Outside of BCD mechanisms, I think what you are trying to describe is called "decimal floating point", and is an IEEE standard. IBM likes to talk about decimal floating point, because even lowly x86 had BCD hardware support since day one (think AAA). IBM even likes to show benchmarks where their hardware DFP is faster than DFP using the same formats on intel. The problem is that DFP has a couple different storage mechanisms, and naturally IBM has compared the one were they do best against the one intel does worst at. In the end, apples to apples, they come out about the same. Surely not worth a 10x price difference for a few percent in IBM's favor, which ends up being application specific anyway.
Yes, you can pull the whole blade, if the application is cluster aware. Just about any modern DB can be run like this, as can your favorite web server. Most "enterprise" applications are designed to run in a loosely coupled environment.
Except that the HP Blade cluster has nothing on the mainframe in terms of reliability and data integrity.
Got any real numbers to back that up?
No? Maybe its because most of the failures, on both are due to software instead of hardware. There is nothing magical about the software on a zseries. If anything it seems more likely for a system programmer to screw up some JCL and destroy something important, than for the esx administrator to accidentally delete an image.
The fusionIO disks, are expensive, so are the texas memory system ones, but they are still an order of magnitude cheaper than anything available for zseries. Oh, and putting a fusionIO PCI board in your server, will probably increase its transaction numbers beyond anything achievable for less than 9 figures in the mainframe space.
So, sure 1TB of memory, 64 processors, a couple 8/16Gb FC adapters/infiniband/etc, fusionIO boards, etc, your x86 might cost $100k, but its going stomp all over a mainframe that costs $1M.
And they (modern x86 servers) have ECC on every bus from the registers to the disk, and cost about 1/10 (or better) than a similar zeries.
There is a reason IBM doesn't publish benchmarks and prices for these things. I recently got a machine from IBM, which costs more than my house, and is slower than a 486. IO rates? Same, its got 8GBit FICON, but doesn't seem to be able to drive more than a small fraction. I'm sure I could spend a few million to unlock the processors, and a few more IO boards, and more than 8GB of RAM. Then it might be fast enough to keep up with the DL580's in the rack next to it.
Plus, the DL580 can have linux installed and running in under an hour, the mainframe? If its running in under a few days its a small miracle. The management aspects, are a total joke. There is more processing power in PCs connected to the thing for management/translation/etc (5 of them!) than it has.
IBM mainframe is far cheaper in cost / transaction
Really? So that is why IBM published tpcc, spec-vm, spec-cpu, etc benchmarks for these machines? No, IBM doesn't publish these benchmarks because they would rather make up their own.
Frankly, I think its been a whispered secret for years that pseries machines best the zseries in every metric.
Having just got my first z114, I call BS. To me it seems that VMWare is actually more advanced. At least I can install an OS in under 30+ hours, and migrating an image around is as easy as a couple clicks. Plus, I don't have to work around the OS's inability to fully utilize disks >54GB.
He he he B&W, by definition, indicates a lack of more money than brains.
A couple years ago I got into it with the sales guy at the local B&W dealer. I was sitting in one of the high end rooms, and I noticed that the "sweet" spot was about 3' wide by about 6" high. I was like WTF kind of speaker has such a narrow sound field. One of the marks of a quality set of speakers is a uniform sound field. If I can hear a huge change in the intensity of the high frequency range by standing up, there is something seriously wrong with your speakers. Basically, my stated option was that just about any cheap walmart speakers were better than the $50k ones they were demoing.
I stay away from those places, I have a tendency to call people out on marketing BS. A few years earlier I got into it with a guy selling a $5k CD player that did "128x" oversampling.
I didn't find javascript to be unbearable. It didn't take to much time, and I learned how to make it do what I wanted. The ugly parts were all DOM related, and even that wasn't too bad once I got the basics. Most of the nasty issues had fairly straightforward workarounds once I found out they were there (for example, the relationship between DOM modifications being posted and when an event completes between browsers). Now there is jquery, which has pretty much hidden all the ugly DOM nasties, and makes javascript nice in most regards. I still dislike the OO model, but I can deal with it.
That said, CSS is an absolute abomination. I can't say anything positive about it. The general idea about separating the content from the layout is a good one. The problem is that CSS is so fundamentally broken that probably 4/5ths of everything I want to do, requires some nasty hack or is simply impossible without using javascript to position something on the fly. I guess my problem is that i've used so many GUI and print layout engines, that I expect there is a way to do things that I imagine, rather than trying to fit my ideas in the constraints of CSS. Nor do I like sprinkling divs all over the place to work around asymmetries in what attributes can be applied to existing tags. Worse yet, are the divs that need to be sprinkled as tag holders. Plus, the fact that I need to use a CSS preprocessor, for basic things seems insane. I also hate its inability to explicitly reference a parent or siblings with respect to simple layout concepts like rightof/leftof, instead requiring layer over layer of divs. I shouldn't have to screw around with the markup to do simple layout changes. I could probably write a hundred page book without even trying, on things that are wrong with CSS.
