OCZ drives have a very high failure rate, and those kinds of failures have nothing to do with write endurance. Of course, products from all brands fail, but you're a lot less likely to have problems with Intel or Samsung drives.
If you're using a consumer SSD for enterprise tasks, you shouldn't be surprised if it wears out more quickly than normal. That's why enterprise-grade drives tend to use SLC, or binned MLC with tons of spare area. If you're talking about consumer-level caching, the same is true there; Intel has a line of SSDs designed for caching, which use a small amount (24GB) of SLC flash in order to keep the endurance high enough. While it is virtually impossible for a consumer to wear out a good quality SSD using anything remotely like normal usage patterns, it's still important to use the best tool for the job.
That said, it's impossible to wear out a decent consumer SSD in a single month. If we take the 240GB Intel 330 drives as an example (which are reported to have a lifecycle of 10k, perhaps because Intel manufactures flash and picks the best bins for themselves), wearing out the drive in a single month (ignoring spare space to compensate for not taking write amplification into account) would require that you write uninterrupted 24/7 at about one gigabyte per second. Considering the drive can't write that fast, your one month figure is likely FUD.
Despite all the, errm, uninformed ladies and gentlemen responding with kneejerk reactions, smaller process sizes really do reduce the program/erase endurance of NAND flash. At 50nm, MLC was at about 10,000 cycles. 34nm took us down to about 5,000 cycles. 25nm took us down to 3,000 to 5,000 cycles, which is where we're at now. So, technically, we are reducing cost at the expense of reliability.
There are mitigating factors, however. Over the same timespan, SSD controllers have improved, substantially reducing write amplification, and capacities have increased, preventing the total *writable* lifespan of drives from decreasing.
As an example, a 60GB drive good for 10,000 cycles with a write amplification factor of 2.0 has a total theoretical write lifespan of 300 TB. On the other hand, a 120GB drive good for 3,000 cycles with a write amplification factor of 1.1 has a theoretical write lifespan of 327 TB. Despite having less than a third of the "reliability" (on a cell level), the drive can actually handle slightly MORE activity overall.
It will always be a balancing act between cost and reliability when it comes to SSDs. As compared to Single Level Cells (SLC), Multi Level Cells (MLC), used by all consumer drives, has a tenth the endurance, but half the cost (storing two bits per cell rather than one). Basically, SLCs store data by trying to differentiate between two voltage levels: high, low. MLC increases that to four states (high, medium-high, medium-low, low). The reduced endurance is because it becomes harder to differentiate the levels sooner. Triple Level Cells (TLC) is starting to show up, and this stores three bits per cell using eight states. It helps density, but once again, at the cost of endurance.
This might be a good time to point out that cell-level "reliability" has no real bearing on the reliability of the entire SSD. Reduced cycle endurance means your drive will wear out faster, but it will still take years to wear them out (if ever), and when they do, they don't lose data, they just stop being able to write. If you're having your SSDs just up and die on you out of the blue, that has nothing to do with the trend towards decreasing write endurance.
Ironic, then, that Apple wasn't willing to pay IBM for a custom design, which would have been a derivative of existing products, but then ended up designing their own CPU from scratch for the A6 (presumably they had nothing to base their designs on, unless ARM provides a reference design with the instruction set license).
The Ivy Bridge power savings aren't too bad. Comparing the 3770k to the 2600k, we see near-identical idle power usage, and an full load difference of 53w for the 3770k and 79.4w for the 2600k. If we factor in the relative performance of the two processors (using multi-threaded cinebench R11), we see the 3770k is 11% faster in terms of absolute performance.
I don't know how much power those chips are using in isolation of the rest of the system, but 11% faster performance at 26 watts lower power usage is pretty substantial. If we argued the CPUs used zero power when idle (which is probably not that far off with modern chips), we could argue that Ivy Bridge was 66% more power efficient than Sandy Bridge, a pretty huge difference.
The thing is, that's under full load. Consumers don't use their processors under full load constantly in the real world; processors spend most of their time in low power modes, so you'd never see these sorts of savings unless you were actually maxing out your processor 24/7 on battery or something.
