I think you meant to say "... performs better with larger files".
You nailed it though. My big gripe with Windows it how it seems to spend more time fiddling with metadata / directory entries than the actual contents. On an SSD with 700mb/sec writes and 0.1 msec access times, I'd expect it to churn through a few thousand files per second at the very least. That's not even factoring the disk cache. All those MFT updates seem to drag it right back down to spinning-disk speeds when dealing with numerous small files. You know, like a source tree or a directory full of images.
As sequential storage performance continues to improve, filesystem overhead is becoming the primary bottleneck.
As they should be. Once you acknowledge the fact that money is an artificial construct, the only realization is that accountants truly create nothing in the enterprise. They don't produce a saleable asset. They don't offer any services to the clients.
If you run a company without an accountant, the only bad thing that will happen is the tax man will get angry.
If you run a company without software, you have no company.
Techreport seems to be slashdotted. Am I being too harsh, or is that horribly embarrassing for a site that focuses on performance testing and overclocking ?
It took the game's consolified low-poly meshes and prorgammatically inflated the poly count, barely enhancing the image quality at all, while driving processing requirements way up. It looks more like one of those "repeat N 1000 times" benchmarks, the kind I write when trying to find bottlenecks in a web page, than any sort of effort to make Crysis 2 not look like the steaming EA-published turd it is.
Crysis 1 still looks better than this half-assed sequel, DX11 be damned.
If this is some bizarre partnership between EA and NVidia to push GPUs, shame on them both. I've come to expect this from EA, by far the greediest chop-shop in the industry, but NVidia... come on, have some self-respect. Treating your customers like blind idiots is NOT going to help move $1500 quad-SLI setups. Between these stunts, and ATI's stagnant driver performance, they are just creating a giant opening for a new contender to come hoover up their marketshare.
Sure, you can do shading like that, if you're only using flat shading - a very ancient technique that was made virtually obsolete when pixel shaders hit the mainstream. A proper shading algorithm does not need distinct polys to apply gradients or shadows, it simply calculates the proper lighting value for each pixel. That's why today's GPUs have hundreds of those tiny processors. To do the same via polygons would require prohibitive amounts of memory and just as much processing, since you would wind up with multiple polys per pixel anyway.
The only thing that makes it "realtime" is that it has a relatively high redraw rate (for a raytracer). That's why there is a lot of fuzz when the camera pans around. It might render only a few thousand rays between redraws, which does give fast feedback but also slows down the rendering process overall. Most raytracers will churn 100k rays before updating the preview.
This is analogous to progressive jpeg decoding, where you start with a very chunky low-res preview and gradually work your way up to the full detail image. I don't see how this could be usable in games, unless you're specifically going for that TV-static effect:P
The problem with home purchases is they are inherently inflationary. The bank loans you $X, you spend $X, some broker pockets 5% of X, yet the house is still worth X. If you turn around and sell it tomorrow, you could get $X back, and pay another broker 5% of your X. The banks make a fortune on interest. The broker makes a fortune plastering his ugly anglo-saxon face on every bus shelter and lawn sign. The entire system is a feedback loop that will ultimately bankrupt everything.
Or do you actually believe it is fundamentally sound for a concrete and lumber box to take 20+ years for a couple to pay ? Did it take two people 20 years to build ? No ? Did it take 20 people two years to build ? Neither ? Then it's not worth 20 years of payments. The grossly unbalanced game of free market economics can only be played for so long; after a while, you run out of new people to extort.
Look, if they're going to offer subsidized internet access to low-income households, I think the real move should be to nationalize it altogether. If Comcast and/or the FCC can acknowledge that it is a public right to have affordable internet access for everyone, then it is high time the profit motive were removed from the equation. Oh, wait, you guys completely fucked that up with the national health care plan... carry on with your nihilistic ledger-padding then!
You're right, 10 watts is pennies, but in my experience the inefficiencies are several orders of magnitude greater than that.
