by a university's choice to go MS -- it doesn't seem like they provide any solutions "big enough iron" so to speak.
Here at CMU (which is not that big a university, ~6000 undergrads) we run a massive distributed filesytem, afs, which is commercially available and was developed in part here (if we reimplemented the "andrew" system today it would probably be around CODA, which is Free iirc). The distributed fileservers have always been Solaris (and also a bit of HP-UX), and I'm not aware you could replace them with NT if you wanted to. AFS is supported by a a wide variety of clients: NT, Linux (and the other Unices), Mac.
The individual departments are not likely to give up their own special types of computers -- Design and Art want their SGIs, many of the professors (and students) use Macs, and the geeks all use Linux and Solaris. What solution based on NT can serve all those clients, for a system with tens of thousands of total users?
I just don't see any other way to run a computing environment the size of a university other than Kerberos and heavy-duty distributed fs stuff. Perhaps I'm missing something?
I certainly don't believe MS could provide anything like CMU's reliability. I've been here over a year now. Once or twice the routing has broken for a few minutes, and once the university blew a power feed and everything on the other side of the street shut down for a day (actually a lot of it was running on backup... but the all-solid-state stuff probably wasn't)
So yeah, there has to be something made clear here: part of the reason Debian works so beautifully is that the distribution itself consists of ~3000 packages maintained by various people, all held to the same standards.
The result is that for 99.5% (at least that's the number on my system) of stuff you want, you use the distro's official version/build, which is guaranteed to work perfectly. On my system, I have ~530 packages installed. There are only 3 that are not native Debian packages -- the only 3 things I've found that I wanted that weren't available. So I downloaded the rpm and installed 'em with alien:)
It's kind of a similar philosophy to BSD, in that the OS consists not just of a kernel and a set of basic stuff, but of everything. In a BSD this is true because everyone's running THE official BSD -- no distros. In Debian, there's the same kind of mentality. By comparison, RedHat's distro is sparse (granted most of the difference is in packages that very very very few people are interested in).
Whenever I find out about a new program I find I need (happens all the time, this is Linux after all:), the first thing I do is a
dpkg -l "*progname*"
to see if there's already a Debian package. Exactly 3 times, there hasn't been one. If Debian wasn't behind on processing developer applications (they're focusing efforts on catching up with bug reports) I'd have already built debs for these apps.
Sometimes the system doesn't work so well: as an example, there hasn't been a working version of XTraceroute (maybe you don't consider that essential... I do:) in like 4 months, since there hasn't been a working version of gtkglarea. Unlike RedHat, there's no rufus repository with tons of builds of different versions of everything. On the other hand, in my experience (ran RedHat for a long time) a lot of stuff on Rufus didn't work because it was impossible to meet its dependency requirements, since there was no central control over the type of machine it was built on. In Debian, the latest version of each package in the Stable or Unstable (woohoo unstable!) trees is always built to work with the latest of everything else. It's only because of this that the auto-installed dependency thing works at all.
The result is impressive.
I've run a lot of both distros. I came to the conclusion Debian was superior because it has everything in a compatible, working version. That makes my life a lot simpler. In addition, Debian packages have Config scripts that make setting up a lot of stuff stuff easy even if you're not familiar with the specific config file formats... let me set up a masquerading ppp line machine in about 3 minutes (of course less 45 minutes toying with isapnp to get the bloody pnp modem to work:). I just answered the config questions, gave it the network info, phone number, account, and password, and BANG! it worked. Hehe, way easier than doing the same thing years ago under Slack 3.2:)
Man, that was way offtopic. I do intend to try Corel's stuff tho. I hope the license on their config stuff is good, cause Debian's idea of a "user-friendly" install is, the "voice" of the person who asks you various questions and tells you what's going on (in text mode) during the install, is pretty chatty and likable:) It's pretty ugly, and not to be approached if you mind rtfm'ing...
I've used 3Coms in Linux boxes for years -- from terrible old 3c501s to 3c509s and now a 3c905. Don Becker's drivers have always worked perfectly for me.
I obviously shoulda checked my references rather than do this off the top of my head....
I'm only saying: I don't see what advantage microII has in embedded applications -- I don't see a particular incentive for embedded chip designers to pick it up.