And yet the most powerful supercomputer in Europe, which is built by IBM, uses... Intel processors. That said, I'll grant you that IBM has an edge in high-performance processor power consumption, but Apple didn't ditch IBM processors because they had a burning desire to switch the x86, they ditched IBM processors because IBM consistently failed to put out anything that could fit in the power envelope required by a laptop.
A more useful comparison might be to benchmark the performance per watt of Apple's A6 versus a comparable Qualcomm Krait; they're very similar architecturally, and use the same instruction set and fab process, but the A6 was laid out by hand while the Krait was done with automated tools.
But even if we were to agree that hand-designed chips didn't have a significant power advantage, Intel still has the full generation lead on process technology. A 22nm part is going to be substantially more power efficient than an identical chip made on a 32nm process... TSMC and Globalfoundries really need to solve that, or Intel is going to eat everybody else's lunch. 22nm is more than double the density of 32nm, even if that doesn't produce a 1:1 power saving...
POWER7+ isn't out yet, so I don't know how you can claim that it's the "highest performing processor right now". It's also a 190W TDP part, as far as I can tell, so if you're trying to compare performance-per-watt, it would have to be more than three times faster than an Ivy Bridge E3-1280V2 to beat it in performance-per-watt. About the only current advantage it would have is physical density, but that's usually not as important since it's not the limiting factor.
They do, but it suffers many of the same problems. It's not just AMD that suffers from being a full process behind Intel either, TSMC and other fabs making ARM chips are suffering from this too. It's one of the ways that Intel is able to start making inroads into mobile phones despite their chips being less efficient; they can get away with it because of their better fab process.
As for AMD's automated designing, they may have changed their behaviour, but it was what lead to the disaster that was Bulldozer:
That's true from a financial standpoint, but according to Anandtech, it takes 30 to 40 extra watts to beat Intel's performance. In many use cases, such as a desktop, 30 extra watts isn't going to matter. But anything mobile or that has heat dissipation constraints (like ultra small form factor or embedded things), it matters more than price.
There are two factors to this, as far as I can tell. The first is that AMD doesn't hand-design their chips anymore (Intel and Apple are the only people doing that, I believe), instead using high-level design tools. Easier to design, but less efficient. The other factor is that AMD is far behind Intel from a manufacturing technology perspective. They're at least a full process generation behind Intel (32nm versus 22nm), and Intel is already using multi-gate transistors (Intel announced they were starting work on them in 2002, AMD announced it in 2003, but only Intel is shipping them yet).
These various factors, among other things, all add up to really hurt AMD's power consumption. They can match Intel's performance, but only by throwing more power at the problem. This means that they can offer a good value to customers, but they can't offer good power consumption for mobile. That, more than anything, prevents them from being considered for big players like Apple.
I'd like to think that saving the lives of the pilots would have been more important, but yes, there would have been a large monetary savings too.
I'm not saying it's expensive, I'm just pointing out that the GP's comment about the airforce not doing anything for $100k isn't relevant, because they're not doing this for $100k.
They tried to renegotiate with Google. Apple wanted a few new things like turn-by-turn, and Google was asking for some stuff in exchange, like increased branding, that Apple wasn't willing to do. Unfortunately, I can't remember where the article I read this was. Anyhow, even if the timing for the contract renewal had worked out, they may not have been able to come to terms on the missing features. Things like turn-by-turn weren't needed (or rather weren't expected) in a smartphone mapping app in 2007, but by 2012 they were expected, and their absence in iOS was notable.
They're at 66.8% in the US, but much higher elsewhere. In Canada, for example, they're above 80% when you combine google.ca and google.com. I'm seeing 80% or more in the global stats I can find, but most of those are skewed one way or another, so it's really hard to get an accurate picture.
Market shares can differ significantly even in countries as close as Canada and the US. AIM was, at least a few years ago, the most popular IM network in the US, but had virtually no presence outside the US. Even in Canada, where AOL did have an active presence as an ISP, MSN Messenger was dominant.
This PDF is now out of date, but it gives you a nice look at the situation for IM market share in 2008:
Note how the situation in the US is not reflected in any other country, although some other countries show similar splits. I suspect the figures are quite different these days; I would expect Google Talk and Facebook Messenger to be much more popular these days, in Canada at least.
They have over 80% of the global search marketshare. That's what makes them a monopoly. There's nothing illegal with being a monopoly, the question is if a company is abusing that monopoly or not.
