File system stack in Windows is slow. No. It's SLOOOOOOWWWW.
It's sometimes _hundreds_ _of_ _times_ slower than in Linux. So it makes less sense to create ultra-fast filesystems if dispatching requests takes longer than time spent in your filesystem.
Also, IFS (Installable FileSystems) layer in Windows is heavily geared towards NTFS, so it's not easy to adapt other filesystems to it.
It makes little sense to compare ARM boards. They are usually power-optimized (with peripherals consuming substantial fraction of power), not performance-optimized. Also, end-user devices might be underclocked and undervoltaged. Besides, if you want performance with non-x86 CPUs, then look at MIPS CPUs (which are usually used for things like AV processing in set-top boxes).
Now, a comparison of end-used devices (netbooks/phones) with ARM/MIPS CPUs might actually make sense.
My old Siemens S65 can back up just fine (using IrDA or Bluetooth). Before that I had Siemens S30 (AFAIR) which also could be backed up just fine (over the serial port).
"Hell, many times putting a very dumb student in a class full of smart students improves the dumb student's grades, because good study habits rub off on them."
Erlang is quite OK for non-distributed programming. Its model of threads exchanging messages is just a natural fit for it. As it is for multicore systems.
PureMPI is _slow_, it's managed and it can't work with raw links.
Windows does _not_ support process migration, clustering services is a different beast (comparable to JBoss Clustering - http://www.jboss.org/jbossclustering/ ).
Also, Windows has no good clustered filesystems (like OCFS2 in Linux). And no analogs of DRBD.
So you can use Windows for parallel processing, but Linux is almost(?) always better for HPC.
"Try programming in linux for one day and then throw parallel processing into it. then throw in some data."
Funny, but I do this every day. Say, does Windows support zero-copy Infiniband links? How about MPI performance? How about fire&forget clustered processes?
Yes, you are correct. However, GDI is capable of accelerating only very basic operations (flood fill, bitblits). It can't accelerate antialiased line drawing and gradient fills.
And it's quite ironic, because internally GDI+ has a lot of hooks for hardware acceleration. Only no one has ever implemented it.
"The idea of fusion and benefits of fusion are tremendous compared to fossil fuels but I've always wondered how long will it last before it starts eating a significant enough portion of the hydrogen to be a concern."
If your fusion powerplants are eating a significant portion of Earth's hydrogen, then it's time to relocate somewhere where the temperature is not high enough to boil oceans.
"So, yes, most German soldiers didn't know jack squat about the extermination, and never rounded up anybody."
Yup. That's why the Wehrmacht happily committed heinous crimes in occupied territories (Russia, Poland, Ukraine). Mostly because Wehrmacht soldiers were not disciplined for crimes against local population.
"No, it isn't. It's within their competence to say that to reduce temperature by X degrees requires Y."
Yes. And the methods for these are mostly within the bounds of climatology. I.e. questions like 'by how much we must decrease CO2 to stop the warming' or 'would stratospheric aerosols help to reduce greenhous effect'.
Slashdot Mirror
File system stack in Windows is slow. No. It's SLOOOOOOWWWW.
It's sometimes _hundreds_ _of_ _times_ slower than in Linux. So it makes less sense to create ultra-fast filesystems if dispatching requests takes longer than time spent in your filesystem.
Also, IFS (Installable FileSystems) layer in Windows is heavily geared towards NTFS, so it's not easy to adapt other filesystems to it.
Nope. It's because UI _is_ a hard task to parallelize.
For example, a lot of programs are still single-threaded. Even the ones like Microsoft Word.
It makes little sense to compare ARM boards. They are usually power-optimized (with peripherals consuming substantial fraction of power), not performance-optimized. Also, end-user devices might be underclocked and undervoltaged. Besides, if you want performance with non-x86 CPUs, then look at MIPS CPUs (which are usually used for things like AV processing in set-top boxes).