Does anyone know what its power needs etc. are? Have there been advanced, low-power versions of this chip produced? The microII 85MHz in my SB3 is pretty power hungry:)
Now an embedded Ultra... that would be interesting. I wouldn't mind an Ultra based subnotebook:)
The MicroSPARCII is not a particularly exciting processor. It's the chip that was used in the SS5, and it ran up to 170MHz in the last SS5 offered. It's not designed for SMP (Turbo- and HyperSPARCs did that in SS10 and SS20). It's a 32bit chip, a bit faster at floating point than an equivalent Pentium. (it's also the chip in my SPARCBook 3:).
So don't expect cheap high powered crazy Suns floating around soon. Sun wants people using MSII's where they're now using R4xxx's and ARMcores and m68k's in PDAs.
An advantage is of course Linux already runs fine on it:) But it does on the ARM, Motorola and MIPS chips too.
SGI has to redefine who it is. Of course the logo thing was a mistake -- long live the old SGI logo!
They have a serious problem: you don't need a crazy SGI station to do graphics anymore. You'll do decently well with a high-end PC, 3DSMax, and good consumer graphics hardware, and for a fraction of the price. There was a reason so many people fell in love with equipment like the original Indigo (mmmm....purple box): it got the job done in a much more primitive age. It was the only way.
I recently heard a story about a little graphics design company a couple of years back (not all that many years, either, maybe 2 or 3). They didn't have money for SGIs, so they invested in PCs and MAX back when MAX was the first product for PC that didn't suck. They would get clients interested with samples, then give a tour of their office. When the clients saw they used PCs, they ran away. So then they got the clients interested, but made excuses why they couldn't visit the office ("We're painting...","We just moved...","There was a water leak, it's a mess..."). After that they started landing jobs and making money:)
But nowadays you can equip a 3D artist with a $2k PC and a $2500 copy of MAX and be in good shape, rather than SGI equipment and software totaling 4x as much.
So that's their problem. I don't know what the solution is. Linux is great, but it runs on cheap PCs too:)
I guess if I knew what the solution was I'd go be the CEO of SGI...
a) Unlikely. It's not necessarily that well suited -- xfs is useful on servers, where fscking for a long time (say over several hundred gigs of RAID diskspace) can be a real turnoff. But xfs is a big-iron solution -- it has overhead that will make it inappropriate for workstations with limited diskspace. In addition, there are efforts (which may benefit from xfs) such as Tweedie's ext3, which will journal and may also be more reasonable for lighter weight applications (you'll have to ask him about that tho... ).
Put it this way: if you haven't lost a day of productivity for a whole team of people due to a server fscking its drives after an unplanned/unclean shutdown, you don't care about xfs. On the other hand, if that's you... you might really want it in there. I think there's a good chance we'll see xfs pretty much straight dumped into the kernel in the near future (perhaps in the 2.5 series). Later we'll see (hopefully) more elegant solutions based on previous work, including parts of xfs.
Anyone care to comment on more precise schedules for these projects?
b) Probably none. SGI's commitment to Linux is too strong. Better yet, as a strong point of Free software, all they have to do is release the source and let the Linux community do the work.
An interesting property of the OS: If MS goes out of biz or decides to drop some OS, you're fsck'd. But while a key programmer on a Linux component losing a job and being forced to go elsewhere to work on unrelated stuff to make ends meet can hurt a Free project, it can't kill it, the way Amiga abandoning its userbase will (IMHO) kill the Amiga community, which can no longer expect any progress on a completely closed system.
Offtopic-bait hint to Amiga: Free the source, if you want it to live...
as well as of massive overdiagnosis of things like ADHD/ADD.
I suspect this psychologist is reading things a bit far into these results. While is possible there are a few highly intelligent technical people who are mildly autistic, I think it is very rare.
I think it's much more likely that parents who were themselves clumsy and socially inept bring up distant, socially uncomfortable children. This may be part of the "nurture" makeup of autism that these researchers are considering, but I have trouble seeing it as a disease.
Speaking from pretty decent samples -- TJHSST and CMU SCS -- I think it's far more likely for nerds to a) Do ballroom. Nerds dance as well as anyone and we're always looking to meet girls. b) Do S'n'S or have other interest/background in acting and theater, including film. c) Be perfectly socially graceful, except for lapses of lust over hardware, and a deep enjoyment of conversations no one else can understand.