It's not like Intel platforms are completely devoid of any power saving features. Processors have power-saving features, drives can be low-power versions (they do make enterprise versions of them) or solid state, and in a cloud environment you can shut entire servers off when they're not needed to handle the load.
Brief summary: muscular dystrophy, no useful movement apart from neck and face, so he controls the mouse with his cheek and earlobe, and the keyboard with a pen in his mouth. He plays stuff like guildwars 2, LoL, etc, and he's way better at them than I am.
Somehow a publicly published idea about how to get people through airport security faster and easier is now a "secret plan for searching you at the airport" and comments equating it to all sorts of nasty things...
It's a public patent, and the goal of the thing is clearly the opposite of what everybody seems to be claiming...
My phone has 64GB of storage, it's not unusual for me to be going on a trip and wanting to copy over tens of gigs of stuff. Doing that over wifi (especially when 2.4GHz is painfully over-saturated and few phones support 5GHz 802.11n) can be rather slow.
The new dock connector will still allow you to charge your device from a standard USB charger just fine, thereby not making chargers obsolete, as was the intention.
Who said anything about powering the TV? I'm talking about powering the MHL to HDMI adapter. Some phones don't require this, some do. This is why most MHL to HDMI adapters have a socket for an external power source.
ARM can still beat Intel when it comes to power efficiency, but I think the Atom-based android smartphones have proven that power efficiency isn't THAT bad.
It's certainly unfortunate that all the mobile vendors went down this path, but it's not an unreasonable provision in this sort of license. That said, the Apple license seems to be from 2008, before miniDP was made part of the DP v1.2 spec. I'm having trouble finding any information on the licensing situation for the full DP spec.
Slashdot Mirror
OCZ drives have a very high failure rate, and those kinds of failures have nothing to do with write endurance. Of course, products from all brands fail, but you're a lot less likely to have problems with Intel or Samsung drives.
If you're using a consumer SSD for enterprise tasks, you shouldn't be surprised if it wears out more quickly than normal. That's why enterprise-grade drives tend to use SLC, or binned MLC with tons of spare area. If you're talking about consumer-level caching, the same is true there; Intel has a line of SSDs designed for caching, which use a small amount (24GB) of SLC flash in order to keep the endurance high enough. While it is virtually impossible for a consumer to wear out a good quality SSD using anything remotely like normal usage patterns, it's still important to use the best tool for the job.
That said, it's impossible to wear out a decent consumer SSD in a single month. If we take the 240GB Intel 330 drives as an example (which are reported to have a lifecycle of 10k, perhaps because Intel manufactures flash and picks the best bins for themselves), wearing out the drive in a single month (ignoring spare space to compensate for not taking write amplification into account) would require that you write uninterrupted 24/7 at about one gigabyte per second. Considering the drive can't write that fast, your one month figure is likely FUD.
Despite all the, errm, uninformed ladies and gentlemen responding with kneejerk reactions, smaller process sizes really do reduce the program/erase endurance of NAND flash. At 50nm, MLC was at about 10,000 cycles. 34nm took us down to about 5,000 cycles. 25nm took us down to 3,000 to 5,000 cycles, which is where we're at now. So, technically, we are reducing cost at the expense of reliability.
There are mitigating factors, however. Over the same timespan, SSD controllers have improved, substantially reducing write amplification, and capacities have increased, preventing the total *writable* lifespan of drives from decreasing.
As an example, a 60GB drive good for 10,000 cycles with a write amplification factor of 2.0 has a total theoretical write lifespan of 300 TB. On the other hand, a 120GB drive good for 3,000 cycles with a write amplification factor of 1.1 has a theoretical write lifespan of 327 TB. Despite having less than a third of the "reliability" (on a cell level), the drive can actually handle slightly MORE activity overall.
It will always be a balancing act between cost and reliability when it comes to SSDs. As compared to Single Level Cells (SLC), Multi Level Cells (MLC), used by all consumer drives, has a tenth the endurance, but half the cost (storing two bits per cell rather than one). Basically, SLCs store data by trying to differentiate between two voltage levels: high, low. MLC increases that to four states (high, medium-high, medium-low, low). The reduced endurance is because it becomes harder to differentiate the levels sooner. Triple Level Cells (TLC) is starting to show up, and this stores three bits per cell using eight states. It helps density, but once again, at the cost of endurance.