Frankly, I was so sick of it, I wrote a layout engine in javascript, that entirely replaced HTML and CSS, with a generic set of XML tags more based on traditional GUI concepts. I learned a whole crap load about javascript and the DOM doing it too. The sad part, is that I never really finished it. Its good enough to draw some simple overlapping windows, and allow the user to reposition them, click buttons, draw things, etc, but its not really good enough for prime time. A big portion of the problem is that its hard to do fast flicker free animation and drawing simply from JS. Sure I can draw a region, and a scroll bar, but when the user scrolls the region, it should be nearly as fast as doing it with just a normal browser and HTML. Doing all the drawing/layout with javascript makes this nearly impossible. Then the single threaded nature of the event system in all the browsers also tends to cause input lag if the redraw process gets involved.
Basically, in the end, there is a reason flash, silverlight, java web start, and random other plugins still exist. Doing heavyweight UI's using JS/HTML/CSS/etc is a PITA.
Besides, the well known reliably issues with SF controllers, i wonder if they fixed the resume problems that seem to crop up on a regular basis. The forums are full of people (including one of my machines) reporting their machines simply blue screen with some fairly high frequency when resuming with these drives. The problems seem to cross vendors too...
I agree -- we need less government, not more. Regulations hurt small businesses, and decades of that crap is well-documented and understood.
WTF, exactly are you basing this on? I hear a certain political element spewing this all the time, but I have yet to hear a concrete provable example of how its actually hurting. I've seen a fair amount of regulation, but invariably the small cost of jumping through some hoop is seriously outweighed by the benefit to society. Not having every house on your block burn down because the building codes require circuit breakers, and minimum wire gauges is a benefit. Not dying from lung cancer because the factory one state over is dumping crap into the atmosphere is also a benefit. Having children that don't' have cognitive problems because the water isn't full of heavy metals might also be an advantage.
Sure i've heard some BS arguments about providing maternity leave causes all this lost productivity and crap like that, but in all the actual cases I ever heard there is actual evidence to the contrary. Especially, since people are going to get pregnant/sick/etc independent of their employers coverage.
If anything, a certain political party that started to gain control over the government more frequently in the late 70's corresponds directly to the weakening of the middle class. Frankly, there is direct evidence that the current tax structure (most of the tax burden on the middle class rather than being shared by business and the upper class), and free trade agreements, and weakening unions, are a triple whammy to the middle class. Its nearly indisputable that trade with low wage countries, and the war on labor unions suppress wages. Claiming that we need to reduce social programs (roads, schools, safety nets, etc) _will_ further weaken the economy in the long run. So, basically those claiming we need more of what has gotten us into the current predicament, is little more than ignorance. Cutting taxes, is really codewords for cutting spending on all the things that provide the backstop for the economy. Sure, the GDP might go up, but it will continue to be an improvement for the top 5% or so, everyone else will get screwed, because that is the result of the current policies. Go ahead, vote for your favorite tax cutter, I'm going to laugh at your ignorance when your 1st grader still can't read, and struggles with math their entire life. I'm also going to laugh at you every-time, you fill up your gas tank. Because, until people start actually thinking, rather than letting faux news fill their heads with propaganda (because that is actually what it is, go study what constitutes propaganda, and then study how its been used in the past, and get back to me...) the state of this country isn't going to get better.
Its all over the place, and looks/acts like someone bolted random paradigms all over it, depending on which subsystem you happen to be using. This primary seems to be a function of multiple groups responsible for different portions of the UI. The guys doing explorer were on one page, the guys doing paint where on another, the guys doing the administrative panels were on a third (or rather didn't see any reason to update their UI from xp). The UI ranges from things like notepad which are using the windows 3.0 (menu only, no button bars) all the way to paint which was updated to match the office paradigm and doesn't even have menu bars by default.
There is obviously no one at MS who is in charge of the look and feel of the product, because they would never have allowed paint to be updated without calculating the time to update the remainder of the OS (at a minimum explorer) and making sure the changes were complete.
