Err, people work in towns. In the UK they frequently use public transport to get there. Cinemas are in town. People go to the cinema. People take work home often, or need to check something up after work.
People don't want to go back to work to pick up their laptop after going to the cinema. They want to go home or to the coffee shop or to the pub.
If you read the article, you would have seen that this type of scan requires a higher dose of radiation, as it is picking up the iodine in the bloodstream.
On the upside, they would have got some really good images.
and finding the entire hostname versus folders/files would be difficult for http URLs as the / delimiter is used throughout. You could use dotted notation:
"intelligence" isn't a (small) hardcoded list of executable names to enable this optimisation, including the industry leading benchmarking application.
This tweak gives Intel Integrated Graphics 20% higher score than ATI Integrated Graphics, despite performing half to one third as fast in real-world gaming tests.
Who is cheating? Intel.
Also the benchmark guidelines forbid it. Intel should add in the required "intelligence" and make it generic throughout their drivers.
Firstly, per-game optimisations are fine if the image quality is not affected. This is often done by all companies, tweaking shaders, etc, to run better on their hardware than the shader that was shipped with the game.
Secondly, offloading some of the graphics pipeline to an underutilised, over-powerful CPU seems logical when the graphics unit utilises unified shaders. In this case, there is a significant performance increase (from "dire" to "mostly naff") without quality dropping. Again, logically this seems fine to me, as long as this type of optimisation isn't targeted at specific titles.
In this case, it was, and not only was it a specific title, it was a benchmark. This is quite bad, and is a type of cheating that the other graphics companies got over a few years ago.
In addition it appears that in the real world of games, the benchmark over-reports the Intel capabilities by a factor of two anyway, even with the application specific optimisation enabled. I.e., the ATI part gets 30fps to the Intel's 10fps/15fps, despite the ATI part's 3dmark score equalling or being below the Intel results.
Showing that when you have low-end graphics, the benchmarks should be run at lower resolutions and quality settings to get a representative comparison against other similar hardware. Running a benchmark in non-typical settings (i.e., 1920x1200 for integrated graphics) can lead to some odd skewings.
It also confirmed how rubbish Intel's integrated graphics are.
Yeah, "stylus hand cramp" is a thing of the past with the iPhone, Android and the Palm Pre, yet the review states that most of the applications require the use of said implement. If that means that the damn thing is as unfriendly and frustrating as the WinMob devices I've used in the past (especially the PIM apps, which were so backwards I don't know how it even got mildly popular as a mobile OS).
The fact is that Microsoft need to remove the existing UI libraries and do what Apple did - create a Touch variant of their current libraries. I.e., a ".NET Touch". All packaged applications need to be implemented in this for consistency throughout the system.
However with Microsoft competing against itself in the mobile OS stakes - Pink Phone UI, Zune UI, WinMob UI, they haven't got a hope in hell of creating a single, decent, developer-friendly and attractive mobile interface.
Totally, I have an additional monitor at work which makes work doable. Just on the road it's awkward (e.g,. travelling to head office on the train). Still, it's not too often. But I do think that 1280x800 is too low a resolution for a 15" laptop still, hence it just gets used for web browsing and email. Why web browsing? The TFT is one of those that can only be seen within a range of 10 degrees from perpendicular, so nobody else can see what's on the screen...
And with the price of 22" to 24" monitors being so low, relative to history, no real excuse for not having one at home even if you're tied to a laptop.
Docking stations are not absolutely necessary to achieve this, but make it a lot more convenient.
I've found the hassle of getting the laptop to dock cleanly or release is roughly the same amount of time that it would take to plug in/out the two USB devices and display that you've mentioned. It also takes up too much space on your desk (the Dell POS dock work supplied extends backwards horribly, meaning you can't have the docked laptop and a keyboard in front of it!).
I can spare the five seconds to plug in a few devices on the desk. The only benefit is that finding the power cable that falls on the floor when unplugged isn't an issue with a dock which is always plugged in. The other argument is that the dock is designed to position the laptop in an ergonomic position - not that the brick of a Dell dock does that.
1280x800 on a 15" screen (my work laptop) is not enough resolution to actually do any modern development work in, and the pixels are just a little bit too large. 1440x900 might be okay. 1680x1050 might have too small pixels. But on a 17" display, 1680x1050 (or 1600x900 for the new 16:9 screens) sounds like a good resolution.
At least you can get your IDE on that without feeling cramped.
These videos are excellent examples of why the internet shouldn't move to be video based.
I don't mind having an additional video, but for the love of all that is decent, create a decent article around the video's contents first that we can all read quickly, without the "sorry, my home lab is a bit crap" filler that wastes time.
I wouldn't care if nVidia had said when asked, "yes, this is a mock-up of the final shipping card". Instead they denied it, apparently eight times, when asked directly before the article was published. At some point afterwards, they finally admitted it was a mock-up.
