When you come down to it, it's amazing how many of the problems that actually require large amounts of processor time _are_ embarrassingly parallel. Presumably because the more calculations you have to perform, the more likely those calculations are to be partitionable into entirely independent data sets.
How long have you had your domain? I've had mine for 10 years now, and I get a really weird combination of addresses. They've built up slowly over time. Some of them are pretty bizarre and totally unrelated to any address I've ever used. Some appear to derive from corrupted address lists that have been copied over and over (my normal address is myname@mydomain, I regular receive stuff to: mynamemyname@mydomain, myname.mydomain@mydomain, yname@mydomain, etc.) Some appear to be guesses of address I might likely use! (I used to contribute to the 'nasm' open source project, and I regularly get spam to 'nasm.source@mydomain' even though I have never used this address.) Still others are other peoples names @mydomain. I regularly get "brewster43@mydomain".
"I create a unique email address for each company I deal with, and each website I register on."
Does nobody of you morons know of mailinator.com?
Why on earth would someone create a mailaddress just to register to a website when mailinator with their gazillion aliases exists?
Just give them mythrowawaylogin@mailinator.com as email address, read it _once_ to click the confirmation link and forget it.
Reason 1: there are plenty of people using services like this - http://www.block-disposable-email.com/cms/ Reason 2: I may want to establish an ongoing relationship with a company (e.g. receive newsletters, etc) rather than just have a fire & forget initial contact Reason 3: Having email coming to my inbox is more convenient than having to open a web site to view it. (I have a regexp-based email setup that allows me to just make up addresses that match a pattern, and I can add individual addresses to my spam filter if they become compromised, so it's actually easier than using mailinator).
The OP also asks for a ROM but never tells us the modell of tablet (except that it is 7'' and from MID, hope they have only one of those).
Unfortunately, this just betrays the fact that he doesn't know what he's talking about: MID isn't a manufacturer at all, but an abbreviation that several of the cheaper manufacturers use to identify a "mobile internet device", i.e. a tablet.
The Android-derived tablet he got is horrible! The battery is the least of its problems.
As a rule of thumb, if someone offers you a tablet for Christmas with resistive touch, you shouldn't even open the box and you should try to return it for a full/partial refund as soon as possible. As to the security issues, the article he pointed to talks about apps being "security risks" or "malware" for requesting GPS permissions when they shouldn't (which is really FUD). In any case, since his tablet doesn't have a GPS chip in it, that issue doesn't apply to him.
Also since he doesn't have access to the official Android Market/Google Play, he should just look on the XDA forums, root his device, install Cyanogen on it, and go through the Cyanogen repo for apps. And he should refrain from installing apps from other locations
You're making an awful lot of assumptions about his tablet when he's told us nothing useful about it other than the price. FWIW, I suspect you're wrong on 2 of the 3 conclusions you draw here: it's perfectly possible to get an $89 tablet that has capacitative touch and Google Play these days, but GPS will likely be out of the question.
But your suggested solution highlights the real problem with buying cheap-ass chinese hardware: there will likely not be an open source build available for it. There's a small chance that this build will work for him, but it's quite unlikely. Probably, nobody even knows what hardware he has in there. So he can't put cyanogenmod, or any of the "kick-ass ROMs" he's asking for, on it. He's stuck with the slightly-dubious build of AOSP the manufacturer came up with. And for his information, "MID" is not the manufacturer -- it's an abbreviation for "mobile internet device" that's used by several manufacturers who prefer not to put their own branding on the devices they make.
if we compare ANY console to any "high end rig" it looks outdated usually before it even comes out by that logic
Judging by benchmarks of bulldozer chips (I've never actually tested one myself), this thing will struggle to outperform the Phenom-II X6 system I put together last year for under £400, which is a long way from being a high end rig. Sure, it has more cores, and is running a more efficient architecture (clock-for-clock benchmarks put Bulldozer about 20% faster than the K10 cores that are in Phenom IIs), but it's clocked *much* slower.