I strongly believe that physicists have a reputation (well deserved IMHO) for being poorly dressed (atrocious prevalence of khakies/sneakers... ) because it's comfortable and they honestly have more important things to do than care. When they do choose to dress well they have no problems doing so...
I've played with Dreamcast -- though I only played 3D Sonic. I loved it. I don't know if I would buy one (haven't seen the other games) but I wasted a lot of time at E3 on it:). The graphics were way ahead of the other consoles, but that's not saying much. I don't think the graphics were up to par with a high end nVidia or 3dfx.
That said, I don't expect to be impressed with PSII. Sure the numbers sound good now (even accounting for the blatant lies that they put out -- don't believe anything like those poly rates, or any other numbers from any 3d hardware company), but the PC industry moves damn quick. GeForce sounds pretty bloody slick (I want... ) and 3dfx and nVidia will be out with new cards in the spring. PSII will realistically be available for XMas'00 (i.e. will definitely be ready). If it's ready this spring, it'll be pretty cutting edge for about 6 months. If it's ready in the fall (I think this is more likely) it'll already be outdated.
Not that I do spend money:) but I'd rather pay $100-150 every year for a new 3d card for my PC that will keep me up to date, and since I already need a new PC every ~18months, than pay $450 for a console that will be cool for a few months, then old for years to come while I can't buy the competitor's console and still play the same games... but then I'm kinda PC-biased...
Cache is not something you get to control directly in code.
When you read a dword from main memory (i.e. not already in the cache) and then do operations on it, the cache controller takes advantage of a free memory bus to go ahead and read the rest of the cache line. Assuming your code is well optimized for cache performance, the next things you read should already be in the cache.
If you're doing a lot of kernel stuff, large chunks of the kernel will be in the cache, as you would want. And if you're running Quake 3, Quake 3 will be in the cache. It's exactly what you want.
I believe (get me if I'm wrong) Rob has reduced the number of moderators from when he originally introduced it. He's said before that he feels that moderation should be a rare thing and only a few comments to an article should ever be moderated in either direction.
I didn't agree initially, but now I do. I try to read everything at least down to 2, and I think this means you get all the intelligent comments, not just the ones lucky enough to make it up to 5.
Of course, there are still a few rare ones that make it up to 5, and some of them don't really deserve it. IMHO moderators should focus on looking for diamonds in the rough rather than further moderating up 3's and 4's.
Perhaps moderating up an article should get harder (requires more points) the higher it starts, so it would cost a moderator a lot of points (or alternatively take several moderators) to bring a 4 up to a 5, but only one to take a 1 to a 2.
Obviously ripping (assuming a good ripper eg. cdParanoia:) should be perfect. You should end up with the same bits on your HD as on the CD. Of course the encoder has something to do with this too (since mp3 is inherently lossy).
You've got it right when you say the real problem is the sound card -- even an SBLive! gold is pretty noisy in analog mode, and the inside of your computer might as well be a freaking radio station. But with the Gold (and I believe certain versions of Diamond's MonsterSound, tho I'm not sure) you can output in digital, which in theory could be hooked up directly to your stereo.
At least one of the digital out's on the SBLive! Gold is a proprietary (i believe, does anyone know the name of the format/standard it uses?) plug that works with the FPS2000 digital speaker set. It's good sound for a computer but not exactly audiophile stuff. I'm not sure if there are any standard (eg. Dolby digital? SPDIF? you tell me...) outs on the SBLive! Gold.
There are however definitely Dolby digital outs on many DVD players, and you can hook those up in digital to a good component system.
It's perfectly possible to do 3D in software. It's even somewhat more interesting, since you're not bound to and 3D card's paradigm. Software is where you get cool stuff like true voxels etc...
Glide probably wasn't out before DX3, but DX3 was pretty much useless (only a very minimum 3D API) so Glide may have beaten anything useful, tho not by much...
------------------------
Replying to another comment: No, an API cannot be faster. But an implementation can. And an API's design can affect an implementation. In any case, Glide (the implementation) is a hell of a lot faster than D3D (the implementation), as per my original comment.
Microsoft writes each version of D3D by asking the manufacturers "What features do you guys need?" and then writing them in.
There were no 3D games bfore D3D. No one had cards.
Just 'cos a card supports D3D doesn't mean you can assume your program will work right. You still have to test and debug each individual card. This is the voice of experience:)
Matrox's bump will be in D3D i'm pretty sure...