The real issue isn't that it is a mock-up, it is lying that it wasn't.
Why would they lie? Presumably to present an image that they're further advanced than they actually are, to get people to hold off buying cards now to see what the nVidia cards are like when released. Sadly these people who thought they would be waiting until December could be waiting until March. And people, once they've started waiting, tend to get doggedly determined to wait it out longer and longer, instead of biting the bullet once it looks like things aren't happening.
It surely has backfired for nVidia, at least in the techie part of the internet. And it's a shame, because Fermi looks to be pretty amazing in many ways.
That's why I said for "traditional gaming graphics" as the ATI arrangement is optimised for that application, i.e., that's probably where an ATI card gets closest to peak performance.
The end of the motherboard was roughly dremmelled off to match the fan enclosure (that is surely the designed fan enclosure for the card). The power connectors were glued on, and didn't match the solder pads for said connectors (indeed one was mostly sawed off).
Prototype? No. This card can't work. Blatant fake presented as a working board? Yes. Back-pedalling and claiming it is a mock up after the fact? Yes.
It wasn't described as a mock up, but as a real working Fermi board.
NVIDIA are quite a way behind in the next generation race (time-wise, not tech-wise), and they had to try and make it look like they were a month or two away from having product availability. This fakery just makes the late Q1 2010 rumours sound more likely...
It also supports C++ and Fortran programming directly - probably with some caveats but nothing major.
Yeah, it started off with tricking the GPU, but soon after NVIDIA and ATI released CUDA and CTM, and now they're supporting a wealth of different APIs (NVIDIA will soon be ahead with Fermi considering the C, C++ and Fortran support, and the developer tools). AMD will probably have a new architecture late next year too (RV870 is a tweaked, enhanced RV770, which was a tweaked, enhanced R600, a good path for them to take, but they can't do it forever).
Well with Fermi starting at around 700 GFLOPS (double precision) in early 2010, and AMD's HD5890 probably achieving 600 GLFOPS, there's still a big benefit for using the $400 - $500 graphics card over three $1000 CPUs.
Both ATI and NVIDIA's cards already support FMA for single and double precision calculations, and thus gain some numerical accuracy over the CPU that doesn't support it!
The AMD HD5000 series is IEEE754-2008 compliant. Fermi will be similarly compliant when it arrives in three or four months. Both support FMA for both 64 and 32 bit operations - FMA has even more numerical accuracy than MAD because it is a fused operation with no intermediate storage/loss of accuracy.
Fermi will have more DP power than HD5000, but less SP.
Fermi also has some extra benefits - ECC throughout the memory hierarchy for example.
The RWT article (linked from the end of the Tech Report article) is also a good read.
Fermi has around 750 GFLOPS (clock speeds weren't released as the product is still around three months away from retail) of *double precision* power. It'll do ray tracing just fine. Indeed real time ray tracing was demonstrated yesterday with Fermi. I definitely see some ray tracing engines getting some CUDA/OpenCL love in the very near future now that DP is so fast.
It's also likely that the 5870 and 5850 could be faster than Fermi for traditional gaming graphics, as these don't need double precision, and the 5000 series has vast single precision power (2.7 TFLOPS for the 5870, 3 TFLOPS for the eventual 5890 that'll emerge when Fermi appears).
The article talks about 512MB chips with a 16ns switching time. Assuming you switch multiple cells in parallel, from a word if you're using it as a memory, more if used as storage, and it has a 16ns cell switching time. Writing bit by bit would allow a 7.4MB/s write speed. Writing 64-bits at a time ups this to nearly 500MB/s. It's not a replacement for DDR3, not even mobile DDR, but could be used as an intermediary memory level between RAM and storage, or the storage itself once capacities increase (when it's made on sub-40nm processes).
The Wiki page talks about a 2006 64MB/512Mb Samsung PCRAM production, which is probably where you're getting the 64MB figure from.
From a cursory overview, it really looks like PCRAM is going to be big, in a few years, as a flash replacement that is simply better in every way. We will see what happens down the line of course, but the development is very promising.
Slashdot Mirror
Err, people work in towns. In the UK they frequently use public transport to get there.
Cinemas are in town.
People go to the cinema.
People take work home often, or need to check something up after work.
People don't want to go back to work to pick up their laptop after going to the cinema. They want to go home or to the coffee shop or to the pub.
Yeah, all good points. I did mainly start to introduce more slashes in though in a mildly sarcastic manner for the original poster ;-)
And it's academic now because all reasonable software will handle the faff with URLs.
If you read the article, you would have seen that this type of scan requires a higher dose of radiation, as it is picking up the iodine in the bloodstream.
On the upside, they would have got some really good images.