Now, a comparison of end-used devices (netbooks/phones) with ARM/MIPS CPUs might actually make sense.
My old Siemens S65 can back up just fine (using IrDA or Bluetooth). Before that I had Siemens S30 (AFAIR) which also could be backed up just fine (over the serial port).
I bought both of them for less than $100.
Yep. And if you take several idiots they are likely to band together and start bullying other kids. Until other kids start imitating idiots.
A well known fact.
Or 'the idiot kid' might become an outcast.
PS: it's OK to mix kids of uneven capabilities, but they should not be VERY uneven.
"Hell, many times putting a very dumb student in a class full of smart students improves the dumb student's grades, because good study habits rub off on them."
Bullshit.
I'm trying to tell that servers are a special case. They are trivially parallelizeable.
We still have not solved more complex parallelization tasks. For example, GUI is still mostly single-threaded.
Erlang is quite OK for non-distributed programming. Its model of threads exchanging messages is just a natural fit for it. As it is for multicore systems.
Check Erlang ;)
Can Apache use 2 CPUs to serve one request? Well, I thought so.
Oracle, AFAIK, can't use several CPUs to service one query too (I might be wrong).
PureMPI is _slow_, it's managed and it can't work with raw links.
Windows does _not_ support process migration, clustering services is a different beast (comparable to JBoss Clustering - http://www.jboss.org/jbossclustering/ ).
Also, Windows has no good clustered filesystems (like OCFS2 in Linux). And no analogs of DRBD.
So you can use Windows for parallel processing, but Linux is almost(?) always better for HPC.
"Try programming in linux for one day and then throw parallel processing into it. then throw in some data."
Funny, but I do this every day. Say, does Windows support zero-copy Infiniband links? How about MPI performance? How about fire&forget clustered processes?
"Chutzpah" is better translated as "naglost'", definitely not "smelost'" ("bravery").
It's true. You can believe me, I'm from Russia.
Yes, you are correct. However, GDI is capable of accelerating only very basic operations (flood fill, bitblits). It can't accelerate antialiased line drawing and gradient fills.
And it's quite ironic, because internally GDI+ has a lot of hooks for hardware acceleration. Only no one has ever implemented it.
"The idea of fusion and benefits of fusion are tremendous compared to fossil fuels but I've always wondered how long will it last before it starts eating a significant enough portion of the hydrogen to be a concern."
If your fusion powerplants are eating a significant portion of Earth's hydrogen, then it's time to relocate somewhere where the temperature is not high enough to boil oceans.
GDI+ does not use hardware acceleration, and AGG is actually faster ;)
The author of Antigrain is a perfectionist.
Just look here - http://www.antigrain.com/svg/index.html . And version 2.4 is under BSD license.
"Czar" title has a distinct advantage. Czars were abolished when the last of them had been put against a wall and the shot.
We have zinc-air elements for decades now, but they are not rechargeable.
Imagine that, say, China attacks USA using robots.
Still thinking that excessive casualties are OK?
But US is a democracy, right? Everyone has a right to keep their own political view, religion, and so on.
Then why communists were singled out and persecuted?
"So, yes, most German soldiers didn't know jack squat about the extermination, and never rounded up anybody."
Yup. That's why the Wehrmacht happily committed heinous crimes in occupied territories (Russia, Poland, Ukraine). Mostly because Wehrmacht soldiers were not disciplined for crimes against local population.
Moving goalposts already?
In any case, you're welcome: http://www.ncdc.noaa.gov/paleo/data.html
Also: http://news.bbc.co.uk/2/hi/science/nature/3090279.stm and http://environment.about.com/od/globalwarming/a/icecore.htm
"No, it isn't. It's within their competence to say that to reduce temperature by X degrees requires Y."
Yes. And the methods for these are mostly within the bounds of climatology. I.e. questions like 'by how much we must decrease CO2 to stop the warming' or 'would stratospheric aerosols help to reduce greenhous effect'.