My 2.5 year old Samsung 275T monitor is currently retailing at the same location for appox $75 more than I paid for it at purchase.
2.5 years old means it's probably not being produced anymore. That model no longer appears on Samsung's LCD monitors page and is sold by very few vendors.
In 30+ years of building systems I think that may be a first.
Pay more attention. I don't even waste my time in building systems and I've seen prices go up on unused, out-of-production hardware (new in box & sold by an authorized retailer).
Really? I seem to remember that the GMA 950 is more like a Voodoo3, lacking even hardware vertex processing. Or has GMA 950 been pretty much discontinued in favor of more powerful GMAs?
The GMA 950 was obsolete in 2006 (when the G965 and GM965 were released) for most applications - the exception being low-powered systems such as Atom. The Atom N450 integrated a newer GPU core (GMA 3150) into the CPU, and was released in December.
Where servers are concerned, conservatism is a virtue, and Debian Stable is my favourite brand of conservatism
Debian support for the previous stable releases is very weak: only one year after the next stable release. Assuming a stable release every two years, that's worse than Ubuntu LTS's 5 years for servers, and worse than RHEL's 7 year support.
I never understood backporting. Sure, it gives the illusion of stability, but you're relying on a much smaller set of developers, those for your OS, who may or may not understand the upstream code well enough to make smart decisions and having them glue code in and call it a "stable" release version.
Considering how much of the development actually comes from Red Hat, it's likely that at least they understand the upstream code well.
One reasons I will always pick Debian - stable over running Red Hat / CentOS with it's 3 year old versions of software that has "backported" fixes.
RHEL/CentOS software is only three years old if you're running a three year old version of RHEL/CentOS. The upcoming RHEL 6 will include newer software.
Red Hat probably spends more time testing their backported fixes than the upstream developers spend testing the original code.
I can also set it up so that, instead of using a swap file, it uses linux's disk swap partition. And that, I suspect, is there the performance comes from.
I've seen huge disk performance gains on NT6 just from running as a VMware ESXi guest with a single disk shared over iSCSI.
The 'swap file' is a truly horrible idea: a fragmentable file on top of a journaled filesystem, vs. a slice of the disk which can be used as directly addressable swap.
Swap partitions are wasteful and antiquated for desktop operating systems. Swap file fragmentation only happens when the page file grows beyond its initial allocation and there is insufficient contiguous space. A filesystem with online defragmentation and special allocation policies will not have any problems with swap file fragmentation.
HUGE difference in performance there
Citation definitely needed.
it's pathetic that MS still hasn't fixed this glaring inadequacy.
It's pathetic that swap partitions are still standard practice on desktop Linux. On systems with sufficient memory, swap space never gets used unless suspend-to-disk is used. And when using suspend-to-disk, the swap partition must be sufficiently large to write everything. The use of encrypted swap partitions (with a random key) also causes problems with hibernation.
Your CPU doesn't have VT-x extensions (hint: only the 3.6GHz "662" and 3.8GHz "672" Pentium 4's have it), so the answer is "yes, but it'll be slower than Christmas."
In VMware, 32-bit Windows guests run faster when not using the VT-x extensions. VMware will only use VT-x/AMD-V when it provides a performance benefit, or is necessary when using a 64-bit guest.
(This might not be the case on newer CPUs with extended/nested page tables.)
Yeah, that's what is sad and amazing. Windows running native on the box also has access to that 400MB of RAM to use as a disk cache, and in theory should do better because it can choose to use the memory for apps or cache depending where the pressure is.
The cache on the host is *in addition* to the cache in the guest. Performance gains can also come from the use of sparse virtual disks - less seek time may be required.
For example, during boot time, the host might load the entire kernel & drivers into cache while the bootloader is still loading it using slow BIOS calls.
Also, it's possible that the virtual disk drivers are lying about blocks being written to the physical disk.
...the amount of spam that actually makes it to an inbox, instead of being dumped into a junk folder or blocked outright?
I don't see much spam in my inbox, but I occasionally get lots of backscatter in my inbox - maybe 10 messages a day for a week, and then nothing for a few months.
In the words of the Wikipedians, you should assume good faith [wikipedia.org]. It is irrelevant whether or not he actually was "lying" unless you had very good reason to believe he was before your original post.
This isn't Wikipedia.
(Why would some random slashdot poster be deliberately trying to deceive people about the encoding of current blu-ray videos?)
Why would some random slashdot poster be deliberately trying to deceive people about the business practices of Microsoft?
No, just a little out of date. I only look at the HD world every once in a while.
Then why did you say "most current titles"?
Some do NOW. But they were originally built out as MPEG-2. Heck, DirectTV was still MPEG-1.5 only a couple of years ago because they didn't want to rip and replace the number of installed receivers required to upgrade.
And now DirecTV and Dish Network HD services are all H.264 as is the biggest IPTV service (U-verse).
Almost certainly as software running on the GPU, i.e. upgradable.