More likely two screens of 480x320 to 640x480, with 4xAA and decent filtering.
Unless they're dumping the dual-screen concept, but it's worked well so far.
Rubbish. More / are needed. Maybe even some backslashes.
http://org/slashdot/tech/article.pl/sid=09\10\14\1219215
Such perfection! You drill down from the root. Also see:
file:/home/hattig/filename.txt
Works! Lovely.
Of course you'd have to stick the port elsewhere:
http:8080/com/example/www/news/2009101414450
and finding the entire hostname versus folders/files would be difficult for http URLs as the / delimiter is used throughout. You could use dotted notation:
https:443/com.example.www/news/2009101414450
Or make DNS return the longest match for the entire URL.
All academic of course, and probably with just as many problems down the line.
"intelligence" isn't a (small) hardcoded list of executable names to enable this optimisation, including the industry leading benchmarking application.
This tweak gives Intel Integrated Graphics 20% higher score than ATI Integrated Graphics, despite performing half to one third as fast in real-world gaming tests.
Who is cheating? Intel.
Also the benchmark guidelines forbid it. Intel should add in the required "intelligence" and make it generic throughout their drivers.
This is the conundrum here.
Firstly, per-game optimisations are fine if the image quality is not affected. This is often done by all companies, tweaking shaders, etc, to run better on their hardware than the shader that was shipped with the game.
Secondly, offloading some of the graphics pipeline to an underutilised, over-powerful CPU seems logical when the graphics unit utilises unified shaders. In this case, there is a significant performance increase (from "dire" to "mostly naff") without quality dropping. Again, logically this seems fine to me, as long as this type of optimisation isn't targeted at specific titles.
In this case, it was, and not only was it a specific title, it was a benchmark. This is quite bad, and is a type of cheating that the other graphics companies got over a few years ago.
In addition it appears that in the real world of games, the benchmark over-reports the Intel capabilities by a factor of two anyway, even with the application specific optimisation enabled. I.e., the ATI part gets 30fps to the Intel's 10fps/15fps, despite the ATI part's 3dmark score equalling or being below the Intel results.
Showing that when you have low-end graphics, the benchmarks should be run at lower resolutions and quality settings to get a representative comparison against other similar hardware. Running a benchmark in non-typical settings (i.e., 1920x1200 for integrated graphics) can lead to some odd skewings.
It also confirmed how rubbish Intel's integrated graphics are.
Yeah, "stylus hand cramp" is a thing of the past with the iPhone, Android and the Palm Pre, yet the review states that most of the applications require the use of said implement. If that means that the damn thing is as unfriendly and frustrating as the WinMob devices I've used in the past (especially the PIM apps, which were so backwards I don't know how it even got mildly popular as a mobile OS).
The fact is that Microsoft need to remove the existing UI libraries and do what Apple did - create a Touch variant of their current libraries. I.e., a ".NET Touch". All packaged applications need to be implemented in this for consistency throughout the system.
However with Microsoft competing against itself in the mobile OS stakes - Pink Phone UI, Zune UI, WinMob UI, they haven't got a hope in hell of creating a single, decent, developer-friendly and attractive mobile interface.
Totally, I have an additional monitor at work which makes work doable. Just on the road it's awkward (e.g,. travelling to head office on the train). Still, it's not too often. But I do think that 1280x800 is too low a resolution for a 15" laptop still, hence it just gets used for web browsing and email. Why web browsing? The TFT is one of those that can only be seen within a range of 10 degrees from perpendicular, so nobody else can see what's on the screen...
And with the price of 22" to 24" monitors being so low, relative to history, no real excuse for not having one at home even if you're tied to a laptop.
Yeah, that's exactly what I want.
Never mind the people that can't watch videos at work (even during lunchtimes, etc) due to the audio aspect.
Docking stations are not absolutely necessary to achieve this, but make it a lot more convenient.
I've found the hassle of getting the laptop to dock cleanly or release is roughly the same amount of time that it would take to plug in/out the two USB devices and display that you've mentioned. It also takes up too much space on your desk (the Dell POS dock work supplied extends backwards horribly, meaning you can't have the docked laptop and a keyboard in front of it!).
I can spare the five seconds to plug in a few devices on the desk. The only benefit is that finding the power cable that falls on the floor when unplugged isn't an issue with a dock which is always plugged in. The other argument is that the dock is designed to position the laptop in an ergonomic position - not that the brick of a Dell dock does that.
1280x800 on a 15" screen (my work laptop) is not enough resolution to actually do any modern development work in, and the pixels are just a little bit too large. 1440x900 might be okay. 1680x1050 might have too small pixels. But on a 17" display, 1680x1050 (or 1600x900 for the new 16:9 screens) sounds like a good resolution.
At least you can get your IDE on that without feeling cramped.
These videos are excellent examples of why the internet shouldn't move to be video based.