One example I encountered last week, involved an accounting tool that queried an entire table, created deep entities for each record, and then performed various trivial sums and sorts on the entire dataset. It ran "well enough" for years, until I added this one client whose member list is about twenty times larger than all the others. Then it didn't work at all, ate up all available memory (48 gigabytes!) and died, because it was using O(n^2) memory and time. So I very easily rewrote it with a simple iterator, such that it uses O(1) memory and runs in O(n) time. So even those "well enough" clients saw a tremendous speedup, and as a result I was able to shut off all but two reporting servers. The original programmer chose the wrong method to traverse the data, probably built the whole thing in half a day without much thought, and for years this company kept throwing more hardware at it. Since they like HP, we're looking at about 60k worth of servers invalidated by about ten minutes of my time, including testing! The power savings pale in comparison to the hardware itself, and by the time they grow again, this gear will be a few years old and largely obsolete.
This low-hanging fruit is pervasive in the enterprise I.T. world. The same client had a $15k MySQL server that was barely delivering 50 queries per second. So they bought five more and set up a master/slave cluster to spread out the reads. Then when the company grew, they kept adding more slave nodes. I logged into one box, noticed it was using 128mb of the 24gb installed, so I tweaked the config to match the hardware, and threw in a shell script ot "preheat" the cache. Surprise, now one box handles 1500 QPS, with plenty of headroom. The (former) network guy thought it was normal performance, that their critical business app was just heavy. Perhaps if he had some concept of how fast the servers "should" have been working, he might have thought to investigate the slowness. Perhaps he didn't even know there was room for improvement. This is the point I'm trying to make, a lot of coders and sysadmins today don't even know what is possible, so how can they possibly know where and when to look for easy optimisations ?
Is it truly cheaper to be sloppy ? Hardware keeps getting cheaper and faster, sure, but not matching the pace at which code is getting slower.
Just look at your average web server. Today's machines are over a hundred times faster than they were 10 years ago, and we're not doing anything significantly different. Serving up text and graphics, processing forms, same old b.s. So then, why aren't we serving 100 times more pages per second ? Apache keeps getting fatter, PHP seems to do a "sleep(100)" after each line, and don't even get me started on Ruby.
There was a time, not so long ago, when I would spend an hour tweaking an oft-used assembler loop, and the end result was a 3x speedup or more. I'm not saying we should rewrite everything in assembler, but I think we're become so far removed from the actual machine, relying on the compiler to "do the right thing", that people don't even have the slighest clue how to distinguish fast code from slow. How often do we use benchmarks to test different solutions to the same problem ? Almost never! People bust out the profiler only when things go wrong, and even then they might say "just add CPU/Ram/SSD" and call it a day.
Or, if we must motivate the hippies, call it "green optimisation". Yes, faster code finishes quicker, using less power to complete the same job. If we're dealing with web servers, faster code would require less cluster nodes, or maybe free up some CPU time for another VM on the host, and those 60A circuits aren't cheap either. If spending an extra day optimizing my code could save me $2000 / mo off my colo bill, I'd be a fool not to invest that time.
Again, Ciscos can do this. I don't care much for the company, but I've a client with more money than brains and they have a HUGE deployment of these things. The WiFi is actually faster than the wired lan, despite having 300+ clients.
So are they testing 4000 versions of the code at any given time ?
I'm fine with a few hundred VMs, heck I have close to a hundred between my desktop and dev servers at home. Several thousand though, for a one-trick pony like eBay, seems gratuitous. Have they even changed much in the last ten years ? Growth, sure, but not many customer-facing changes.
802.11n already does this, they call it "beam-forming". Cisco features it in their high-end access points, using multiple antennae to send the same payload but with varying phase shift, which recombine at the receiver to produce a stronger coherent wave.
I love how the summary introduces him as the "inventor of the Quicktime codec". Yeah, he provided the RPZA ("road pizza") codec, which is so damn simple it made Bink Video look like fine art, back in the day.
We were combining multiple languages long before Windows even existed. Ever heard of a sneaky little program called a linker ?
It was just as "effortless" as your beloved.NET. You simply wrote a makefile that called the various compilers for all your bits of code, then linked them all together to produce the binary. Even if you use only one language, the linker is still used to hook up with system-wide libraries, and generate proper signatures and stubs.
And if you're about to say that makefiles are not effortless, well what do you think your shiny $1500 IDE does when you click "Build" ? Chances are, it produces a Makefile from all the metadata in your project and hands it to a hidden command-line build tool.
The one great strength Firefox has is its plugin system. Can the browser people please stop fucking with the interface and concentrate on the damn HTML renderer ? If I were to get hit in the head with the douche stick and suddenly wanted my Firefox to look like Chrome, I'd install a Chrome skin. If they're so obsessed with minimalist interfaces, fire the interface guy and let him release the dumb thing as a 3rd party plugin.