Yah, the win8 dev previews are pushing me to buy the first mac I've owned since the early '90s. The vista/win7 abomination (7 only seems good if you had to use vista, the UI is a clusterfsk *). The whole thing is probably on purpose, they will come out with win9, and reverse a couple decisions and everyone will praise it for being great. I think business is really pushing back right now, even against 7. My wife started a new job a few months back, and her brand new PC was running XP, the IBM CE shows up at work, and he is running XP, everywhere I look outside of the local coffee shop full of college kids I see XP.
Then there are the "cord cutters". I see them at the trendy places, and at my friends houses. They don't even use there PC's anymore. The ipads have replaced them.
*, I pointed this out to a co-worker by starting 6 or 8 different apps that come with it and pointing out it looks like a history of computing, some of the UI is win3, some of it is 98, some of it is XP, some of it is vista, some of it is office 07. If your going to convert the UI to a ribbon, _EVERYTHING_ needs to be a ribbon...
Its a visual representation of the complete lack of management at MS.
Yah me three, I haven't had to screw with monitor timings in 15 years until I decided to try plugging my PC into a HDMI TV. For some reason 1080p from the HDMI port on my PC isn't the same as 1080p from my entertainment stack.
What we haven't done yet is discovered any lossless parallelizeable compression schemes.
Uh, there are a lot of ways to parallelize loss-less compression schemes. I've been involved in projects doing this a couple times over the last decade. One example out of a half dozen I can think off of the top of my head is the history buffer search in LZ77 can be parallelized. How you go about that will make a huge difference in how fast it is.
Is that its not really useful for learning OpenCL. Sure it will teach you the syntax and how to write an OpenCL program. That isn't the problem. The problem is that if your writing something in OpenCL you probably want it to be fast. Learning the language is doable by someone with C experience in just a couple hours with just the SDKs shipped by AMD/Nividia/Intel. Learning how to optimize a routine for a particular GPU/etc is the hard part, and is application specific. It also requires knowledge of how compute device actually work at an extremely low level. I don't believe this book teaches that. Save your money, download the spec and a SDK for your device. Start reading the architecture docs..
Coding for it in OpenCL isn't much different than writing C code that is just a wrapper around some assembly. There is no reason a MUCH more human friendly interface couldn't be made with the compiler taking care of using the appropriate memory and instructions to optimize for GPU usage.
As someone who has actually done some OpenCL programming, I can tell you why your wrong. Learning openCL syntax isn't hard, if you know C# you can probably write some useful openCL code in just an hour or two. It is after all, a C-like language just like C# is a C like language.
That said, don't expect your openCL code to run faster than similar C code compiled with SSE. Thats because making OpenCL run fast is an exercise is looking at memory access patterns, understanding how to share data between hundreds of threads efficiently, etc. My first openCL program was actually slower (by 1/2) than a similar program using all 8 cores of my CPU. I got it on par with the CPU using a top of the line AMD GPU within a day or so, and then spent another two weeks trying different things until finally finding the magic bullet which removed a memory collision I was having and by itself increased the performance of my routine by ~32x. Running the same code on an nvidia GPU put me back in the ballpark of my CPUs again, requiring more time to make it fast on those GPUs. Time I wasn't willing to spend.
The bottom line is that OpenCL could be any language, but, what is necessarily is the ability to make changes which affect how data is laid out in memory, and how that data is being read/written. Furthermore, you need the ability to specify where the memory is used, because GPU's have unforgiving memory hierarchy. So if your not comfortable with the nitty gritty details of how computers (or in this case GPUs) actually work (not some CompSci abstraction) your not going to write good OpenCL code. You also need a gut feeling for how fast something could be, based on the specification of a particular device. Otherwise you won't know when to give up.
They removed all the hard to implement features that require tradeoffs, and dumbed down the system to the point where a decent subset of their power users will continue to use NTFS...
All the while missing the single most important thing to happen in storage in the last 50 years... Aka SSDs and the ability to seek quickly.
I really can't see how bombers are going to survive. They have enough problems with manned fighters. If a nation can build surface to air missiles that are tracking the bombers on multiple wavelengths (visible and otherwise), and have the ability to detect movement (rotation/translation/etc) with nearly 100% accuracy from both below and above, what chance does a bomber have? Especially against a hail of a dozens of devices built for 1/1000 the price of the bomber?
A bomber is going to need the equivalent of a carrier support group... Or possibly an arsenal of active countermeasures (aka themselves, self guided air to air missiles).
It is my firm belief that the next real war (if it lasts more than a few minutes) will totally remove the concept of air superiority, as well as above surface ships. I can't imagine that any serious national defense group isn't looking at totally automated surface to air and surface to surface missiles with computer vision and countermeasure filtering/avoidance systems good enough that they can hit ships and airborne aircraft larger than a few meters traveling at less than say 1500 km/h.
