...but pretty soon Solais 10 is going to be a big competitor to Linux on laptops, especially the 64-bit AMD ones.
It already works pretty well on the Acer Ferrari 3000 series. Most stuff "just works" (wifi, USB, firewire, card reader, dvd writer etc.) and JDS is a fairly tolerable desktop if you can put up with Sun's pointy-haired decision to replace a lot of the native GNOME applets with (inferior) ones written in Java.
I think they are working on refining power management now.
WINE can use native Windows DLLs, in case you need one that isn't yet implemented fully by WINE itself, or if there is some particular quirk of the native DLL that you need to have.
Back in the '80's MS-DOS viruses used to intercept system calls, take over interrupt vectors, patch DOS itself etc. to hide themselves. There were features of viruses called "tunnelling" and "stealth".
Emulators are not a bad choice. I'd much rather install an emulator for the target platform in a light-weight laptop than lug around a huge beige box all day...
An interesting piece of information for all the raving zealots on both sides of the argument to note: the first AMD64 (Opteron/Athlon 64) port of Linux was done on a software simulator, before any hardware was available.
Software isn't a bad choice, especially nowadays with good free emulators such as bochs. The kind of things you'll be doing for a class assignment will not require cutting-edge performance, merely the ability to execute the code correctly. All the extra debugging facilities that come with emulators can be very useful. Especially if you're very used to one instruction set and then have to code for a completely different one, it's all very well reading the manuals, but when you actually try it, you'll be surprised. The difference between x86 and RISC (e.g. PowerPC, MIPS, ARM, SPARC) is enormous.
Although there are still issues about what makes a machine when there is a very tightly coupled network, this actually makes the most sense. After all, the major flaw in the per CPU (chip) but not per core argument is that it allows some companies (Intel, for example) to put multiple processors into a machine that only needs one license, but prevents another company (Asus, for example) to build a motherboard (machine) that takes multiple processors by acomplish the exact same end. By what logic should an Intel motherboard running one Intel chip but containing four complete core processors pay a lower licensing fee that an Asus motherboiard with (for example) two AMD cores, each one on it's own chip, for a total of only two cores?
Point of information: AMD announced dual core Opterons months before intel announced dual core Pentiums.
The engineering samples have been out for a long, long time.
The 387 wasn't a seperate processor. It was more like an add-on extention to the microcode ROM of the 386 to implement the floating-point instructions, in addition to a few extra pieces of low-level circuitry, and it also contained the floating-point specific registers.
Please note that "Hyperthreading" processors are not dual-core. They just have hackery to make a single pipeline look like two.
You may jest, but I once worked for an old assembly-laguage coder. He used to put random dbs into his source to confuse casual would-be reverse engineers. I went home and wrote a disassembler that followed program flow...
How about a pentagram and a picture of a goat sitting on a wooden throne, blood raining down and all kinds of beasts and monsters engaged in acts of wanton carnal depravity while bodies of sinners roast half-dead over bonfires?
Just look at the scaling graphs here for parallel bzip2. Note the almost linear scaling on proper SMP and NUMA architectures vs. the embarrassing curve on intel's old-fashioned bus architecture with dual pentium 4 xeons with hyperthreading. Also notice the high clock frequency on the intel processors compared with the performace achieved on the "slower" ones...
IBM has not ported some of it's application software to Solaris 10, which could be a problem for Sun.
Solaris garantees backwards binaray compatability. You can run all your old Solaris x86 32-bit binaries unchanged and with better performance on a nice shiny new Opteron or Athlon 64 with 64-bit Solaris 10. This move by IBM is not a threat to Sun. It just shows that IBM is running scared from Solaris 10.
Solaris 10 is set to take a lot of customers away from IBM. IBM is very afraid.
The problem with Linux is that over the years things have changed and broken binray compatability. This isn't a show-stopper usually, but if you do have some closed-source software from 5 years ago that you still want to run today, you are going to find all kinds of library dependency problems.
The thing about Linux is that most of your applications are Open Source or Free, so they get updated and recompiled incrementally as time goes on.
I bought some Loki games for Linux a long time ago. Some of them haven't worked in years because they depend on obsolete and deprocated libraries. If I had lots of time on my hands (which I don't have nowadays) I could probably spend several days looking out old source tarballs and doing a bit of porting, but life's too short.
Most people or businesses who buy software or computers to do a job need specific version of specific kernels with specific libraries and utilities and specific versions of applications that have been integrated, tested and certified to work together.
Windows is very poor at this. Linux is a bit better, but if you're using Linux commercially, you're probably using RedHat Enterprise Linux (or maybe SuSE), you've payed hundreds or thousands of dollars for the software license (for the OS), you've probalby spent tens of thousands on the hardware, you have a support contract, you'll have spent thousands on the applications and you'll have trained clued-up staff to deal with it all.
Does Red Hat garantee backwards compatability?
Can I get Red Hat ES today and Oracle and be garanteed that in 5 years time, my Oracle that I bought will still run, unchanged (same binary), still supported etc.?
Linux is much, much better than Windows, but no Linux company has solved this problem yet.
Well, see, he made Adam out of mud, and Eve out of one of Adam's ribs. Then when he wanted to make a son for himself the next time, instead of getting out his modelling clay again (mud) he got poor old Mary up the duff. Can you imagine the stigma attached to that at the time? Poor Mary. Cruel, wicked, vindictive god.
