To me, Windows 95 was when Microsoft turned from a somewhat serious business (cheap, shoddy, but mostly adequate software at the time) into a cult. It was all about the new and shiny, while the true geeks could see beyond the hype. Eventually, Windows became the boring business status quo, but at the time it was not so different from what Apple is now.
It's not a compatibility layer. It's that programs compiled against 32 bit shared libraries need those 32 bit shared libraries to run, and this is going to be difficult to get past, short of making everything a static binary.
True, it's not like emulating another architecture, because the CPU can run in different modes. But you still need some level of OS support for this mode switch.
I thought the case with 16-bit on 32-bit Windows was exactly the same, since the CPU can run in 16-bit mode as well. Or is there a key difference?
A few years ago in Finland, an Information Society Party was being organized, but it did not go very far. Both the name and the idea were considered bland, as most other parties had something to say about the information society as well.
However, in the past two years or so, PP Finland has made great progress. It has been registered for the upcoming parliamentary election, and its number of members puts it in the same league with the smallest parties currently in parliament.
Besides the catchier name, it must be said that the political climate has changed as well. Finland used to be one of the least corrupt countries in the world, but the past year or so has revealed a huge amount of corruption in the leading parties, including the former PM himself.
The idea of an integrity party and other alternatives are thus gaining interest.
Of course, privacy and freedom of speech have also become increasingly important, for example due to our Internet censorship.
Going to 64 bit means your instructions will be 64 bit, which means doubling the cache mem usuage.
Depending on how the os/app uses the cache you may even find an slow down on performance.
AFAIK, cache is also used for data, and data is not doubled in size by the architecture change. But I agree that memory/cache usage is somewhat increased when going from 32 to 64 bits.
However, x86-64 has other improvements over 32-bit x86, so you often get a net speedup. Other posters have already mentioned extra registers and guaranteed SSE2.
In other architectures, such as PPC, there are none of these extra improvements. Thus many people choose to run a 32-bit userland on 64-bit PPC (usually with a 64-bit kernel).
32-bit Windows had extra layers to support 16-bit binaries. So this kind of backwards compatibility is not a new idea, but it should not be needed in open source at all. I personally like having a clean system with a single architecture.
Intel graphics also have opensource drivers, including the necessary kernel bits in the vanilla Linux tree. They have been this way for years, and have exposed most of the functionality, unlike the open ATI/AMD drivers. Though it must be said that the latter are improving a lot; I'm currently typing on a Powerbook with Radeon graphics running Gentoo. For some reason, binary Linux PPC drivers are hard to find;)
For my uses, Intel GPUs have been powerful enough for years, and this includes HD video playback (using OpenGL for video and subs on separate planes) and simple molecular modelling. An i855 laptop from 2005 had slight issues with HD video, but an i945 is doing fine, and even that is ancient today.
Recently I have become interested in GPGPU, so I'm currently looking for a quiet Mini-ITX board with a full-size PCIe slot. So far only the VIA EPIA VB8001 has caught my attention, but newer versions of the Nano processor should be on their way.
The Core Duo processor was basically a dual Pentium M with SSE3 on a single die. The Core microarchitecture, on the other hand, was found in Core 2 processors.
This naming scheme will probably continue as long as processors have cores. Makes you wonder how processors worked at all, before the Cores were introduced.
You can feel when you're spinning around, even though you could do a coordinate transformation where you are standing still. The explanation is that you're spinning relative to the rest of the universe.
I think the binary compatibility you mention is the key point. MS could release Windows for ARM, but most applications are closed source for x86. To put it bluntly, open source does not have this problem, and its users can change architectures based on actual technical merits.
Moreover, x86 is only cheap because these users of closed software demand it. If you look at actual manufacturing cost at equal volumes, ARM should be considerably cheaper due to smaller die sizes and simpler/older processes.
I agree that Intel has been really, really good at improving x86. Imagine what Intel could do with a cleaner architecture like ARM.
The processor isn't going to be a substantial proportion of the power usage on a full-scale system like that; it's not (commercially) worth going non-x86 for the proportionally small savings. VIA makes systems very much like what you're describing, but they run x86.
Silence is another desirable trait, besides power consumption. For example, a VIA Mini-ITX systems spends most of its power in the small CPU, which needs a whiny fan. So even if that power is less than half of the total, it is still worth reducing the concentration of heating power. ARM's lack of heatsinks and fans also means interesting possibilities for case design.