I'll grant its GPU may be better, but I suspect if I'd shelled out another £100 on my graphics card I'd see comparable performance. Haven't looked at the benchmarks, though, so I'm just guessing there.
The resolution for TVs will still be 1080p so its not like theres going to be much need for more GPU horsepower and anything more than 30fps is wasted on a TV anyway.
GPU power increases are not only used to driver higher resolutions or framerates; they are also used to increase model detail level, add additional effects, and otherwise improve the quality of the graphics. Also, many modern TVs are able to display 50 and/or 60 fps quite comfortably, so suggesting faster than 30fps is wasted on a TV seems illogical -- no more so than it would be wasted on a monitor.
When the average teachers in Chicago are making ~75k / yr with incredible benefits
Citation to a credible source needed. The only place I see figures of $75K are news articles quoting a biased source. Unbiased sources (e.g. the various salary surveys) are reporting $55K or thereabouts.
Let's face it --- this wouldn't be the first time an employer has inflated claims about how much he's paying in an attempt to discredit unions negotiating for a better deal.
Not in any context that's relevant to the discussion you're posting to. It has been talking about America's shortcomings in response to an article about France's purported shortcomings.
The company's head office is in Canada Square, London. It is listed on the LSE. It may originally have been based in Hong Kong, but it moved because regulations would have prevented them taking over the Midland Bank if they weren't British.
Apparently I forgot to include the tag. I guess a reference to a really old DOS based golf game and a really old text based browser weren't enough of a clue.
The most recent version of the former was released in 2004, the latter in 2010. Neither of these qualify as "really old" IMO.
Whereas in the good, honest Blue States of America, we just threaten hackers for political gain until they commit suicide. Much less expensive.
Probably somewhat influenced by the fact that the democrat AG only had a 4% lead on the republican candidate in the last election -- must be trying to impress the borderline voters.
Analysis cost is about $20k per genome for good enough coverage to distinguish CNV (and I work for a DNA sequencing company).
Really? My understanding was that whole genome sequencing was now available for $6k (as long as you're buying tests in bulk). Is there some reason a typical commercial whole-genome sequence would not be suitable for this purpose?
To catch those mutations, they won't be analyzing signatures, I don't think. A signature only works if you're sure that's where the mutations are. But they won't know any of that. They need to sequence the whole damn genome. That's around $50k a pop, right?
There seems to be an assumption that the innocent twin has CONVINCING evidence of his brother's guilt, and/or his own innocence. I might have a butt-ton load of evidence that I'm happy to share, but none of it is convincing. What then?
Well, thankfully the prosecutor has to share convincing evidence that you are guilty, which seems to be lacking in this case.
I'm pretty sure you haven't actually seen that. In practice, what the HDD industry as a whole does is:
Well, actually I have. I got the company wrong, though -- it was Maxtor who used this definition. Quoting from the manual above:
KILOBYTE (K) - A unit of measure consisting of 1,024 (2^10) bytes [...] MEGABYTE (MB) - A unit of measurement consisting of 1,000 kilobytes or 1,024,000 bytes
Renault has a rather good safety record compared to other cars in the same class and price range and this is not how a "normal" Renault Laguna would handle.
"Consider this: Once you've put progress on a bar, you can't take it off. Suppose you start a process that should take 20 minutes, and do the first 5 minutes, progress is now at 25%. But then, partway through, something unexpected happens and you realize the process is actually going to take 40 minutes. You can't take the progress "back" now, that would disorient the user. So you have to rescale the remainder of the bar."
But that's not what this person's app did. Whoever wrote their progress bar did it wrong.
The problem with many of these progress indicators is that they only calculate the estimated time once. Then if the app gets stuck or slows down at, say, 75%, it might say "about one minute" for 10 minutes. But that should never happen.
Instead, estimated remaining time -- when it is appropriate, which is a lot less often than it is actually used -- should be updated right along with the % progress. So yeah, as you say, if done properly, it can start out at 5 minutes and creep up to 20 or whatever. But it should not just freeze at some time and stay there a while. If it's going to do that, the app should probably not be trying to estimate remaining time at all.