Computer vision systems are simply getting too good, and they are getting better every year. Given enough time and a system that can reacquire targets even if they are flying barely above ground level, it seems that it has to become a computer vs computer type problem. The nation that can build the largest number of small/fast/cheap missiles wins.
And yes, if he survived a 108 MPH accident without a seat belt and the air bag saved him without major injuries it's nearly a guarantee he'd have the same outcome with a seat belt.
I'm not so sure. _IF_ you hit the airbag just right its probably better than having a seatbelt on. That is because the deceleration of automobile crumple zones, and the airbag pressure is probably less than the seatbelt impact and the crumple zones. Seatbelts save lives because they keep you in the automobile, and they keep you from being impaled by something in the automobile (think gearshift). With seatbelts, your deceleration rate is mostly dependent on the crumble zones. With the airbag, your getting a foot or so extra deceleration space added to the crumple zones.
All that said, In his case it sounds like he was _REALLY_ lucky as the car flipped. Those are the kinds of situations where you really want to have the seatbelt.
First reason, redundancy. If your backup policy specifies that your making copies of the data on a regular basis, and you then proceed to delete all the copies but one, why are you making the copies in the first place? Maybe instead of dedupe you need better backup software that can detect the redundancy and simply choose not to back up what hasn't changed.
Second, performance. By its very nature dedupe degrades over time, the dedupe vendors battle this in a number of ways (secondary caches, inverted dedupe streams, etc) . Eventually though streams of data on become sequences of tokens scattered all over the storage system in nearly random patterns. Its the same as having a file system where every 4k of data in a file requires head seeks.
Scale, as the amount of physical capacity increases the need to maintain hash lookups for that physical capacity increases. These hash lookups pretty much must resides in physical RAM. This leads to expensive dedupe nodes or massive dedupe inefficiency if the data is split across multiple standalone dedupe appliances.
Price, the above limits tend to drive the price per deduped MB up beyond the price of RAID arrays from 2nd tier vendors like nexan, acnc, etc.
Now all that said, there are places where dudup provides an advantage, but they are few and far between. That is because not everyone get 30:1 (especially if their backup software isnt doing full backups). Generally the dedupe systems are at the bottom of the performance curve and not everyone is willing to grow their backup window or slow down there apps either. That tends to leave them in fairly low end enviroments often better served by inexpensive raid boxes.
Its about any field, the more you study any particular area, the more you realize how wrong congress is about it. I started really noticing about 20 years ago that it seemed that I disagreed with the vast majority of their activity, back then, I wrote some of it off as my own personal ignorance, but the more I learn about economics, technology, etc the more I understand just how wrong they are about those areas. From that I am forced to cynically extrapolate that they are probably just as wrong about areas in which I am not knowledgeable.
you cannot win this war. Its like the virus scanners. no virus scanner really knows the most recent virus. so no website will be able to provide ads in a way that cannot be blocked.
Are you sure you don't have this backwards? The ad blockers are the reactionary ones in this case (the virus scanners), as I can grab the latest version of the ad block software and tune my site to avoid it and its blacklist. Then I get a small window where everyone sees my ads, but in the meantime I can come up with another way to beat it, then when the ad block blacklist gets published, I simply update my site again.
Of course economics eventually kicks in and the one or the other looses, but the ease of updating the blacklist on ad blocking software is much less than the effort of building new ad display platforms. Sometimes the engine needs updating too, but probably only evens the playing field.
So is the garbage collection mechanism of Java: memory is a shared resource.
And that is one of the 3 or so major flaws in garbage collection, especially as implemented by Java. The memory requirements of the application are not reflected in the memory footprint being requested of the OS. This can cause _SERIOUS_ performance consequences in systems where there are multiple applications (one or more not using as shared jvm).
I saw it a few years ago in an application I worked on. The load of the application varied over time, and on different machines. Invariably though the JVM would have gigs of memory allocated for a basically idle app and cause another application on the machine to be paged, when the actual footprint requirements were fairly minor. Then the java app would scale up for a few hours. The end result was pretty much throw 10x the hardware at the problem, because convincing the JVM to actually back down the memory footprints at the appropriate times required our app to be able to poll the memory situation of the host machine and make determinations about its state, which wasn't nearly as fine grained as the OS was capable of. The resulting C version of the java app, quite literally ran on a single machine vs the dozen or so the java app required.