I think x86 compatibility is a much bigger issue. People "choose" x86 because it "needs" to run Windows.
But really, I wonder if the article person confuses (sampling) frequency with bandwidth? Not that they used to be sampled, but that's a common measure. In that case, these would be equivalent to 6.6 kHz sampling rate. I know, like it makes a big difference. But still...
Many people seem to confuse bandwidth with data transmission rate, so you can imagine the confusion if the author had actually used the correct terminology.
My thoughts exactly. One problem is that non-technical people tend to use "bandwidth" as a synonym for channel capacity (data transfer rate), instead of its original meaning (the width of a range of frequencies). So perhaps it is safer to avoid using that word altogether.
There was also Tegra-based Toshiba AC100 on display, with a lot more nerd-appeal IMHO.
I might buy one of these, if you can install a proper distro instead of this silly little smartphone OS. Tegra as such is relatively open, and there are instructions on getting Ubuntu and Gentoo on the devkit, but the netbook may be a different matter.
Here "relatively open" means that you need to run Nvidia's own kernel branch based on 2.6.29, and the video drivers seem to be closed blobs. I think
I'll get a GNU/RMS-approved Lemote Yeeloong instead.
Slashdot Mirror
I always thought it as Dr. Watson having constipation.
To me, Windows 95 was when Microsoft turned from a somewhat serious business (cheap, shoddy, but mostly adequate software at the time) into a cult. It was all about the new and shiny, while the true geeks could see beyond the hype. Eventually, Windows became the boring business status quo, but at the time it was not so different from what Apple is now.
In other news, Finland embargos Rare Exports to the rest of the world...
The key is "unencrypted". From http://www.blackmagic-design.com/products/intensity/techspecs/ :
"For legal reasons HDMI input is unable to capture from copy protected HDMI sources."
It's not a compatibility layer. It's that programs compiled against 32 bit shared libraries need those 32 bit shared libraries to run, and this is going to be difficult to get past, short of making everything a static binary.
True, it's not like emulating another architecture, because the CPU can run in different modes. But you still need some level of OS support for this mode switch.
I thought the case with 16-bit on 32-bit Windows was exactly the same, since the CPU can run in 16-bit mode as well. Or is there a key difference?
A few years ago in Finland, an Information Society Party was being organized, but it did not go very far. Both the name and the idea were considered bland, as most other parties had something to say about the information society as well.
However, in the past two years or so, PP Finland has made great progress. It has been registered for the upcoming parliamentary election, and its number of members puts it in the same league with the smallest parties currently in parliament.
Besides the catchier name, it must be said that the political climate has changed as well. Finland used to be one of the least corrupt countries in the world, but the past year or so has revealed a huge amount of corruption in the leading parties, including the former PM himself. The idea of an integrity party and other alternatives are thus gaining interest. Of course, privacy and freedom of speech have also become increasingly important, for example due to our Internet censorship.
Going to 64 bit means your instructions will be 64 bit, which means doubling the cache mem usuage.
Depending on how the os/app uses the cache you may even find an slow down on performance.
AFAIK, cache is also used for data, and data is not doubled in size by the architecture change. But I agree that memory/cache usage is somewhat increased when going from 32 to 64 bits.
However, x86-64 has other improvements over 32-bit x86, so you often get a net speedup. Other posters have already mentioned extra registers and guaranteed SSE2.
In other architectures, such as PPC, there are none of these extra improvements. Thus many people choose to run a 32-bit userland on 64-bit PPC (usually with a 64-bit kernel).
32-bit Windows had extra layers to support 16-bit binaries. So this kind of backwards compatibility is not a new idea, but it should not be needed in open source at all. I personally like having a clean system with a single architecture.
Optus's ad campaign promises 'supersonic' speeds Well, I'd expect that. I wouldn't like a ping time of 6 seconds per kilometer distance!
Nobody expects a sonic transmission!
Intel graphics also have opensource drivers, including the necessary kernel bits in the vanilla Linux tree. They have been this way for years, and have exposed most of the functionality, unlike the open ATI/AMD drivers. Though it must be said that the latter are improving a lot; I'm currently typing on a Powerbook with Radeon graphics running Gentoo. For some reason, binary Linux PPC drivers are hard to find ;)
For my uses, Intel GPUs have been powerful enough for years, and this includes HD video playback (using OpenGL for video and subs on separate planes) and simple molecular modelling. An i855 laptop from 2005 had slight issues with HD video, but an i945 is doing fine, and even that is ancient today.
