It's not just about the capacity, though the highest capacity chips generally have the lowest yield rates.
The new flash is cheaper, faster and higher capacity because yields are expected to be higher due to the smaller process (65nm). You get 496 gross die per wafer vs 450ish on 70nm.
The low chip count is due to the quad packaging used for flash in SSD. Each flash is quad stacked (okay dual-dual) and price wise is 4x+ more expensive.
Hmm...should be interesting to see how well the new MLC process stacks up to the SLC stuff they have.
The new flash is cheaper, faster and higher capacity because yields are expected to be higher due to the smaller process (65nm). You get 496 gross die per wafer vs 450ish on 70nm.
The low chip count is due to the quad packaging used for flash in SSD. Each flash is quad stacked (okay dual-dual) and price wise is 4x+ more expensive.
Hmm...should be interesting to see how well the new MLC process stacks up to the SLC stuff they have.
Anyone care to translate this? NoO offence, Vinea, you clearly have considerable knowledge but your geek-speak is gobbledy-gook to me!
Tiger, Leopard, and other apps DON"T come on a flash drive. Until they DO, we still NEED a DVD drive. How long will that be? I don't see that happening soon.
I got a "free" 1GB USB flash drive with my copy of Tax Cut. 4GB drives are $17. 8GB are a $28. A year ago these were probably $10-$15 or so retail. Not so bad for a $40-$50 software package.
I'd say Apple could go no optical in a year or so if they wanted to and provide a Blu-Ray UberDrive or current SuperDrive as an option. If they really wanted to, they COULD do so sooner but the cost would be higher.
They'd have to update iTunes to include regular software in addition to software for the iPods and iPhones. Then provide that service to 3rd party developers for a nominal fee.
Anyone care to translate this? NoO offence, Vinea, you clearly have considerable knowledge but your geek-speak is gobbledy-gook to me!
I know that the ML and SL in MLC and SLC stands for multi-level and single-level. So to me this seems like a denser flash technology (i.e. more flash in the same amount of space). When he talks about 65nm process, he's talking about the manufacturing process to make the chips, using a laser (I think) that is 65nm wide, or else 65nm refers to the size of the transisters, themselves that they are able to manufacture (I'm unsure off-hand).
This information was also posted to /. where some people in the comments mentioned that the MLC flash chips have unknown durability. As in, we know how many read/write cycles before failure on SLC chips, but since the MLC chips are so new we don't even know anything about how many times we can write to them before they fail. I would think that Samsung probably has estimates on this, but depending on how different the technology is it might require real world conditions before we know more. But of course, the tech is so new that there probably isn't readily available information on that kind of stuff to people outside of the industry (and even then, maybe only to certain people in a particular industry niche).
Although I don't think that Samsung would release a product to market that had an unknown or unpredictable MTBF (Mean Time Before Failure), you might be taking your chances in putting mission critical data on these drives right as the hit they market. (Though if you have a good enough backup scheme, you should be ok)
Anyone care to translate this? NoO offence, Vinea, you clearly have considerable knowledge but your geek-speak is gobbledy-gook to me!
Ah...the new process is smaller so you can cram more chips on one wafer. When there's a defect it screws up a smaller precentage of the total output so the yields are higher. At least for random defects on a wafer...
Yields are also usually lower on a new process as folks figure out how to fix new issues. For example jumping from 90nm to 65nm they had to deal with leakage and other design-related yield losses. You might get a 70% good wafer but only 35% actually meets your performance spec for power use due to leakage.
But Samsung is claiming lower costs off the bat. So I'm guessing they're getting pretty good yields on their 65nm process.
You bring my 5 MB Apple Profile Hard Drive to mind on my 256 KB RAM Apple /// ... Did that ever make my three floppy disk drive Apple ][ look lame!
Talking of future possibilities ... I can't wait to see if the Wi-Fi manufacturers get hold of the current standard TV spectrum when the FCC auction it off. I've read (but can't believe even though I'd like to) it could enable a five mile radius personal wi-fi range.
That 20 MB drive under my Mac+ made me feel like a King!
I've just ordered a MBA for summer library research work in Japan. I debated and debated the SSG, but finally went with the regular HD on the theory that if this thing proves out, I'd be back in the market when the miracle drives finally make the shelves.
Of course, the iPhoneAIR might be coming next month...