The GPUs have specific video decoding and processing hardware.
Regular iPods don't even have a publicly available SDK. Downloadable game support only came to the 5th generation iPod (2005), and only a few dozen games were ever released for it.
I don't know how much evidence there is of Apple abusing its iPod position. They haven't stopped other online music services from selling music from major labels. They haven't stopped other companies from making and selling MP3 players. iTunes Store sells unencrypted MPEG4 AAC music. The only thing they do is make a single program that syncs music with their music players, and this application only syncs with Apple devices.
Are you sure about that? After all, hardly anything has "hardware" TnL nowadays, even though everything supports it - the old calls go through a purpose-made shader program which emulates old style TnL. Similarly with video postprocessing on many current GPUs, running on shaders...so why not video decoding?
The lower-end cards aren't powerful enough to decode video and postprocess it, so just about all cards get dedicated decoding. Also, it is more efficient to use dedicated video processing that runs at a slower clock rate. The Radeon HD 2400 had partial MPEG-2 acceleration in shaders, and it was unable to both decode and deinterlace MPEG-2 at 1080i, yet could handle H.264 at 1080i just fine because it had a separate H.264 decoder.
Look up PureVideo or UVD or the recent video card block diagrams for more info.
You could still accelerate the parts of the decoding that you could. The older Intel 950 GPUs didn't accelerate all the functions in H.264, but did enough that the CPU could do the rest.
I am unaware of any H.264 decoding features in the GMA 950. Playback of H.264 on the GMA 950 is a function of CPU speed and memory bandwidth (I've successfully played 1080p on the GMA 900 and GMA 950).
The few GPUs that advertised partial H.264 decode acceleration provided a small reduction in CPU usage compared to those which include full bitstream decode acceleration.
Awesome, just fucking awesome... I never realized the Atom had *this much* power (and with that I mean processing, the rated power-use is like 8 watts). They should be at least 10 times more efficient as most desktop processors.
Atom isn't really that fast. Everything I've seen indicates that a single core 1.6 GHz Atom compares to something like a 600 MHz single Core 2 or 900 MHz Pentium M; a 25W TDP Core 2 Duo at 2.53 GHz is significantly faster and probably more efficient per watt.
Atom is designed to be *cheap*; Intel's best high-performance low-power CPUs are the Core ULV branded CPUs.
Dedicated DSP chips are common, but they are common in things like camcorders, Blu-Ray players, DirecTV receivers, cable set-top boxes, TiVos, TV tuner cards, etc. Such devices generally don't provide general-purpose web browsers.
And with H.264 these devices can play the same video streams as other devices - like many set-top boxes and TVs already can. They might stream them through a non-HTML interface, or via something like DNLA, or play via an attached USB device. And the devices that encode H.264 (camcorders, videoconference systems) can generate streams playable on other devices.
Most devices capable of web browsing have a full scale GPU. Even portable devices like the iPhone family or netbooks have GPUs to some degree. They don't generally use specialized single-codec decoder silicon.
Most of these devices do have specialized hardware to decode H.264. Even the programmable ones are designed to have enough power to decode H.264 at a given maximum specification. The video decoding on modern PC GPUs is a special-purpose part of the GPU, not an implementation of decoding using OpenCL or CUDA.
And most current titles use VC-1 and old ones used MPEG-2, not H.264.
You are lying.
Some DVB might have added H.264 in addition to MPEG-2 but when they launched they were also just MPEG-2.
HDTV DVB broadcasts use H.264, as do IPTV systems.
Netflix will use whatever is deployed as they aren't any sort of standard and don't really have dependencies on much physical hardware yet.
Netflix support has been in some Blu-ray players, TVs, and other devices for probably a year now, and now supports iPhone/iPad.
No, H.264 as a format that isn't just encapsulated in some other locked DRM hell like Flash, the Netflix player, a cable company settop box, etc. is almost entirely an Apple only thing at this point.
H.264 is used in almost every commercial video application. I use it every day without using Apple products. H.264 support is in my Intel, nVidia, and Intel GPUs, in my non-Apple phones, in non-Apple videoconferencing systems, in non-Apple Blu-ray players.
Slashdot Mirror
2.5 years old means it's probably not being produced anymore. That model no longer appears on Samsung's LCD monitors page and is sold by very few vendors.
Pay more attention. I don't even waste my time in building systems and I've seen prices go up on unused, out-of-production hardware (new in box & sold by an authorized retailer).
And the set of users who upgrade systems with integrated graphics is even smaller - that's what this socket is primarily used for.
Sure. And you can upgrade to a CPU about as fast as what you can install in a literally years old LGA 775 system...
The GMA 950 was obsolete in 2006 (when the G965 and GM965 were released) for most applications - the exception being low-powered systems such as Atom. The Atom N450 integrated a newer GPU core (GMA 3150) into the CPU, and was released in December.
Debian support for the previous stable releases is very weak: only one year after the next stable release. Assuming a stable release every two years, that's worse than Ubuntu LTS's 5 years for servers, and worse than RHEL's 7 year support.