Slashdot Mirror
It already works pretty well on the Acer Ferrari 3000 series. Most stuff "just works" (wifi, USB, firewire, card reader, dvd writer etc.) and JDS is a fairly tolerable desktop if you can put up with Sun's pointy-haired decision to replace a lot of the native GNOME applets with (inferior) ones written in Java.
I think they are working on refining power management now.
WINE can use native Windows DLLs, in case you need one that isn't yet implemented fully by WINE itself, or if there is some particular quirk of the native DLL that you need to have.
Nothing much changes it would seem.
+1 Insightful.
An interesting piece of information for all the raving zealots on both sides of the argument to note: the first AMD64 (Opteron/Athlon 64) port of Linux was done on a software simulator, before any hardware was available.
Software isn't a bad choice, especially nowadays with good free emulators such as bochs. The kind of things you'll be doing for a class assignment will not require cutting-edge performance, merely the ability to execute the code correctly. All the extra debugging facilities that come with emulators can be very useful. Especially if you're very used to one instruction set and then have to code for a completely different one, it's all very well reading the manuals, but when you actually try it, you'll be surprised. The difference between x86 and RISC (e.g. PowerPC, MIPS, ARM, SPARC) is enormous.
Make that three. Brian Potter fans too?
Point of information: AMD announced dual core Opterons months before intel announced dual core Pentiums.
The engineering samples have been out for a long, long time.
The 387 wasn't a seperate processor. It was more like an add-on extention to the microcode ROM of the 386 to implement the floating-point instructions, in addition to a few extra pieces of low-level circuitry, and it also contained the floating-point specific registers.
Please note that "Hyperthreading" processors are not dual-core. They just have hackery to make a single pipeline look like two.
You may jest, but I once worked for an old assembly-laguage coder. He used to put random dbs into his source to confuse casual would-be reverse engineers. I went home and wrote a disassembler that followed program flow...
Then it will start to die.
All of the great things in the world were built out of love, not money.
You could glue one of those little calendar things to the bottom of it and give it to mummy and daddy for Christmas!
Time for my medication...
Just look at the scaling graphs here for parallel bzip2. Note the almost linear scaling on proper SMP and NUMA architectures vs. the embarrassing curve on intel's old-fashioned bus architecture with dual pentium 4 xeons with hyperthreading. Also notice the high clock frequency on the intel processors compared with the performace achieved on the "slower" ones...
IBM is nothing more than the technological arm of the Salvation Army
Well, maybe their friend Bill will buy them out and M$ will release a version of NT with a Solaris kernel to run on EM64T.
Solaris garantees backwards binaray compatability. You can run all your old Solaris x86 32-bit binaries unchanged and with better performance on a nice shiny new Opteron or Athlon 64 with 64-bit Solaris 10. This move by IBM is not a threat to Sun. It just shows that IBM is running scared from Solaris 10.
Solaris 10 is set to take a lot of customers away from IBM. IBM is very afraid.
Solaris 10 comes with MySQL.
MySQL is already in Solaris 10.
Elvis, is that you? The king of Rock and Roll? Are you risen and living amongst us?
Just like Solaris, then :-)
If only 'twere true.
The problem with Linux is that over the years things have changed and broken binray compatability. This isn't a show-stopper usually, but if you do have some closed-source software from 5 years ago that you still want to run today, you are going to find all kinds of library dependency problems.
The thing about Linux is that most of your applications are Open Source or Free, so they get updated and recompiled incrementally as time goes on.
I bought some Loki games for Linux a long time ago. Some of them haven't worked in years because they depend on obsolete and deprocated libraries. If I had lots of time on my hands (which I don't have nowadays) I could probably spend several days looking out old source tarballs and doing a bit of porting, but life's too short.
Most people or businesses who buy software or computers to do a job need specific version of specific kernels with specific libraries and utilities and specific versions of applications that have been integrated, tested and certified to work together.
Windows is very poor at this. Linux is a bit better, but if you're using Linux commercially, you're probably using RedHat Enterprise Linux (or maybe SuSE), you've payed hundreds or thousands of dollars for the software license (for the OS), you've probalby spent tens of thousands on the hardware, you have a support contract, you'll have spent thousands on the applications and you'll have trained clued-up staff to deal with it all.
Does Red Hat garantee backwards compatability?
Can I get Red Hat ES today and Oracle and be garanteed that in 5 years time, my Oracle that I bought will still run, unchanged (same binary), still supported etc.?
Linux is much, much better than Windows, but no Linux company has solved this problem yet.
/me ducks
Shhhhhh....You're not supposed to diparage IBM here, silly! Slashdot runs on DB2, and IBM rulez whereas Sun SuX0rz, don't you know?
Well, see, he made Adam out of mud, and Eve out of one of Adam's ribs. Then when he wanted to make a son for himself the next time, instead of getting out his modelling clay again (mud) he got poor old Mary up the duff. Can you imagine the stigma attached to that at the time? Poor Mary. Cruel, wicked, vindictive god.
What point are you trying to make?
Do you feel threatened by my beliefs and opinions?
Why do you feel that you have to silence me?
Why is my contribution to this discussion not worthy of consideration?