I don't mind having an additional video, but for the love of all that is decent, create a decent article around the video's contents first that we can all read quickly, without the "sorry, my home lab is a bit crap" filler that wastes time.
I wouldn't care if nVidia had said when asked, "yes, this is a mock-up of the final shipping card". Instead they denied it, apparently eight times, when asked directly before the article was published. At some point afterwards, they finally admitted it was a mock-up.
The real issue isn't that it is a mock-up, it is lying that it wasn't.
Why would they lie? Presumably to present an image that they're further advanced than they actually are, to get people to hold off buying cards now to see what the nVidia cards are like when released. Sadly these people who thought they would be waiting until December could be waiting until March. And people, once they've started waiting, tend to get doggedly determined to wait it out longer and longer, instead of biting the bullet once it looks like things aren't happening.
It surely has backfired for nVidia, at least in the techie part of the internet. And it's a shame, because Fermi looks to be pretty amazing in many ways.
That's why I said for "traditional gaming graphics" as the ATI arrangement is optimised for that application, i.e., that's probably where an ATI card gets closest to peak performance.
What about the motherboard being sawn short, all the ports being glued on, NVIDIA saying it was a mockup after this report came out, etc?
The end of the motherboard was roughly dremmelled off to match the fan enclosure (that is surely the designed fan enclosure for the card). The power connectors were glued on, and didn't match the solder pads for said connectors (indeed one was mostly sawed off).
Prototype? No. This card can't work.
Blatant fake presented as a working board? Yes.
Back-pedalling and claiming it is a mock up after the fact? Yes.
It wasn't described as a mock up, but as a real working Fermi board.
NVIDIA are quite a way behind in the next generation race (time-wise, not tech-wise), and they had to try and make it look like they were a month or two away from having product availability. This fakery just makes the late Q1 2010 rumours sound more likely...
Fermi supports IEEE-754.
It also supports C++ and Fortran programming directly - probably with some caveats but nothing major.
Yeah, it started off with tricking the GPU, but soon after NVIDIA and ATI released CUDA and CTM, and now they're supporting a wealth of different APIs (NVIDIA will soon be ahead with Fermi considering the C, C++ and Fortran support, and the developer tools). AMD will probably have a new architecture late next year too (RV870 is a tweaked, enhanced RV770, which was a tweaked, enhanced R600, a good path for them to take, but they can't do it forever).
Well with Fermi starting at around 700 GFLOPS (double precision) in early 2010, and AMD's HD5890 probably achieving 600 GLFOPS, there's still a big benefit for using the $400 - $500 graphics card over three $1000 CPUs.
Both ATI and NVIDIA's cards already support FMA for single and double precision calculations, and thus gain some numerical accuracy over the CPU that doesn't support it!
The AMD HD5000 series is IEEE754-2008 compliant. Fermi will be similarly compliant when it arrives in three or four months. Both support FMA for both 64 and 32 bit operations - FMA has even more numerical accuracy than MAD because it is a fused operation with no intermediate storage/loss of accuracy.
Fermi will have more DP power than HD5000, but less SP.
Fermi also has some extra benefits - ECC throughout the memory hierarchy for example.
The RWT article (linked from the end of the Tech Report article) is also a good read.
Fermi has around 750 GFLOPS (clock speeds weren't released as the product is still around three months away from retail) of *double precision* power. It'll do ray tracing just fine. Indeed real time ray tracing was demonstrated yesterday with Fermi. I definitely see some ray tracing engines getting some CUDA/OpenCL love in the very near future now that DP is so fast.
It's also likely that the 5870 and 5850 could be faster than Fermi for traditional gaming graphics, as these don't need double precision, and the 5000 series has vast single precision power (2.7 TFLOPS for the 5870, 3 TFLOPS for the eventual 5890 that'll emerge when Fermi appears).
AMD is also pushing the Bullet physics API running via OpenCL as a non-proprietary way forward.
How does that compare to cmol/LoCm (light speed moles per Library of Congress Metre).
The article talks about 512MB chips with a 16ns switching time. Assuming you switch multiple cells in parallel, from a word if you're using it as a memory, more if used as storage, and it has a 16ns cell switching time. Writing bit by bit would allow a 7.4MB/s write speed. Writing 64-bits at a time ups this to nearly 500MB/s. It's not a replacement for DDR3, not even mobile DDR, but could be used as an intermediary memory level between RAM and storage, or the storage itself once capacities increase (when it's made on sub-40nm processes).
The Wiki page talks about a 2006 64MB/512Mb Samsung PCRAM production, which is probably where you're getting the 64MB figure from.
From a cursory overview, it really looks like PCRAM is going to be big, in a few years, as a flash replacement that is simply better in every way. We will see what happens down the line of course, but the development is very promising.