support real world full speed transfers. Heck our 9140's fell over at ~1.5Gbps total throughput, pretty sad for a switch which has 8 'dedicated' 2Gbps ports and another 32 shared ports.
It's WAY faster than the equivalent $ chips today. Two Xeon 5570's get 25,000 SAPS vs FOUR Xeon 7460's (that cost twice as much each) getting 25,830. They really are that much better, and I can't wait till HP starts shipping the DL380 G6 at the end of this month because I have a bunch of projects that can really use that kind of compute density and performance per core (damn Oracle licensing).
Hehe, not likely unless you want to pay Cisco premiums for both the hardware AND the maintenance. I'll stick to my HP C-class blade enclosures TYVM. I get 3 years of 6 hour call to repair service for about 10% of the cost of the blade center, Cisco costs more than 10% per YEAR for SmartNet and it isn't nearly as good of service IMHO.
The old location bar addon is MUCH closer to FF2 style bar wrt functionality as well as looks but since the dev's don't like the idea of people reverting the functionality they have left it in "experimental" status since June of last year forcing you to signup for a (free) account.
The spec is a set of necessary conditions but for many people it will be a sufficient set, they expect a filesystem to be as bulletproof as possible in every situation.