Slashdot Mirror
Outside of BCD mechanisms, I think what you are trying to describe is called "decimal floating point", and is an IEEE standard. IBM likes to talk about decimal floating point, because even lowly x86 had BCD hardware support since day one (think AAA). IBM even likes to show benchmarks where their hardware DFP is faster than DFP using the same formats on intel. The problem is that DFP has a couple different storage mechanisms, and naturally IBM has compared the one were they do best against the one intel does worst at. In the end, apples to apples, they come out about the same. Surely not worth a 10x price difference for a few percent in IBM's favor, which ends up being application specific anyway.
Yes, you can pull the whole blade, if the application is cluster aware. Just about any modern DB can be run like this, as can your favorite web server. Most "enterprise" applications are designed to run in a loosely coupled environment.
Except that the HP Blade cluster has nothing on the mainframe in terms of reliability and data integrity.
Got any real numbers to back that up?
No? Maybe its because most of the failures, on both are due to software instead of hardware. There is nothing magical about the software on a zseries. If anything it seems more likely for a system programmer to screw up some JCL and destroy something important, than for the esx administrator to accidentally delete an image.
The fusionIO disks, are expensive, so are the texas memory system ones, but they are still an order of magnitude cheaper than anything available for zseries. Oh, and putting a fusionIO PCI board in your server, will probably increase its transaction numbers beyond anything achievable for less than 9 figures in the mainframe space.
So, sure 1TB of memory, 64 processors, a couple 8/16Gb FC adapters/infiniband/etc, fusionIO boards, etc, your x86 might cost $100k, but its going stomp all over a mainframe that costs $1M.
And they (modern x86 servers) have ECC on every bus from the registers to the disk, and cost about 1/10 (or better) than a similar zeries.
There is a reason IBM doesn't publish benchmarks and prices for these things. I recently got a machine from IBM, which costs more than my house, and is slower than a 486. IO rates? Same, its got 8GBit FICON, but doesn't seem to be able to drive more than a small fraction. I'm sure I could spend a few million to unlock the processors, and a few more IO boards, and more than 8GB of RAM. Then it might be fast enough to keep up with the DL580's in the rack next to it.
Plus, the DL580 can have linux installed and running in under an hour, the mainframe? If its running in under a few days its a small miracle. The management aspects, are a total joke. There is more processing power in PCs connected to the thing for management/translation/etc (5 of them!) than it has.
IBM mainframe is far cheaper in cost / transaction
Really? So that is why IBM published tpcc, spec-vm, spec-cpu, etc benchmarks for these machines? No, IBM doesn't publish these benchmarks because they would rather make up their own.
Frankly, I think its been a whispered secret for years that pseries machines best the zseries in every metric.
Having just got my first z114, I call BS. To me it seems that VMWare is actually more advanced. At least I can install an OS in under 30+ hours, and migrating an image around is as easy as a couple clicks. Plus, I don't have to work around the OS's inability to fully utilize disks >54GB.
The list goes on and on.
He he he B&W, by definition, indicates a lack of more money than brains.
A couple years ago I got into it with the sales guy at the local B&W dealer. I was sitting in one of the high end rooms, and I noticed that the "sweet" spot was about 3' wide by about 6" high. I was like WTF kind of speaker has such a narrow sound field. One of the marks of a quality set of speakers is a uniform sound field. If I can hear a huge change in the intensity of the high frequency range by standing up, there is something seriously wrong with your speakers. Basically, my stated option was that just about any cheap walmart speakers were better than the $50k ones they were demoing.
I stay away from those places, I have a tendency to call people out on marketing BS. A few years earlier I got into it with a guy selling a $5k CD player that did "128x" oversampling.
I didn't find javascript to be unbearable. It didn't take to much time, and I learned how to make it do what I wanted. The ugly parts were all DOM related, and even that wasn't too bad once I got the basics. Most of the nasty issues had fairly straightforward workarounds once I found out they were there (for example, the relationship between DOM modifications being posted and when an event completes between browsers). Now there is jquery, which has pretty much hidden all the ugly DOM nasties, and makes javascript nice in most regards. I still dislike the OO model, but I can deal with it.
That said, CSS is an absolute abomination. I can't say anything positive about it. The general idea about separating the content from the layout is a good one. The problem is that CSS is so fundamentally broken that probably 4/5ths of everything I want to do, requires some nasty hack or is simply impossible without using javascript to position something on the fly. I guess my problem is that i've used so many GUI and print layout engines, that I expect there is a way to do things that I imagine, rather than trying to fit my ideas in the constraints of CSS. Nor do I like sprinkling divs all over the place to work around asymmetries in what attributes can be applied to existing tags. Worse yet, are the divs that need to be sprinkled as tag holders. Plus, the fact that I need to use a CSS preprocessor, for basic things seems insane. I also hate its inability to explicitly reference a parent or siblings with respect to simple layout concepts like rightof/leftof, instead requiring layer over layer of divs. I shouldn't have to screw around with the markup to do simple layout changes. I could probably write a hundred page book without even trying, on things that are wrong with CSS.
