Haven't you ever recompiled gcc with gcc? Granted, C isn't interpreted. However I've been forced to run Scheme interpreters that were written in Scheme. Scheme types adore the notion of recursion, thus they found this beautiful. Drop the word "performance" and they go through the roof, though...
The reason QNX and x86 Linux have a 4GB memory limit is that i386 is a 32-bit architecture. It is natural to use word-sized pointers, and 32-bit fields can access precisely 2^32 bits of memory which equals 4 GB. This limit is rather fundamental to the hardware architecture. Perhaps you could write an system which used 64-bit pointers on a 32-bit chip, but the real fix is to migrate to a 64 bit architecture.
I'm sorry you find this limitation "annoying", but if you really feel your applications need more than 4 Gbyte, than you should probably be using a 64-bit architecture anyway.
Certainly some day 4GB will look small, just as 640K looks small now, but the fundamental problem lies in our hardware, not our software.
I know of no "day job" for him. And some people would say that "selling out" came with Red Hat and Caldera. The real purists use Debian or maybe Slackware.
I use Red Hat, so I suppose I've sold out in some people's eyes. --Lenny
>Are they still owned by IBM? What's their market share now?
They never were owned by IBM. The chips say IBM/Cyrix because they are fabricated by IBM, since Cyrix has no fab of it's own. They've been leasing IBM fabs for some time, but the companies are distinct and, to my knowledge, IBM has had nothing to do with the design of the processors...only the production (not that fabrication is a small feet).
>Is it true that the Cyrix worked by parsing the x86 instructions on the front-end, then sending them through a RISC section?
Probably so. A number of x86 chips work internally on a RISC core, with a rather complicated translator on each side. It's rather ugly to implement, but building a modern, heavily pipe-lined processor around the x86 spec is so nasty, that many engineers have decided to implement a translation internally. I believe that my old AMD K5 was designed this way.
Also, I believe most x86 chips these days use microcode rather than straight logic to implement the more complicated functions. This can be thought of as breaking up a single, complex x86 instruction into multiple, simpler steps internally so that the ALU doesn't need to be quite so complex (and hence can run at higher clock). I've seen schematics of the K7 and it's ALU's "direct path" (strait logic) vs. it's "vector path" (microcode).
As for Cyrix...in my eyes, the only thing that keeps them relevent is the MediaGX line. I haven't really run into any fans of the MII, and most of my friends have bad tastes in their mouths from the old 686's.
Anyone know if Enlightenment is likewise approaching stability? I assume that RH wants both Gnome and Enlightenment (and let's not forget the 2.2 kernel) for 6.0 .
That was pretty funny, and really made me nostalgic. You know, modern games like Grim Fandango are awful cool, but they just aren't the same as the old text-games. Games like Zork had some real character. Also: I kind of like the idea of an adventurer trapped in an office, murking around with printers and evil coworkers, reading notes from bosses. I understand that people still write these games, maybe I should try my hand some time...
This almost makes me want to reboot into DOS...guess it's time to install dosemu on this puppy.
As someone above mentioned, these machines were designed specifically for use as workstations and came with nice, big, expensive monitors. But you are correct: only minimal video hardware is necessary on a server. My servers have TOTAL crap in the way of video hardware, and no dedicated head.
And please be careful who you call a Moron. Here you are an AC, but in life the mistake could cost you.
By the way...I like Suns, but this video thing still bothers me.
A journalling filesystem would be nice, although that will likely be saved until Linux 3.0 . ext2fs seems pretty decent, though I know little about the topic. Hmmm...wonder where I could find some good literature on filesystems (theory and comparisons)?
Anyone know when they are starting the 2.3 tree? I'm quite content with my shiny new 2.2 kernel (my/dev/sound finally works!), but I'd also like to see what they've got in the works for the future. Have they begun a tentative feature list for 2.3?
As I stated, Apple could have done just fine as a software company, letting others produce hardware, but they wanted to control the WHOLE market. Think about how powerful/wealthy MS is right now because they make the unifying OS for disparate hardware clone makers (yes: I run Linux too, but that's not the point).
IBM wouldn't have lost out so much if they had written their own OS. As it stood, when the clones came along, they bought their chips from Intel, their OS from Microsoft and nothing from IBM. The Mac clones still bought Apple software, and, in general, built better hardware than Apple. Letting them live would have made Mac-like hardware much more attractive to consumers (look what's happened in the PC sector: more competition = more variety, lower cost). This may have decreased sales of Apple hardware, but could have led to a much LARGER market for Apple software. Again I state: look at MS...I think that we'll all agree that they've been financially successful.