People use Glide rather than D3D 'cos it's way faster. Speed really is all that matters...
I read another article (the one on rivaextreme... pretty good article), I can add a few more comments.
GeForce has environment mapping (iirc so does Permidia3) but not bump mapping.
It can do 8 free hardware accel'd lights... imho this is kind of limiting... we'll see.
A Voodoo3 on a fast machine under Glide can handle about a 10-12kpoly scene lighted textured w/effects and phsyics running about 20-25 fps on a 450a. I'll be very impressed if GeForce can do twice that -- 25kpolys at 25fps, or about 500k real polys/sec (BTW a Voodoo3 under ideal conditions w/out features can do 500k "fake" polys/sec... I again expect GeForce to better that...).
But 15million polys/sec is the kind of bloated number that usually comes out of graphics shops. Don't believe it for a second.
As for 100kpoly models lighted w/fx running smoothly, i'll believe it when I see it.
if DaveS or DaveR wants to correct me on any of this stuff, go right ahead guys...
So basically nVidia chose to make a high fill-rate card with hardware lighting and transforms (geometry acceleration). These aren't innovative directions -- they were the obvious ones. None the less, the other major player, 3dfx, has pulled back from these choices. I'll explain why:
nVidia has a card which can do supported operations fast. It obviously has a lot of fill. It'll be a good board. Of course it'll still be slow in D3D... everything is (we once demonstrated that it's physically impossible under DX6 to be faster than a Voodoo3 under Glide). There are some downsides: if you want to do crazy weird stuff with your lighting (eg. wrong faster stuff, funky effects) you may not be able to get it to work. Similarly with geometry -- special fast cases will become normal cases. So there may be a 50%-100% gain in triangle rate, but it's unlikely geometry acceleration will ever be able to provide much more than that.
nVidia seems to have chosen not to support the hardware bump mapping of the Matrox G400, an extremely high fill (runs beautifully bump mapped in a window in 1600x1200x32bpp) card without geom accel. 3DLabs' long awaited Permidia3 will also have some kind of hardware bump. IMHO this is a relatively flexible feature -- you could do a lot with it. It remains to be seen how flexible nVidia's lighting and geom turn out to be.
I'll be impressed if D3D ever delivers real hardware geometry benafits. We have yet to see a single benefit of DX6 over DX5 (not screwing with the fp control word especially) actually work. I'm highly suspect of anything MS sez.
So what about the remaining behemoth, 3dfx? Their Voodoo4 is supposed to be an extremely high fill card (fill has always been their hallmark). It may not support any more hardware features (eg. bump, lighting, geom accel), but it will fill like crazy. It's supposed to do full screen anti-aliasing... 3dfx talked about putting a geometry accelerator on V4 but I believe they backed off from it. Voodoo4 is however still an SST and therefore still a true descendent of the original Voodoo chipset conceived as a flexible, long-term solution for both PCs and arcade games.
I'm eagerly awaiting the new generation. But I expect the real crazy stuff to start happening in the following generation... it may be finally time to kill some very old paradigms in 3d hardware...
I've been running Q3A reasonably happily on my TNT 2 Ultra for a couple of weeks now. The machine is a Celeron 450a.
The GLX renderer isn't amazing -- I have to turn off lightmaps, and I run in 640x480 (haven't really messed w/higher res modes, the card may have plenty of fill to do them...). I'm still waiting for DRI to get full performance out of the board.
Nonetheless, q3 ought to run acceptably on even a TNT with a few more of the settings tuned down...
Okay, there's not a Linux hacker on the face of the planet who wouldn't kill to have your job.
- Paid by RHADL. - Wake up when you want, work when you want. - Go to all the big trade shows. - Work with the likes of Linus and all the other regulars. - Get free toys (and I mean good toys) like PA-RISC systems from HP and Athlons from AMD.
But it wasn't always that way. Back before even RedHat paid you, back when you hacked on your aging spare equipment, what drew you to it? How did you know this was what you wanted to do before you knew about all the perqs involved, or that there would ever be perqs?
A language without excellent, maintainable, preferrably native C/C++ bindings is useless.
I can't walk into a contract job, inform them that I'll code their solution in Obj-C, and then do it and split. I'd be leaving them with something they have no prayer of maintaining or extending.