Comments
It's not just about the capacity, though the highest capacity chips generally have the lowest yield rates.
The new flash is cheaper, faster and higher capacity because yields are expected to be higher due to the smaller process (65nm). You get 496 gross die per wafer vs 450ish on 70nm.
The low chip count is due to the quad packaging used for flash in SSD. Each flash is quad stacked (okay dual-dual) and price wise is 4x+ more expensive.
Hmm...should be interesting to see how well the new MLC process stacks up to the SLC stuff they have.
The new flash is cheaper, faster and higher capacity because yields are expected to be higher due to the smaller process (65nm). You get 496 gross die per wafer vs 450ish on 70nm.
The low chip count is due to the quad packaging used for flash in SSD. Each flash is quad stacked (okay dual-dual) and price wise is 4x+ more expensive.
Hmm...should be interesting to see how well the new MLC process stacks up to the SLC stuff they have.
Anyone care to translate this? NoO offence, Vinea, you clearly have considerable knowledge but your geek-speak is gobbledy-gook to me!
Tiger, Leopard, and other apps DON"T come on a flash drive. Until they DO, we still NEED a DVD drive. How long will that be? I don't see that happening soon.
I got a "free" 1GB USB flash drive with my copy of Tax Cut. 4GB drives are $17. 8GB are a $28. A year ago these were probably $10-$15 or so retail. Not so bad for a $40-$50 software package.
I'd say Apple could go no optical in a year or so if they wanted to and provide a Blu-Ray UberDrive or current SuperDrive as an option. If they really wanted to, they COULD do so sooner but the cost would be higher.
They'd have to update iTunes to include regular software in addition to software for the iPods and iPhones. Then provide that service to 3rd party developers for a nominal fee.
Anyone care to translate this? NoO offence, Vinea, you clearly have considerable knowledge but your geek-speak is gobbledy-gook to me!
I know that the ML and SL in MLC and SLC stands for multi-level and single-level. So to me this seems like a denser flash technology (i.e. more flash in the same amount of space). When he talks about 65nm process, he's talking about the manufacturing process to make the chips, using a laser (I think) that is 65nm wide, or else 65nm refers to the size of the transisters, themselves that they are able to manufacture (I'm unsure off-hand).
This information was also posted to /. where some people in the comments mentioned that the MLC flash chips have unknown durability. As in, we know how many read/write cycles before failure on SLC chips, but since the MLC chips are so new we don't even know anything about how many times we can write to them before they fail. I would think that Samsung probably has estimates on this, but depending on how different the technology is it might require real world conditions before we know more. But of course, the tech is so new that there probably isn't readily available information on that kind of stuff to people outside of the industry (and even then, maybe only to certain people in a particular industry niche).
Although I don't think that Samsung would release a product to market that had an unknown or unpredictable MTBF (Mean Time Before Failure), you might be taking your chances in putting mission critical data on these drives right as the hit they market. (Though if you have a good enough backup scheme, you should be ok)
Anyone care to translate this? NoO offence, Vinea, you clearly have considerable knowledge but your geek-speak is gobbledy-gook to me!
Ah...the new process is smaller so you can cram more chips on one wafer. When there's a defect it screws up a smaller precentage of the total output so the yields are higher. At least for random defects on a wafer...
Yields are also usually lower on a new process as folks figure out how to fix new issues. For example jumping from 90nm to 65nm they had to deal with leakage and other design-related yield losses. You might get a 70% good wafer but only 35% actually meets your performance spec for power use due to leakage.
But Samsung is claiming lower costs off the bat. So I'm guessing they're getting pretty good yields on their 65nm process.
Soooo true.
You bring my 5 MB Apple Profile Hard Drive to mind on my 256 KB RAM Apple /// ... Did that ever make my three floppy disk drive Apple ][ look lame!
Talking of future possibilities ... I can't wait to see if the Wi-Fi manufacturers get hold of the current standard TV spectrum when the FCC auction it off. I've read (but can't believe even though I'd like to) it could enable a five mile radius personal wi-fi range.
That 20 MB drive under my Mac+ made me feel like a King!
I've just ordered a MBA for summer library research work in Japan. I debated and debated the SSG, but finally went with the regular HD on the theory that if this thing proves out, I'd be back in the market when the miracle drives finally make the shelves.
Of course, the iPhoneAIR might be coming next month...