Frankly, I was so sick of it, I wrote a layout engine in javascript, that entirely replaced HTML and CSS, with a generic set of XML tags more based on traditional GUI concepts. I learned a whole crap load about javascript and the DOM doing it too. The sad part, is that I never really finished it. Its good enough to draw some simple overlapping windows, and allow the user to reposition them, click buttons, draw things, etc, but its not really good enough for prime time. A big portion of the problem is that its hard to do fast flicker free animation and drawing simply from JS. Sure I can draw a region, and a scroll bar, but when the user scrolls the region, it should be nearly as fast as doing it with just a normal browser and HTML. Doing all the drawing/layout with javascript makes this nearly impossible. Then the single threaded nature of the event system in all the browsers also tends to cause input lag if the redraw process gets involved.
Basically, in the end, there is a reason flash, silverlight, java web start, and random other plugins still exist. Doing heavyweight UI's using JS/HTML/CSS/etc is a PITA.
Besides, the well known reliably issues with SF controllers, i wonder if they fixed the resume problems that seem to crop up on a regular basis. The forums are full of people (including one of my machines) reporting their machines simply blue screen with some fairly high frequency when resuming with these drives. The problems seem to cross vendors too...
I agree -- we need less government, not more. Regulations hurt small businesses, and decades of that crap is well-documented and understood.
WTF, exactly are you basing this on? I hear a certain political element spewing this all the time, but I have yet to hear a concrete provable example of how its actually hurting. I've seen a fair amount of regulation, but invariably the small cost of jumping through some hoop is seriously outweighed by the benefit to society. Not having every house on your block burn down because the building codes require circuit breakers, and minimum wire gauges is a benefit. Not dying from lung cancer because the factory one state over is dumping crap into the atmosphere is also a benefit. Having children that don't' have cognitive problems because the water isn't full of heavy metals might also be an advantage.
Sure i've heard some BS arguments about providing maternity leave causes all this lost productivity and crap like that, but in all the actual cases I ever heard there is actual evidence to the contrary. Especially, since people are going to get pregnant/sick/etc independent of their employers coverage.
If anything, a certain political party that started to gain control over the government more frequently in the late 70's corresponds directly to the weakening of the middle class. Frankly, there is direct evidence that the current tax structure (most of the tax burden on the middle class rather than being shared by business and the upper class), and free trade agreements, and weakening unions, are a triple whammy to the middle class. Its nearly indisputable that trade with low wage countries, and the war on labor unions suppress wages. Claiming that we need to reduce social programs (roads, schools, safety nets, etc) _will_ further weaken the economy in the long run. So, basically those claiming we need more of what has gotten us into the current predicament, is little more than ignorance. Cutting taxes, is really codewords for cutting spending on all the things that provide the backstop for the economy. Sure, the GDP might go up, but it will continue to be an improvement for the top 5% or so, everyone else will get screwed, because that is the result of the current policies. Go ahead, vote for your favorite tax cutter, I'm going to laugh at your ignorance when your 1st grader still can't read, and struggles with math their entire life. I'm also going to laugh at you every-time, you fill up your gas tank. Because, until people start actually thinking, rather than letting faux news fill their heads with propaganda (because that is actually what it is, go study what constitutes propaganda, and then study how its been used in the past, and get back to me...) the state of this country isn't going to get better.
Question: What is wrong with the Win 7 UI?
I thought I had answered in the original posting.
One simple word: "Inconsistent"
Its all over the place, and looks/acts like someone bolted random paradigms all over it, depending on which subsystem you happen to be using. This primary seems to be a function of multiple groups responsible for different portions of the UI. The guys doing explorer were on one page, the guys doing paint where on another, the guys doing the administrative panels were on a third (or rather didn't see any reason to update their UI from xp). The UI ranges from things like notepad which are using the windows 3.0 (menu only, no button bars) all the way to paint which was updated to match the office paradigm and doesn't even have menu bars by default.
There is obviously no one at MS who is in charge of the look and feel of the product, because they would never have allowed paint to be updated without calculating the time to update the remainder of the OS (at a minimum explorer) and making sure the changes were complete.