> The difference between PCs/Macs/OS and so on is getting less well defined as we go on.
...that may be, but I don't think that Apple is keen on the progression. Even the workstation market is starting to use PCI, AGP and other "PC" architectures...have you ever seen the inside of one of the new, low end Sun's? This is a very real market pressure...PCI equipment is common and cheap, giving Intel a large advantage (as if they didn't have one already). The only way to combat that in the price sector is to use the same gear. PCI is alright, but it's not as good as what most workstations used to use. It is very cost conscious, however.
I agree with the above poster: Apple really doesn't like being open. They like controlling ALL aspects of their little universe. Many of my Mac-addict friends agree that Apple would have done much better by releasing MacOS on x86 back in the Windows 1.0/2.0 days. The software was vastly superior to the early Windows stuff, and they could have taken advantage of the more common hardware to make a killing. But they wanted to be in the software AND hardware businesses. And don't get me started about how they killed off all the Mac clones...
As visionary as they are in many ways, I think that Apple is very short-sighted and stubborn in the strict business sense.
I was browsing the link someone posted to the "Center for Nanoscale Science and Technology" at Rice University, and they have a link where you can buy nanotubules created at Rice for research purposes. Of course, it would probably take an electron microscope to SEE the things, but somehow this strikes me as cool. Tubes for sale:
I was in a Microcenter this weekend, and I wandered through the Apple room. There were gobs of elementary school kids climbing ALL OVER the iMacs. They would crawl on top of an iMac box to play with one of the iMacs set up on display. I swear, the kids thought of them as playground equipment. And they were clicking around, exploring the systems' desktops too. Maybe brightly colored "cute" computers are the key to getting kids into technology.
wouldn't have believed it if I hadn't seen it. --Lenny
Yes, you are correct. Very good. I poked around/opt/kde/bin and found some game that was tetris-like (though no fun at all) and I just assumed that that was their replacement for the old Tetris.
I suppose I should load KWM and see what all is in the new menu...I've just been running K apps from Window Maker.
foundit on a random FTP mirror and didn't even know what it was...couldn't find mention of a version 5.3 anywhere. So OF COURSE I installed it... I'm rather pleased with it, except that KDE 1.1 no longer has Tetris in it, which is about the only game that I ever played. It seems considerably faster than when I ran Mandrake 5.1 . And KDE is nice to have, but Window Maker is far better than KWM...
What are you basing this on? Modern architectures have to have very deep pipelines (I think that the K7 will have 7) in order to maintain competitive speed. If it weren't for speculative execution, the processor would have to stall its pipeline every time a jump instruction was encountered. This is a rather serious performance hit, which would occur quite frequently in practice. I've never designed a processor myself, but according to all that I've read, neglecting speculative execution in this day and age is a rather Bad Idea.
Also: for what it's worth, the IA-64 is supposed to be a VLIW processor (or, as Intel calls it, EPIC), not a RISC design. These are more philosophies than set guidelines, however, so the matter is somewhat open to interpretation.
--Lenny
Seriously: What brand HD do you trust?
on
100gig HDs Coming
·
· Score: 1
Could definately make a poll question. I'd probably vote IBM, though my experience is limited.
It's nice that they have their page in multiple languages, but their English wasn't too much of an improvement over Babelfish.
From the looks of things, they aren't very concerned with things like the "Linux community". I see no words regarding licensing, or history. I don't expect that this will go very far, but I wish them luck.
...this is something I've thought about off and on. Plenty of people hate the internet because of the amount of porn out there, and many people are very VERY offended when they (or their child) happen upon an "adult" page by accident. It seems to me that it would be a lot easier for people who don't like porn to avoid it/filter it off of their subnets if all "adult" sites had their own top level domain category like '.adu' or something. That way, if you saw that extension, you knew what you were getting into. Also, it would be trivial to set up filters to throw up a warning when you click some mystery link that contains such a URL.
Problems: * then one has to judge what is "smut" and what isn't, which is a terribly thorny issue * enforcement would be nigh impossible * people could link with strait IP addresses, making DNS a moot point.
I don't think idea is very realistic, but I think it would be nice. I expect that it would help different factions coexist peacefully on the internet.