Let me get this straight. The gimmick is you rent out your place to a total stranger, you don't even meet them face-to-face, and expect them not to run away with all your phat loot ? Moronic. Hotels don't trust them anywhere near as much. They sure as shit don't leave anything of real value in closets, despite the cameras on each floor and at all exits.
What happened to EJ is truly vile, but what the fuck was she expecting ? She probably felt generous thinking 3% of Airbnb users would be vile, but she got the math wrong. Yes, 3% might be wanted criminals, but then about 90% are opportunist scum, and the remaining 7% are people like EJ with their heads in the clouds. All the locks and home insurance in the world are pointless if you're handing your keys to any stranger with a credit card.
After reading that post, I almost think she was asking for it, that it was all a set-up to show how dangerous this thing can be. Heck, I could do the same: I'll just write my door code on the lock itself, then leave for a week. By the time I return, I guarantee you there will be nothing left of my apartment, not even the fancy lock! They'll even smile at my cameras as they walk out with my used underwear.
Inventing a farcical business model, backed by a handful of dot-com profiteers is not going to change the fact that people are, by default, selfish, destructive, competitive swine until proven otherwise. People are greedy little shits, and nothing is going to change that as long as we worship possessions and wealth.
Slashdot Mirror
I think you meant to say "... performs better with larger files".
You nailed it though. My big gripe with Windows it how it seems to spend more time fiddling with metadata / directory entries than the actual contents. On an SSD with 700mb/sec writes and 0.1 msec access times, I'd expect it to churn through a few thousand files per second at the very least. That's not even factoring the disk cache. All those MFT updates seem to drag it right back down to spinning-disk speeds when dealing with numerous small files. You know, like a source tree or a directory full of images.
As sequential storage performance continues to improve, filesystem overhead is becoming the primary bottleneck.
As they should be. Once you acknowledge the fact that money is an artificial construct, the only realization is that accountants truly create nothing in the enterprise. They don't produce a saleable asset. They don't offer any services to the clients.
If you run a company without an accountant, the only bad thing that will happen is the tax man will get angry.
If you run a company without software, you have no company.
Correction: Bethesda's lawyers are being total assholes here.
Nothing surprising, I'm sure most law firms reject any résumé that does not include "Being a stubborn asshole" in the first paragraph.
Techreport seems to be slashdotted. Am I being too harsh, or is that horribly embarrassing for a site that focuses on performance testing and overclocking ?
Just sayin...
It took the game's consolified low-poly meshes and prorgammatically inflated the poly count, barely enhancing the image quality at all, while driving processing requirements way up. It looks more like one of those "repeat N 1000 times" benchmarks, the kind I write when trying to find bottlenecks in a web page, than any sort of effort to make Crysis 2 not look like the steaming EA-published turd it is.
Crysis 1 still looks better than this half-assed sequel, DX11 be damned.
If this is some bizarre partnership between EA and NVidia to push GPUs, shame on them both. I've come to expect this from EA, by far the greediest chop-shop in the industry, but NVidia... come on, have some self-respect. Treating your customers like blind idiots is NOT going to help move $1500 quad-SLI setups. Between these stunts, and ATI's stagnant driver performance, they are just creating a giant opening for a new contender to come hoover up their marketshare.
Sure, you can do shading like that, if you're only using flat shading - a very ancient technique that was made virtually obsolete when pixel shaders hit the mainstream. A proper shading algorithm does not need distinct polys to apply gradients or shadows, it simply calculates the proper lighting value for each pixel. That's why today's GPUs have hundreds of those tiny processors. To do the same via polygons would require prohibitive amounts of memory and just as much processing, since you would wind up with multiple polys per pixel anyway.
The only thing that makes it "realtime" is that it has a relatively high redraw rate (for a raytracer). That's why there is a lot of fuzz when the camera pans around. It might render only a few thousand rays between redraws, which does give fast feedback but also slows down the rendering process overall. Most raytracers will churn 100k rays before updating the preview.
This is analogous to progressive jpeg decoding, where you start with a very chunky low-res preview and gradually work your way up to the full detail image. I don't see how this could be usable in games, unless you're specifically going for that TV-static effect :P
The words do whatever I tell them to do.