Yah, the win8 dev previews are pushing me to buy the first mac I've owned since the early '90s. The vista/win7 abomination (7 only seems good if you had to use vista, the UI is a clusterfsk *). The whole thing is probably on purpose, they will come out with win9, and reverse a couple decisions and everyone will praise it for being great. I think business is really pushing back right now, even against 7. My wife started a new job a few months back, and her brand new PC was running XP, the IBM CE shows up at work, and he is running XP, everywhere I look outside of the local coffee shop full of college kids I see XP.
Then there are the "cord cutters". I see them at the trendy places, and at my friends houses. They don't even use there PC's anymore. The ipads have replaced them.
*, I pointed this out to a co-worker by starting 6 or 8 different apps that come with it and pointing out it looks like a history of computing, some of the UI is win3, some of it is 98, some of it is XP, some of it is vista, some of it is office 07. If your going to convert the UI to a ribbon, _EVERYTHING_ needs to be a ribbon...
Its a visual representation of the complete lack of management at MS.
Yah me three, I haven't had to screw with monitor timings in 15 years until I decided to try plugging my PC into a HDMI TV. For some reason 1080p from the HDMI port on my PC isn't the same as 1080p from my entertainment stack.
What we haven't done yet is discovered any lossless parallelizeable compression schemes.
Uh, there are a lot of ways to parallelize loss-less compression schemes. I've been involved in projects doing this a couple times over the last decade. One example out of a half dozen I can think off of the top of my head is the history buffer search in LZ77 can be parallelized. How you go about that will make a huge difference in how fast it is.
Is that its not really useful for learning OpenCL. Sure it will teach you the syntax and how to write an OpenCL program. That isn't the problem. The problem is that if your writing something in OpenCL you probably want it to be fast. Learning the language is doable by someone with C experience in just a couple hours with just the SDKs shipped by AMD/Nividia/Intel. Learning how to optimize a routine for a particular GPU/etc is the hard part, and is application specific. It also requires knowledge of how compute device actually work at an extremely low level. I don't believe this book teaches that. Save your money, download the spec and a SDK for your device. Start reading the architecture docs..
Coding for it in OpenCL isn't much different than writing C code that is just a wrapper around some assembly. There is no reason a MUCH more human friendly interface couldn't be made with the compiler taking care of using the appropriate memory and instructions to optimize for GPU usage.
As someone who has actually done some OpenCL programming, I can tell you why your wrong. Learning openCL syntax isn't hard, if you know C# you can probably write some useful openCL code in just an hour or two. It is after all, a C-like language just like C# is a C like language.
That said, don't expect your openCL code to run faster than similar C code compiled with SSE. Thats because making OpenCL run fast is an exercise is looking at memory access patterns, understanding how to share data between hundreds of threads efficiently, etc. My first openCL program was actually slower (by 1/2) than a similar program using all 8 cores of my CPU. I got it on par with the CPU using a top of the line AMD GPU within a day or so, and then spent another two weeks trying different things until finally finding the magic bullet which removed a memory collision I was having and by itself increased the performance of my routine by ~32x. Running the same code on an nvidia GPU put me back in the ballpark of my CPUs again, requiring more time to make it fast on those GPUs. Time I wasn't willing to spend.
The bottom line is that OpenCL could be any language, but, what is necessarily is the ability to make changes which affect how data is laid out in memory, and how that data is being read/written. Furthermore, you need the ability to specify where the memory is used, because GPU's have unforgiving memory hierarchy. So if your not comfortable with the nitty gritty details of how computers (or in this case GPUs) actually work (not some CompSci abstraction) your not going to write good OpenCL code. You also need a gut feeling for how fast something could be, based on the specification of a particular device. Otherwise you won't know when to give up.
They removed all the hard to implement features that require tradeoffs, and dumbed down the system to the point where a decent subset of their power users will continue to use NTFS...
All the while missing the single most important thing to happen in storage in the last 50 years... Aka SSDs and the ability to seek quickly.
I really can't see how bombers are going to survive. They have enough problems with manned fighters. If a nation can build surface to air missiles that are tracking the bombers on multiple wavelengths (visible and otherwise), and have the ability to detect movement (rotation/translation/etc) with nearly 100% accuracy from both below and above, what chance does a bomber have? Especially against a hail of a dozens of devices built for 1/1000 the price of the bomber?
A bomber is going to need the equivalent of a carrier support group... Or possibly an arsenal of active countermeasures (aka themselves, self guided air to air missiles).
The next air superiority fighter will be a drone.
If by drone you mean missile...