Slashdot Mirror
Haven't you ever recompiled gcc with gcc? Granted, C isn't interpreted. However I've been forced to run Scheme interpreters that were written in Scheme. Scheme types adore the notion of recursion, thus they found this beautiful. Drop the word "performance" and they go through the roof, though...
BTW- I like PERL
--Lenny
The reason QNX and x86 Linux have a 4GB memory limit is that i386 is a 32-bit architecture. It is natural to use word-sized pointers, and 32-bit fields can access precisely 2^32 bits of memory which equals 4 GB. This limit is rather fundamental to the hardware architecture. Perhaps you could write an system which used 64-bit pointers on a 32-bit chip, but the real fix is to migrate to a 64 bit architecture.
I'm sorry you find this limitation "annoying", but if you really feel your applications need more than 4 Gbyte, than you should probably be using a 64-bit architecture anyway.
Certainly some day 4GB will look small, just as 640K looks small now, but the fundamental problem lies in our hardware, not our software.
--Lenny
I know of no "day job" for him. And some people would say that "selling out" came with Red Hat and Caldera. The real purists use Debian or maybe Slackware.
I use Red Hat, so I suppose I've sold out in some people's eyes.
--Lenny
>Are they still owned by IBM? What's their market share now?
They never were owned by IBM. The chips say IBM/Cyrix because they are fabricated by IBM, since Cyrix has no fab of it's own. They've been leasing IBM fabs for some time, but the companies are distinct and, to my knowledge, IBM has had nothing to do with the design of the processors...only the production (not that fabrication is a small feet).
>Is it true that the Cyrix worked by parsing the x86 instructions on the front-end, then sending them through a RISC section?
Probably so. A number of x86 chips work internally on a RISC core, with a rather complicated translator on each side. It's rather ugly to implement, but building a modern, heavily pipe-lined processor around the x86 spec is so nasty, that many engineers have decided to implement a translation internally. I believe that my old AMD K5 was designed this way.
Also, I believe most x86 chips these days use microcode rather than straight logic to implement the more complicated functions. This can be thought of as breaking up a single, complex x86 instruction into multiple, simpler steps internally so that the ALU doesn't need to be quite so complex (and hence can run at higher clock). I've seen schematics of the K7 and it's ALU's "direct path" (strait logic) vs. it's "vector path" (microcode).
As for Cyrix...in my eyes, the only thing that keeps them relevent is the MediaGX line. I haven't really run into any fans of the MII, and most of my friends have bad tastes in their mouths from the old 686's.
--Lenny
Anyone know if Enlightenment is likewise approaching stability? I assume that RH wants both Gnome and Enlightenment (and let's not forget the 2.2 kernel) for 6.0 .
--Lenny
That was pretty funny, and really made me nostalgic. You know, modern games like Grim Fandango are awful cool, but they just aren't the same as the old text-games. Games like Zork had some real character. Also: I kind of like the idea of an adventurer trapped in an office, murking around with printers and evil coworkers, reading notes from bosses. I understand that people still write these games, maybe I should try my hand some time...
This almost makes me want to reboot into DOS...guess it's time to install dosemu on this puppy.
--Lenny
As someone above mentioned, these machines were designed specifically for use as workstations and came with nice, big, expensive monitors. But you are correct: only minimal video hardware is necessary on a server. My servers have TOTAL crap in the way of video hardware, and no dedicated head.
And please be careful who you call a Moron. Here you are an AC, but in life the mistake could cost you.
By the way...I like Suns, but this video thing still bothers me.
--Lenny
A journalling filesystem would be nice, although that will likely be saved until Linux 3.0 . ext2fs seems pretty decent, though I know little about the topic. Hmmm...wonder where I could find some good literature on filesystems (theory and comparisons)?
--Lenny
Anyone know when they are starting the 2.3 tree? I'm quite content with my shiny new 2.2 kernel (my /dev/sound finally works!), but I'd also like to see what they've got in the works for the future. Have they begun a tentative feature list for 2.3?
--Lenny
As I stated, Apple could have done just fine as a software company, letting others produce hardware, but they wanted to control the WHOLE market. Think about how powerful/wealthy MS is right now because they make the unifying OS for disparate hardware clone makers (yes: I run Linux too, but that's not the point).
IBM wouldn't have lost out so much if they had written their own OS. As it stood, when the clones came along, they bought their chips from Intel, their OS from Microsoft and nothing from IBM. The Mac clones still bought Apple software, and, in general, built better hardware than Apple. Letting them live would have made Mac-like hardware much more attractive to consumers (look what's happened in the PC sector: more competition = more variety, lower cost). This may have decreased sales of Apple hardware, but could have led to a much LARGER market for Apple software. Again I state: look at MS...I think that we'll all agree that they've been financially successful.