Considering how much of the development actually comes from Red Hat, it's likely that at least they understand the upstream code well.
RHEL/CentOS software is only three years old if you're running a three year old version of RHEL/CentOS. The upcoming RHEL 6 will include newer software.
Red Hat probably spends more time testing their backported fixes than the upstream developers spend testing the original code.
I've seen huge disk performance gains on NT6 just from running as a VMware ESXi guest with a single disk shared over iSCSI.
Swap partitions are wasteful and antiquated for desktop operating systems.
Swap file fragmentation only happens when the page file grows beyond its initial allocation and there is insufficient contiguous space. A filesystem with online defragmentation and special allocation policies will not have any problems with swap file fragmentation.
Citation definitely needed.
It's pathetic that swap partitions are still standard practice on desktop Linux. On systems with sufficient memory, swap space never gets used unless suspend-to-disk is used. And when using suspend-to-disk, the swap partition must be sufficiently large to write everything.
The use of encrypted swap partitions (with a random key) also causes problems with hibernation.
Mac OS X uses swap files too.
In VMware, 32-bit Windows guests run faster when not using the VT-x extensions. VMware will only use VT-x/AMD-V when it provides a performance benefit, or is necessary when using a 64-bit guest.
(This might not be the case on newer CPUs with extended/nested page tables.)
The cache on the host is *in addition* to the cache in the guest. Performance gains can also come from the use of sparse virtual disks - less seek time may be required.
For example, during boot time, the host might load the entire kernel & drivers into cache while the bootloader is still loading it using slow BIOS calls.
Also, it's possible that the virtual disk drivers are lying about blocks being written to the physical disk.
I find this neither sad nor amazing. Host RAM can be used as disk cache.
This has an easy solution: tax the harmful products in proportion to the added cost of care. Tobacco products are already separately taxed.
Are you new here?
I don't see much spam in my inbox, but I occasionally get lots of backscatter in my inbox - maybe 10 messages a day for a week, and then nothing for a few months.
This isn't Wikipedia.
Why would some random slashdot poster be deliberately trying to deceive people about the business practices of Microsoft?
I don't know, but it happens a lot.
Except this isn't a bank account we are talking about here. This is a subject the poster admitted having no interest in.
Of course he knew his knowledge was out of date - he admitted to this, and yet originally said he was talking about the current situation.
Except where he admitted to using out of date knowledge to base his statement on what current titles use.
Then why did you say "most current titles"?
And now DirecTV and Dish Network HD services are all H.264 as is the biggest IPTV service (U-verse).
The GPUs have specific video decoding and processing hardware.
Is this still the case (at least for Safari users)?
Flash appears to run as a separate process in Safari 4, and even then I haven't seen any crashes of the plugin processes.
The lower-end cards aren't powerful enough to decode video and postprocess it, so just about all cards get dedicated decoding. Also, it is more efficient to use dedicated video processing that runs at a slower clock rate. The Radeon HD 2400 had partial MPEG-2 acceleration in shaders, and it was unable to both decode and deinterlace MPEG-2 at 1080i, yet could handle H.264 at 1080i just fine because it had a separate H.264 decoder.
Look up PureVideo or UVD or the recent video card block diagrams for more info.
I am unaware of any H.264 decoding features in the GMA 950. Playback of H.264 on the GMA 950 is a function of CPU speed and memory bandwidth (I've successfully played 1080p on the GMA 900 and GMA 950).
The few GPUs that advertised partial H.264 decode acceleration provided a small reduction in CPU usage compared to those which include full bitstream decode acceleration.
Atom isn't really that fast. Everything I've seen indicates that a single core 1.6 GHz Atom compares to something like a 600 MHz single Core 2 or 900 MHz Pentium M; a 25W TDP Core 2 Duo at 2.53 GHz is significantly faster and probably more efficient per watt.
Atom is designed to be *cheap*; Intel's best high-performance low-power CPUs are the Core ULV branded CPUs.
And with H.264 these devices can play the same video streams as other devices - like many set-top boxes and TVs already can. They might stream them through a non-HTML interface, or via something like DNLA, or play via an attached USB device. And the devices that encode H.264 (camcorders, videoconference systems) can generate streams playable on other devices.
Most of these devices do have specialized hardware to decode H.264. Even the programmable ones are designed to have enough power to decode H.264 at a given maximum specification. The video decoding on modern PC GPUs is a special-purpose part of the GPU, not an implementation of decoding using OpenCL or CUDA.
You are lying.
HDTV DVB broadcasts use H.264, as do IPTV systems.
Netflix support has been in some Blu-ray players, TVs, and other devices for probably a year now, and now supports iPhone/iPad.
H.264 is used in almost every commercial video application. I use it every day without using Apple products. H.264 support is in my Intel, nVidia, and Intel GPUs, in my non-Apple phones, in non-Apple videoconferencing systems, in non-Apple Blu-ray players.