This might be a good time to point out that cell-level "reliability" has no real bearing on the reliability of the entire SSD. Reduced cycle endurance means your drive will wear out faster, but it will still take years to wear them out (if ever), and when they do, they don't lose data, they just stop being able to write. If you're having your SSDs just up and die on you out of the blue, that has nothing to do with the trend towards decreasing write endurance.
It was worth about $20 million, that's serious business.
Ironic, then, that Apple wasn't willing to pay IBM for a custom design, which would have been a derivative of existing products, but then ended up designing their own CPU from scratch for the A6 (presumably they had nothing to base their designs on, unless ARM provides a reference design with the instruction set license).
The Ivy Bridge power savings aren't too bad. Comparing the 3770k to the 2600k, we see near-identical idle power usage, and an full load difference of 53w for the 3770k and 79.4w for the 2600k. If we factor in the relative performance of the two processors (using multi-threaded cinebench R11), we see the 3770k is 11% faster in terms of absolute performance.
I don't know how much power those chips are using in isolation of the rest of the system, but 11% faster performance at 26 watts lower power usage is pretty substantial. If we argued the CPUs used zero power when idle (which is probably not that far off with modern chips), we could argue that Ivy Bridge was 66% more power efficient than Sandy Bridge, a pretty huge difference.
The thing is, that's under full load. Consumers don't use their processors under full load constantly in the real world; processors spend most of their time in low power modes, so you'd never see these sorts of savings unless you were actually maxing out your processor 24/7 on battery or something.
And yet the most powerful supercomputer in Europe, which is built by IBM, uses... Intel processors. That said, I'll grant you that IBM has an edge in high-performance processor power consumption, but Apple didn't ditch IBM processors because they had a burning desire to switch the x86, they ditched IBM processors because IBM consistently failed to put out anything that could fit in the power envelope required by a laptop.
A more useful comparison might be to benchmark the performance per watt of Apple's A6 versus a comparable Qualcomm Krait; they're very similar architecturally, and use the same instruction set and fab process, but the A6 was laid out by hand while the Krait was done with automated tools.
But even if we were to agree that hand-designed chips didn't have a significant power advantage, Intel still has the full generation lead on process technology. A 22nm part is going to be substantially more power efficient than an identical chip made on a 32nm process... TSMC and Globalfoundries really need to solve that, or Intel is going to eat everybody else's lunch. 22nm is more than double the density of 32nm, even if that doesn't produce a 1:1 power saving...
POWER7+ isn't out yet, so I don't know how you can claim that it's the "highest performing processor right now". It's also a 190W TDP part, as far as I can tell, so if you're trying to compare performance-per-watt, it would have to be more than three times faster than an Ivy Bridge E3-1280V2 to beat it in performance-per-watt. About the only current advantage it would have is physical density, but that's usually not as important since it's not the limiting factor.
They do, but it suffers many of the same problems. It's not just AMD that suffers from being a full process behind Intel either, TSMC and other fabs making ARM chips are suffering from this too. It's one of the ways that Intel is able to start making inroads into mobile phones despite their chips being less efficient; they can get away with it because of their better fab process.
As for AMD's automated designing, they may have changed their behaviour, but it was what lead to the disaster that was Bulldozer:
http://www.xbitlabs.com/news/cpu/display/20111013232215_Ex_AMD_Engineer_Explains_Bulldozer_Fiasco.html
That's true from a financial standpoint, but according to Anandtech, it takes 30 to 40 extra watts to beat Intel's performance. In many use cases, such as a desktop, 30 extra watts isn't going to matter. But anything mobile or that has heat dissipation constraints (like ultra small form factor or embedded things), it matters more than price.
There are two factors to this, as far as I can tell. The first is that AMD doesn't hand-design their chips anymore (Intel and Apple are the only people doing that, I believe), instead using high-level design tools. Easier to design, but less efficient. The other factor is that AMD is far behind Intel from a manufacturing technology perspective. They're at least a full process generation behind Intel (32nm versus 22nm), and Intel is already using multi-gate transistors (Intel announced they were starting work on them in 2002, AMD announced it in 2003, but only Intel is shipping them yet).