--Lenny
> The difference between PCs/Macs/OS and so on is getting less well defined as we go on.
...that may be, but I don't think that Apple is keen on the progression. Even the workstation market is starting to use PCI, AGP and other "PC" architectures...have you ever seen the inside of one of the new, low end Sun's? This is a very real market pressure...PCI equipment is common and cheap, giving Intel a large advantage (as if they didn't have one already). The only way to combat that in the price sector is to use the same gear. PCI is alright, but it's not as good as what most workstations used to use. It is very cost conscious, however.
I agree with the above poster: Apple really doesn't like being open. They like controlling ALL aspects of their little universe. Many of my Mac-addict friends agree that Apple would have done much better by releasing MacOS on x86 back in the Windows 1.0/2.0 days. The software was vastly superior to the early Windows stuff, and they could have taken advantage of the more common hardware to make a killing. But they wanted to be in the software AND hardware businesses. And don't get me started about how they killed off all the Mac clones...
As visionary as they are in many ways, I think that Apple is very short-sighted and stubborn in the strict business sense.
--Lenny
I was browsing the link someone posted to the "Center for Nanoscale Science and Technology" at Rice University, and they have a link where you can buy nanotubules created at Rice for research purposes. Of course, it would probably take an electron microscope to SEE the things, but somehow this strikes me as cool. Tubes for sale:
http://cnst.rice.edu/tubes/
--Lenny
I was in a Microcenter this weekend, and I wandered through the Apple room. There were gobs of elementary school kids climbing ALL OVER the iMacs. They would crawl on top of an iMac box to play with one of the iMacs set up on display. I swear, the kids thought of them as playground equipment. And they were clicking around, exploring the systems' desktops too. Maybe brightly colored "cute" computers are the key to getting kids into technology.
wouldn't have believed it if I hadn't seen it.
--Lenny
Yes, you are correct. Very good. I poked around /opt/kde/bin and found some game that was tetris-like (though no fun at all) and I just assumed that that was their replacement for the old Tetris.
I suppose I should load KWM and see what all is in the new menu...I've just been running K apps from Window Maker.
--Lenny
foundit on a random FTP mirror and didn't even know what it was...couldn't find mention of a version 5.3 anywhere. So OF COURSE I installed it... I'm rather pleased with it, except that KDE 1.1 no longer has Tetris in it, which is about the only game that I ever played. It seems considerably faster than when I ran Mandrake 5.1 . And KDE is nice to have, but Window Maker is far better than KWM...
--Lenny
What are you basing this on? Modern architectures have to have very deep pipelines (I think that the K7 will have 7) in order to maintain competitive speed. If it weren't for speculative execution, the processor would have to stall its pipeline every time a jump instruction was encountered. This is a rather serious performance hit, which would occur quite frequently in practice. I've never designed a processor myself, but according to all that I've read, neglecting speculative execution in this day and age is a rather Bad Idea.
Also: for what it's worth, the IA-64 is supposed to be a VLIW processor (or, as Intel calls it, EPIC), not a RISC design. These are more philosophies than set guidelines, however, so the matter is somewhat open to interpretation.
--Lenny
Could definately make a poll question. I'd probably vote IBM, though my experience is limited.
--Lenny
It's nice that they have their page in multiple languages, but their English wasn't too much of an improvement over Babelfish.
From the looks of things, they aren't very concerned with things like the "Linux community". I see no words regarding licensing, or history. I don't expect that this will go very far, but I wish them luck.
--Lenny
...this is something I've thought about off and on. Plenty of people hate the internet because of the amount of porn out there, and many people are very VERY offended when they (or their child) happen upon an "adult" page by accident. It seems to me that it would be a lot easier for people who don't like porn to avoid it/filter it off of their subnets if all "adult" sites had their own top level domain category like '.adu' or something. That way, if you saw that extension, you knew what you were getting into. Also, it would be trivial to set up filters to throw up a warning when you click some mystery link that contains such a URL.
Problems:
* then one has to judge what is "smut" and what isn't, which is a terribly thorny issue
* enforcement would be nigh impossible
* people could link with strait IP addresses, making DNS a moot point.
I don't think idea is very realistic, but I think it would be nice. I expect that it would help different factions coexist peacefully on the internet.
I can dream, can't I?
--Lenny