Assembler, motherfucker. That's what we called it in the 80's while you were just a glint in James Gosling's lazy eye.
I'm going to laugh that one into the weekend... Cheers mate!
The problem with home purchases is they are inherently inflationary. The bank loans you $X, you spend $X, some broker pockets 5% of X, yet the house is still worth X. If you turn around and sell it tomorrow, you could get $X back, and pay another broker 5% of your X. The banks make a fortune on interest. The broker makes a fortune plastering his ugly anglo-saxon face on every bus shelter and lawn sign. The entire system is a feedback loop that will ultimately bankrupt everything.
Or do you actually believe it is fundamentally sound for a concrete and lumber box to take 20+ years for a couple to pay ? Did it take two people 20 years to build ? No ? Did it take 20 people two years to build ? Neither ? Then it's not worth 20 years of payments. The grossly unbalanced game of free market economics can only be played for so long; after a while, you run out of new people to extort.
Look, if they're going to offer subsidized internet access to low-income households, I think the real move should be to nationalize it altogether. If Comcast and/or the FCC can acknowledge that it is a public right to have affordable internet access for everyone, then it is high time the profit motive were removed from the equation. Oh, wait, you guys completely fucked that up with the national health care plan... carry on with your nihilistic ledger-padding then!
You're right, 10 watts is pennies, but in my experience the inefficiencies are several orders of magnitude greater than that.
One example I encountered last week, involved an accounting tool that queried an entire table, created deep entities for each record, and then performed various trivial sums and sorts on the entire dataset. It ran "well enough" for years, until I added this one client whose member list is about twenty times larger than all the others. Then it didn't work at all, ate up all available memory (48 gigabytes!) and died, because it was using O(n^2) memory and time. So I very easily rewrote it with a simple iterator, such that it uses O(1) memory and runs in O(n) time. So even those "well enough" clients saw a tremendous speedup, and as a result I was able to shut off all but two reporting servers. The original programmer chose the wrong method to traverse the data, probably built the whole thing in half a day without much thought, and for years this company kept throwing more hardware at it. Since they like HP, we're looking at about 60k worth of servers invalidated by about ten minutes of my time, including testing! The power savings pale in comparison to the hardware itself, and by the time they grow again, this gear will be a few years old and largely obsolete.
This low-hanging fruit is pervasive in the enterprise I.T. world. The same client had a $15k MySQL server that was barely delivering 50 queries per second. So they bought five more and set up a master/slave cluster to spread out the reads. Then when the company grew, they kept adding more slave nodes. I logged into one box, noticed it was using 128mb of the 24gb installed, so I tweaked the config to match the hardware, and threw in a shell script ot "preheat" the cache. Surprise, now one box handles 1500 QPS, with plenty of headroom. The (former) network guy thought it was normal performance, that their critical business app was just heavy. Perhaps if he had some concept of how fast the servers "should" have been working, he might have thought to investigate the slowness. Perhaps he didn't even know there was room for improvement. This is the point I'm trying to make, a lot of coders and sysadmins today don't even know what is possible, so how can they possibly know where and when to look for easy optimisations ?
What ever happened to good old fashioned public execution ? This guy has a white-crime rap sheet longer than most wall street traders.
Just off him already. Maybe that will scare some of the other spammers away.
Is it truly cheaper to be sloppy ? Hardware keeps getting cheaper and faster, sure, but not matching the pace at which code is getting slower.
Just look at your average web server. Today's machines are over a hundred times faster than they were 10 years ago, and we're not doing anything significantly different. Serving up text and graphics, processing forms, same old b.s. So then, why aren't we serving 100 times more pages per second ? Apache keeps getting fatter, PHP seems to do a "sleep(100)" after each line, and don't even get me started on Ruby.
There was a time, not so long ago, when I would spend an hour tweaking an oft-used assembler loop, and the end result was a 3x speedup or more. I'm not saying we should rewrite everything in assembler, but I think we're become so far removed from the actual machine, relying on the compiler to "do the right thing", that people don't even have the slighest clue how to distinguish fast code from slow. How often do we use benchmarks to test different solutions to the same problem ? Almost never! People bust out the profiler only when things go wrong, and even then they might say "just add CPU/Ram/SSD" and call it a day.