It is my firm belief that the next real war (if it lasts more than a few minutes) will totally remove the concept of air superiority, as well as above surface ships. I can't imagine that any serious national defense group isn't looking at totally automated surface to air and surface to surface missiles with computer vision and countermeasure filtering/avoidance systems good enough that they can hit ships and airborne aircraft larger than a few meters traveling at less than say 1500 km/h.
Computer vision systems are simply getting too good, and they are getting better every year. Given enough time and a system that can reacquire targets even if they are flying barely above ground level, it seems that it has to become a computer vs computer type problem. The nation that can build the largest number of small/fast/cheap missiles wins.
And yes, if he survived a 108 MPH accident without a seat belt and the air bag saved him without major injuries it's nearly a guarantee he'd have the same outcome with a seat belt.
I'm not so sure. _IF_ you hit the airbag just right its probably better than having a seatbelt on. That is because the deceleration of automobile crumple zones, and the airbag pressure is probably less than the seatbelt impact and the crumple zones. Seatbelts save lives because they keep you in the automobile, and they keep you from being impaled by something in the automobile (think gearshift). With seatbelts, your deceleration rate is mostly dependent on the crumble zones. With the airbag, your getting a foot or so extra deceleration space added to the crumple zones.
All that said, In his case it sounds like he was _REALLY_ lucky as the car flipped. Those are the kinds of situations where you really want to have the seatbelt.
But dedupe is 80% snake oil.
First reason, redundancy. If your backup policy specifies that your making copies of the data on a regular basis, and you then proceed to delete all the copies but one, why are you making the copies in the first place? Maybe instead of dedupe you need better backup software that can detect the redundancy and simply choose not to back up what hasn't changed.
Second, performance. By its very nature dedupe degrades over time, the dedupe vendors battle this in a number of ways (secondary caches, inverted dedupe streams, etc) . Eventually though streams of data on become sequences of tokens scattered all over the storage system in nearly random patterns. Its the same as having a file system where every 4k of data in a file requires head seeks.
Scale, as the amount of physical capacity increases the need to maintain hash lookups for that physical capacity increases. These hash lookups pretty much must resides in physical RAM. This leads to expensive dedupe nodes or massive dedupe inefficiency if the data is split across multiple standalone dedupe appliances.
Price, the above limits tend to drive the price per deduped MB up beyond the price of RAID arrays from 2nd tier vendors like nexan, acnc, etc.
Now all that said, there are places where dudup provides an advantage, but they are few and far between. That is because not everyone get 30:1 (especially if their backup software isnt doing full backups). Generally the dedupe systems are at the bottom of the performance curve and not everyone is willing to grow their backup window or slow down there apps either. That tends to leave them in fairly low end enviroments often better served by inexpensive raid boxes.
Its about any field, the more you study any particular area, the more you realize how wrong congress is about it. I started really noticing about 20 years ago that it seemed that I disagreed with the vast majority of their activity, back then, I wrote some of it off as my own personal ignorance, but the more I learn about economics, technology, etc the more I understand just how wrong they are about those areas. From that I am forced to cynically extrapolate that they are probably just as wrong about areas in which I am not knowledgeable.
you cannot win this war. Its like the virus scanners. no virus scanner really knows the most recent virus. so no website will be able to provide ads in a way that cannot be blocked.
Are you sure you don't have this backwards? The ad blockers are the reactionary ones in this case (the virus scanners), as I can grab the latest version of the ad block software and tune my site to avoid it and its blacklist. Then I get a small window where everyone sees my ads, but in the meantime I can come up with another way to beat it, then when the ad block blacklist gets published, I simply update my site again.
Of course economics eventually kicks in and the one or the other looses, but the ease of updating the blacklist on ad blocking software is much less than the effort of building new ad display platforms. Sometimes the engine needs updating too, but probably only evens the playing field.
So is the garbage collection mechanism of Java: memory is a shared resource.
And that is one of the 3 or so major flaws in garbage collection, especially as implemented by Java. The memory requirements of the application are not reflected in the memory footprint being requested of the OS. This can cause _SERIOUS_ performance consequences in systems where there are multiple applications (one or more not using as shared jvm).
I saw it a few years ago in an application I worked on. The load of the application varied over time, and on different machines. Invariably though the JVM would have gigs of memory allocated for a basically idle app and cause another application on the machine to be paged, when the actual footprint requirements were fairly minor. Then the java app would scale up for a few hours. The end result was pretty much throw 10x the hardware at the problem, because convincing the JVM to actually back down the memory footprints at the appropriate times required our app to be able to poll the memory situation of the host machine and make determinations about its state, which wasn't nearly as fine grained as the OS was capable of. The resulting C version of the java app, quite literally ran on a single machine vs the dozen or so the java app required.