These various factors, among other things, all add up to really hurt AMD's power consumption. They can match Intel's performance, but only by throwing more power at the problem. This means that they can offer a good value to customers, but they can't offer good power consumption for mobile. That, more than anything, prevents them from being considered for big players like Apple.
I'd like to think that saving the lives of the pilots would have been more important, but yes, there would have been a large monetary savings too.
I'm not saying it's expensive, I'm just pointing out that the GP's comment about the airforce not doing anything for $100k isn't relevant, because they're not doing this for $100k.
They tried to renegotiate with Google. Apple wanted a few new things like turn-by-turn, and Google was asking for some stuff in exchange, like increased branding, that Apple wasn't willing to do. Unfortunately, I can't remember where the article I read this was. Anyhow, even if the timing for the contract renewal had worked out, they may not have been able to come to terms on the missing features. Things like turn-by-turn weren't needed (or rather weren't expected) in a smartphone mapping app in 2007, but by 2012 they were expected, and their absence in iOS was notable.
Per aircraft. They built 195 of them to date, so the program to fix the whole fleet would cost about $20 million.
It is blocked at my workplace (because it has "games" in the hostname), but any games.slashdot.org slashdot can also be accessed from slashdot.org.
They're at 66.8% in the US, but much higher elsewhere. In Canada, for example, they're above 80% when you combine google.ca and google.com. I'm seeing 80% or more in the global stats I can find, but most of those are skewed one way or another, so it's really hard to get an accurate picture.
Market shares can differ significantly even in countries as close as Canada and the US. AIM was, at least a few years ago, the most popular IM network in the US, but had virtually no presence outside the US. Even in Canada, where AOL did have an active presence as an ISP, MSN Messenger was dominant.
This PDF is now out of date, but it gives you a nice look at the situation for IM market share in 2008:
http://billionsconnected.com/blog/wp-content/uploads/2008/08/global_im_market_share_stats_july_08.pdf
Note how the situation in the US is not reflected in any other country, although some other countries show similar splits. I suspect the figures are quite different these days; I would expect Google Talk and Facebook Messenger to be much more popular these days, in Canada at least.
They have over 80% of the global search marketshare. That's what makes them a monopoly. There's nothing illegal with being a monopoly, the question is if a company is abusing that monopoly or not.
It's not like Intel platforms are completely devoid of any power saving features. Processors have power-saving features, drives can be low-power versions (they do make enterprise versions of them) or solid state, and in a cloud environment you can shut entire servers off when they're not needed to handle the load.
You wouldn't be the first, this guy already games with no hands:
http://www.twitch.tv/aieron/videos
Brief summary: muscular dystrophy, no useful movement apart from neck and face, so he controls the mouse with his cheek and earlobe, and the keyboard with a pen in his mouth. He plays stuff like guildwars 2, LoL, etc, and he's way better at them than I am.
Somehow a publicly published idea about how to get people through airport security faster and easier is now a "secret plan for searching you at the airport" and comments equating it to all sorts of nasty things...
It's a public patent, and the goal of the thing is clearly the opposite of what everybody seems to be claiming...
My phone has 64GB of storage, it's not unusual for me to be going on a trip and wanting to copy over tens of gigs of stuff. Doing that over wifi (especially when 2.4GHz is painfully over-saturated and few phones support 5GHz 802.11n) can be rather slow.
The new dock connector will still allow you to charge your device from a standard USB charger just fine, thereby not making chargers obsolete, as was the intention.
Who said anything about powering the TV? I'm talking about powering the MHL to HDMI adapter. Some phones don't require this, some do. This is why most MHL to HDMI adapters have a socket for an external power source.
ARM can still beat Intel when it comes to power efficiency, but I think the Atom-based android smartphones have proven that power efficiency isn't THAT bad.
No, what people care is what they can charge from. An iPhone charges from any standard USB charger.
It's certainly unfortunate that all the mobile vendors went down this path, but it's not an unreasonable provision in this sort of license. That said, the Apple license seems to be from 2008, before miniDP was made part of the DP v1.2 spec. I'm having trouble finding any information on the licensing situation for the full DP spec.
By your logic, charging my iPhone with an Amazon Kindle charger means that I have a full-sized USB connector somewhere on my iPhone.