Or, if we must motivate the hippies, call it "green optimisation". Yes, faster code finishes quicker, using less power to complete the same job. If we're dealing with web servers, faster code would require less cluster nodes, or maybe free up some CPU time for another VM on the host, and those 60A circuits aren't cheap either. If spending an extra day optimizing my code could save me $2000 / mo off my colo bill, I'd be a fool not to invest that time.
That lib requires cooperative event handling in the kid class. I much prefer the longer, but deterministic form:
if ( $myLawn->getContents()->filter({type: kid})->count() > 0 ) {
$myShotgun = new Shotgun()
$myShotgun->loadAmmo();
$myLawn->getOwner()->wieldObject($myShotgun);
for( $i = 5; $i>0; $i--) { sleep(1000); }
while ( $myLawn->getContents()->filter({type: kid})->count() > 0 ) {
$myShotgun->fire();
}
}
Again, Ciscos can do this. I don't care much for the company, but I've a client with more money than brains and they have a HUGE deployment of these things. The WiFi is actually faster than the wired lan, despite having 300+ clients.
This.
It's wine. Old wine. Probably tastes like ass.
Now if a truck full of Mill St. Tankhouse beer were to sink to the bottom of the river, I'm strapping on the O2 tank!
So are they testing 4000 versions of the code at any given time ?
I'm fine with a few hundred VMs, heck I have close to a hundred between my desktop and dev servers at home. Several thousand though, for a one-trick pony like eBay, seems gratuitous. Have they even changed much in the last ten years ? Growth, sure, but not many customer-facing changes.
802.11n already does this, they call it "beam-forming". Cisco features it in their high-end access points, using multiple antennae to send the same payload but with varying phase shift, which recombine at the receiver to produce a stronger coherent wave.
I love how the summary introduces him as the "inventor of the Quicktime codec". Yeah, he provided the RPZA ("road pizza") codec, which is so damn simple it made Bink Video look like fine art, back in the day.
We were combining multiple languages long before Windows even existed. Ever heard of a sneaky little program called a linker ?
It was just as "effortless" as your beloved .NET. You simply wrote a makefile that called the various compilers for all your bits of code, then linked them all together to produce the binary. Even if you use only one language, the linker is still used to hook up with system-wide libraries, and generate proper signatures and stubs.
And if you're about to say that makefiles are not effortless, well what do you think your shiny $1500 IDE does when you click "Build" ? Chances are, it produces a Makefile from all the metadata in your project and hands it to a hidden command-line build tool.
TFA reads like a thinly-veiled promo for Nimbus Data Systems, which I can only guess are pushing a Linux-based SAN appliance full of SSDs. Big whoop.
What I would love to know is: Why does eBay need 4000 VMs ?
They were scanning for a sunken ship carrying champagne... you know, because that's totally worth the man-years collectively wasted on this article.
The one great strength Firefox has is its plugin system. Can the browser people please stop fucking with the interface and concentrate on the damn HTML renderer ? If I were to get hit in the head with the douche stick and suddenly wanted my Firefox to look like Chrome, I'd install a Chrome skin. If they're so obsessed with minimalist interfaces, fire the interface guy and let him release the dumb thing as a 3rd party plugin.
Let me get this straight. The gimmick is you rent out your place to a total stranger, you don't even meet them face-to-face, and expect them not to run away with all your phat loot ? Moronic. Hotels don't trust them anywhere near as much. They sure as shit don't leave anything of real value in closets, despite the cameras on each floor and at all exits.
What happened to EJ is truly vile, but what the fuck was she expecting ? She probably felt generous thinking 3% of Airbnb users would be vile, but she got the math wrong. Yes, 3% might be wanted criminals, but then about 90% are opportunist scum, and the remaining 7% are people like EJ with their heads in the clouds. All the locks and home insurance in the world are pointless if you're handing your keys to any stranger with a credit card.
After reading that post, I almost think she was asking for it, that it was all a set-up to show how dangerous this thing can be. Heck, I could do the same: I'll just write my door code on the lock itself, then leave for a week. By the time I return, I guarantee you there will be nothing left of my apartment, not even the fancy lock! They'll even smile at my cameras as they walk out with my used underwear.
Inventing a farcical business model, backed by a handful of dot-com profiteers is not going to change the fact that people are, by default, selfish, destructive, competitive swine until proven otherwise. People are greedy little shits, and nothing is going to change that as long as we worship possessions and wealth.