The problem with this is that the application can only recalculate estimated remaining time when it has new information. It doesn't get new information in the middle of performing an operation, only when operations are complete. Sure, some operations (e.g. copying a file) provide meaningful progress data at very regular intervals, but others don't. So what probably happens in the cases you're talking about it is that some indivisible operation that is supposed to take only a short time actually ends up taking a long time (e.g. opening a file on a CDROM should only take a few 10s of milliseconds, but if the system hits a bad block it will repeatedly spin the disc down, back up, and retry, an operation that can take over a minute), and therefore the progress indicator is suddenly wrong *but the system has no way to know it until the operation completes*.
I'm not going to argue about single/multi-core, except to say that any application that requires the kind of performance we're talking about is likely to be of the easily-parallelizable kind (image manipulation, etc), so is likely to get above average speed-up. But this, I will:
I would think you memory bandwidth is higher as well [on the desktop machine].
Nope. The desktop has a 667MT DDR3 x 64-bit interface = 5.3GB^-s. The phone's is 800MT and the same width, so 6.4GB^-s.
I've seen at least one hard disk manufacturer -- I think it was WD, but I'm not sure -- using these hybrid units, too, i.e. selling disks with 1GB == 1,024,000,000 bytes.
Slashdot Mirror
mmm.... Embarrassingly Parallel...
When you come down to it, it's amazing how many of the problems that actually require large amounts of processor time _are_ embarrassingly parallel. Presumably because the more calculations you have to perform, the more likely those calculations are to be partitionable into entirely independent data sets.
Are they firing programmers now? Last I knew, some penny-pincher figured out it was cheaper to just take them out back and shoot them..
Why d'you think they call it "firing"?
Blizzard has in fact proved - with the most popular MMO in history
Interesting. I didn't know Blizzard wrote Runescape.
How long have you had your domain? I've had mine for 10 years now, and I get a really weird combination of addresses. They've built up slowly over time. Some of them are pretty bizarre and totally unrelated to any address I've ever used. Some appear to derive from corrupted address lists that have been copied over and over (my normal address is myname@mydomain, I regular receive stuff to: mynamemyname@mydomain, myname.mydomain@mydomain, yname@mydomain, etc.) Some appear to be guesses of address I might likely use! (I used to contribute to the 'nasm' open source project, and I regularly get spam to 'nasm.source@mydomain' even though I have never used this address.) Still others are other peoples names @mydomain. I regularly get "brewster43@mydomain".
"I create a unique email address for each company I deal with, and each website I register on."
Does nobody of you morons know of mailinator.com?
Why on earth would someone create a mailaddress just to register to a website when mailinator with their gazillion aliases exists?
Just give them mythrowawaylogin@mailinator.com as email address, read it _once_ to click the confirmation link and forget it.
Reason 1: there are plenty of people using services like this - http://www.block-disposable-email.com/cms/
Reason 2: I may want to establish an ongoing relationship with a company (e.g. receive newsletters, etc) rather than just have a fire & forget initial contact
Reason 3: Having email coming to my inbox is more convenient than having to open a web site to view it. (I have a regexp-based email setup that allows me to just make up addresses that match a pattern, and I can add individual addresses to my spam filter if they become compromised, so it's actually easier than using mailinator).
It's highly unlikely the random "MID" tablet you've picked is the right one. He could have any of these:
http://www.manta.com.pl/pl/mid08/
http://www.rakuten.com/prod/android-4-0-ice-cream-sandwich-mid-7-capacitive-touch-screen-wi-fi-g/230139359.html
http://www.youtube.com/watch?v=VfH_AZ3AAUs
http://www.amazon.com/Google-Android-Capacitive-Gsensor-MID70404B/dp/B008XLCUF6
or thousands of others all using the name "MID".
The OP also asks for a ROM but never tells us the modell of tablet (except that it is 7'' and from MID, hope they have only one of those).