Recently I have become interested in GPGPU, so I'm currently looking for a quiet Mini-ITX board with a full-size PCIe slot. So far only the VIA EPIA VB8001 has caught my attention, but newer versions of the Nano processor should be on their way.
The Core Duo processor was basically a dual Pentium M with SSE3 on a single die. The Core microarchitecture, on the other hand, was found in Core 2 processors.
http://en.wikipedia.org/wiki/Core_(microarchitecture)
This naming scheme will probably continue as long as processors have cores. Makes you wonder how processors worked at all, before the Cores were introduced.
You can feel when you're spinning around, even though you could do a coordinate transformation where you are standing still. The explanation is that you're spinning relative to the rest of the universe.
http://en.wikipedia.org/wiki/Mach's_principle
In principle, a keyboard should be able to generate some electricity from the movement. But has anyone actually seen these IRL?
http://www.theregister.co.uk/1999/07/22/keyboard_generates_electricity_for_notebooks/
NOT TRUE!!! SCREAMING IN AN ARGUMENT MAKES YOU MORE RIGHT!!!
<Morbo> ARGUMENTS DO NOT WORK THAT WAY!
(This bit down here is to appease /.'s filters, as it thinks I'm yelling...no sense of humor at all.)
I think the binary compatibility you mention is the key point. MS could release Windows for ARM, but most applications are closed source for x86. To put it bluntly, open source does not have this problem, and its users can change architectures based on actual technical merits.
Moreover, x86 is only cheap because these users of closed software demand it. If you look at actual manufacturing cost at equal volumes, ARM should be considerably cheaper due to smaller die sizes and simpler/older processes.
I agree that Intel has been really, really good at improving x86. Imagine what Intel could do with a cleaner architecture like ARM.
The processor isn't going to be a substantial proportion of the power usage on a full-scale system like that; it's not (commercially) worth going non-x86 for the proportionally small savings. VIA makes systems very much like what you're describing, but they run x86.
Silence is another desirable trait, besides power consumption. For example, a VIA Mini-ITX systems spends most of its power in the small CPU, which needs a whiny fan. So even if that power is less than half of the total, it is still worth reducing the concentration of heating power. ARM's lack of heatsinks and fans also means interesting possibilities for case design.
I think x86 compatibility is a much bigger issue. People "choose" x86 because it "needs" to run Windows.
But really, I wonder if the article person confuses (sampling) frequency with bandwidth? Not that they used to be sampled, but that's a common measure. In that case, these would be equivalent to 6.6 kHz sampling rate. I know, like it makes a big difference. But still...
Many people seem to confuse bandwidth with data transmission rate, so you can imagine the confusion if the author had actually used the correct terminology.
My thoughts exactly. One problem is that non-technical people tend to use "bandwidth" as a synonym for channel capacity (data transfer rate), instead of its original meaning (the width of a range of frequencies). So perhaps it is safer to avoid using that word altogether.
Ha... if 3.3KHz was the accepted frequency all they'll hear probably sounds like bird chirping.
No, that 3.3 kHz would be a constant sine wave. Turning that sound on and off requires some bandwidth.
Nokia and Linus are Finnish. Microsoft is American.
I indented the code to make it readable. That's so obvious I don't need a comment to remind me.
(pun indented)
There was also Tegra-based Toshiba AC100 on display, with a lot more nerd-appeal IMHO.
I might buy one of these, if you can install a proper distro instead of this silly little smartphone OS. Tegra as such is relatively open, and there are instructions on getting Ubuntu and Gentoo on the devkit, but the netbook may be a different matter.
Here "relatively open" means that you need to run Nvidia's own kernel branch based on 2.6.29, and the video drivers seem to be closed blobs. I think I'll get a GNU/RMS-approved Lemote Yeeloong instead.
http://tegradeveloper.nvidia.com/tegra/forums/tegra-forums/linux-development
Since there isn't a video driver that uses the Cell's SPUs,
http://www.mesa3d.org/cell.html
Silicon diodes have a voltage drop of about 0.7 V. Most Nokias seem to charge at 3.7 V, so the power loss due to one diode is pretty significant.