Unfortunately, this just betrays the fact that he doesn't know what he's talking about: MID isn't a manufacturer at all, but an abbreviation that several of the cheaper manufacturers use to identify a "mobile internet device", i.e. a tablet.
The Android-derived tablet he got is horrible! The battery is the least of its problems.
As a rule of thumb, if someone offers you a tablet for Christmas with resistive touch, you shouldn't even open the box and you should try to return it for a full/partial refund as soon as possible. As to the security issues, the article he pointed to talks about apps being "security risks" or "malware" for requesting GPS permissions when they shouldn't (which is really FUD). In any case, since his tablet doesn't have a GPS chip in it, that issue doesn't apply to him.
Also since he doesn't have access to the official Android Market/Google Play, he should just look on the XDA forums, root his device, install Cyanogen on it, and go through the Cyanogen repo for apps. And he should refrain from installing apps from other locations
You're making an awful lot of assumptions about his tablet when he's told us nothing useful about it other than the price. FWIW, I suspect you're wrong on 2 of the 3 conclusions you draw here: it's perfectly possible to get an $89 tablet that has capacitative touch and Google Play these days, but GPS will likely be out of the question.
But your suggested solution highlights the real problem with buying cheap-ass chinese hardware: there will likely not be an open source build available for it. There's a small chance that this build will work for him, but it's quite unlikely. Probably, nobody even knows what hardware he has in there. So he can't put cyanogenmod, or any of the "kick-ass ROMs" he's asking for, on it. He's stuck with the slightly-dubious build of AOSP the manufacturer came up with. And for his information, "MID" is not the manufacturer -- it's an abbreviation for "mobile internet device" that's used by several manufacturers who prefer not to put their own branding on the devices they make.
if we compare ANY console to any "high end rig" it looks outdated usually before it even comes out by that logic
Judging by benchmarks of bulldozer chips (I've never actually tested one myself), this thing will struggle to outperform the Phenom-II X6 system I put together last year for under £400, which is a long way from being a high end rig. Sure, it has more cores, and is running a more efficient architecture (clock-for-clock benchmarks put Bulldozer about 20% faster than the K10 cores that are in Phenom IIs), but it's clocked *much* slower.
I'll grant its GPU may be better, but I suspect if I'd shelled out another £100 on my graphics card I'd see comparable performance. Haven't looked at the benchmarks, though, so I'm just guessing there.
The resolution for TVs will still be 1080p so its not like theres going to be much need for more GPU horsepower and anything more than 30fps is wasted on a TV anyway.
GPU power increases are not only used to driver higher resolutions or framerates; they are also used to increase model detail level, add additional effects, and otherwise improve the quality of the graphics. Also, many modern TVs are able to display 50 and/or 60 fps quite comfortably, so suggesting faster than 30fps is wasted on a TV seems illogical -- no more so than it would be wasted on a monitor.
When the average teachers in Chicago are making ~75k / yr with incredible benefits
Citation to a credible source needed. The only place I see figures of $75K are news articles quoting a biased source. Unbiased sources (e.g. the various salary surveys) are reporting $55K or thereabouts.
Let's face it --- this wouldn't be the first time an employer has inflated claims about how much he's paying in an attempt to discredit unions negotiating for a better deal.
Not in any context that's relevant to the discussion you're posting to. It has been talking about America's shortcomings in response to an article about France's purported shortcomings.
No it isn't.
Hence the name Hong Kong and Shanghai Banking.
The company's head office is in Canada Square, London. It is listed on the LSE. It may originally have been based in Hong Kong, but it moved because regulations would have prevented them taking over the Midland Bank if they weren't British.
Because most businesses selling products through Amazon.com haven't laboriously modeled all the details of all the products that they sell in Blender
That's ok, because all you need is a set of photographs taken from a variety of angles, and that model can be built automatically.
Apparently I forgot to include the tag. I guess a reference to a really old DOS based golf game and a really old text based browser weren't enough of a clue.
The most recent version of the former was released in 2004, the latter in 2010. Neither of these qualify as "really old" IMO.
Funny, NeXTstep seemed pretty real time to me.
It's amazing how much performance you get when you spend a quarter of the national median annual salary on a single workstation.
Whereas in the good, honest Blue States of America, we just threaten hackers for political gain until they commit suicide. Much less expensive.
Probably somewhat influenced by the fact that the democrat AG only had a 4% lead on the republican candidate in the last election -- must be trying to impress the borderline voters.
Analysis cost is about $20k per genome for good enough coverage to distinguish CNV (and I work for a DNA sequencing company).
Really? My understanding was that whole genome sequencing was now available for $6k (as long as you're buying tests in bulk). Is there some reason a typical commercial whole-genome sequence would not be suitable for this purpose?
To catch those mutations, they won't be analyzing signatures, I don't think. A signature only works if you're sure that's where the mutations are. But they won't know any of that. They need to sequence the whole damn genome. That's around $50k a pop, right?
More like $6K.
There seems to be an assumption that the innocent twin has CONVINCING evidence of his brother's guilt, and/or his own innocence. I might have a butt-ton load of evidence that I'm happy to share, but none of it is convincing. What then?
Well, thankfully the prosecutor has to share convincing evidence that you are guilty, which seems to be lacking in this case.
I'm pretty sure you haven't actually seen that. In practice, what the HDD industry as a whole does is:
Well, actually I have. I got the company wrong, though -- it was Maxtor who used this definition. Quoting from the manual above:
Renault has a rather good safety record compared to other cars in the same class and price range and this is not how a "normal" Renault Laguna would handle.
You say that, but the last time we had a similar story here was a renault too: http://slashdot.org/story/04/10/05/1539203/a-car-with-a-mind-of-its-own
"Consider this: Once you've put progress on a bar, you can't take it off. Suppose you start a process that should take 20 minutes, and do the first 5 minutes, progress is now at 25%. But then, partway through, something unexpected happens and you realize the process is actually going to take 40 minutes. You can't take the progress "back" now, that would disorient the user. So you have to rescale the remainder of the bar."
But that's not what this person's app did. Whoever wrote their progress bar did it wrong.
The problem with many of these progress indicators is that they only calculate the estimated time once. Then if the app gets stuck or slows down at, say, 75%, it might say "about one minute" for 10 minutes. But that should never happen.
Instead, estimated remaining time -- when it is appropriate, which is a lot less often than it is actually used -- should be updated right along with the % progress. So yeah, as you say, if done properly, it can start out at 5 minutes and creep up to 20 or whatever. But it should not just freeze at some time and stay there a while. If it's going to do that, the app should probably not be trying to estimate remaining time at all.
The problem with this is that the application can only recalculate estimated remaining time when it has new information. It doesn't get new information in the middle of performing an operation, only when operations are complete. Sure, some operations (e.g. copying a file) provide meaningful progress data at very regular intervals, but others don't. So what probably happens in the cases you're talking about it is that some indivisible operation that is supposed to take only a short time actually ends up taking a long time (e.g. opening a file on a CDROM should only take a few 10s of milliseconds, but if the system hits a bad block it will repeatedly spin the disc down, back up, and retry, an operation that can take over a minute), and therefore the progress indicator is suddenly wrong *but the system has no way to know it until the operation completes*.
I'm not going to argue about single/multi-core, except to say that any application that requires the kind of performance we're talking about is likely to be of the easily-parallelizable kind (image manipulation, etc), so is likely to get above average speed-up. But this, I will:
I would think you memory bandwidth is higher as well [on the desktop machine].
Nope. The desktop has a 667MT DDR3 x 64-bit interface = 5.3GB^-s. The phone's is 800MT and the same width, so 6.4GB^-s.
I've seen at least one hard disk manufacturer -- I think it was WD, but I'm not sure -- using these hybrid units, too, i.e. selling disks with 1GB == 1,024,000,000 bytes.