Java is unacceptable -- too slow, no particular reason for existence.
Obj-C is unacceptable -- if it's as good as Smalltalk, it probably belongs in the same catagory (along w/ML and others) of academic languages of no practical value.
Tell me what Obj-C does so much better than C++ that it's worth the cost of conversion.
I think the best way to succinctly describe OpenStep is to say it is to software developers what the Macintosh is to normal users.
Considering I can't for the life of me figure out how to use a Mac (it doesn't have any nipples...it's not intuitive in my book and I couldn't find the man(1) command) I don't think I'd enjoy it.
Seriously, the Mac's a pretty limiting environment. It's a padded room. Is that really where you want to be coding? I sure don't.
but I'll attempt to justify myself anyhow. I'm choosing to skip the whole display driver architecture question and focus on application programming.
First of all, I don't believe XML/whateverML/Mozilla to be the answer. I don't want my application written in a scripting language. Running everything through your browser is a bad idea, and everyone except MicroSoft thinks so. Frankly, I haven't seen a browser that doesn't suck (except Lynx!) and I'm not convinced I ever will. I'm not that impressed (yet) by the Mozilla efforts, tho I support them.
I'm only going to consider compiled widget toolkits, no Java. I'm sorry. Java sucks. It involves way too much typing, and it's slow. I don't approve of anything that makes code slow. The whole point is to be fast.
What does that leave?
1. Motif I once sat thru a college lecture on Motif. In an hour, the lecturer had managed to cover use of the clipboard. No function name was less than 20 characters, and it took 15 calls to get the most basic thing done. No wonder Netscape under Motif was ~1million lines. 'Nuff said.
2. MFC MFC isn't really so bad as a design. It can be annoying, and it often lacks flexibility, but it's not completely unusable. However, I find that using MFC (specifically in VisualC++, which conveniently writes the hard/annoying parts for you) is a good way to kill a project. It forces you to use their class heirarchy and objects, and if the generated code gets screwed up you can be in real trouble. The usability is also heavily dependent on spotty MS documentation -- if the documentation is bad (eg. the MDI multimedia devices in VC5) figuring it out can be a chore. A mediocre choice at best. At least it has nice built in database drivers.
3. Visual Basic has got to be the quickest way of throwing together a UI I've ever seen. I really like it, especially for prototyping. I have no speed problems (since it's compiled now and officially in my OK book), but BASIC can be a dreadfully limiting language. Writing the interface in VB and all the hard code in a DLL isn't a half bad idea, tho. At the very least useful for trying out interface designs on people. Some built in database drivers.
4. PowerSoft Power++, Borland CBuilder I've only ever used Power++. It's really just a wrapper around MFC classes, but it cleans a lot of stuff up and does all the hard stuff. It still pushes you to use MFC-type objects, but they're quite a bit less ugly to work with, and the docs are better. But if the rest of you're code's not in Watcom, Power++ might cause compiler issues (which is why I eventually dropped it). I've never use CBuilder, but it's more popular than Power++ so likely as good or better. Both of these are probably very good, fast ways of getting an application written. Mad build in DB drivers.
5. Gtk+ A very nicely written, madly flexible widget toolkit with the added benefit of almost no baggage. In other words, at no point do you subclass some dialog box to start writing code. It's C dammit, there's nothing to subclass! This has great advantages, even for a bigotted C++ coder like myself -- it means that my application's structure, instead of the widget toolkit's, determines the structure of the code. The result is a much more maintainably laid out project. I also find gtk+ quite fast to work with. You can pack the dialogs in code and know exactly how they will turn out, write your handlers and you're done. Unfortunately there is no built in DB support anywhere -- it's a pure widget set, and you're on your own. You could always use direct SQL, or pick up one of several DB libraries I've never used (any recommendations?). Also, the function names are kind of long...oh well. Another plus: Gtk+ has a bajillion language bindings.
I've never used Qt so I'm afraid I can't comment...
I'll choose Gtk+ as the preferred widgetset. Applications written with it (rather than around it like some others) end up far more maintainable IMHO, and that's the most important thing for app developers.
Slashdot Mirror
by a university's choice to go MS -- it doesn't seem like they provide any solutions "big enough iron" so to speak.
... but the all-solid-state stuff probably wasn't)
Here at CMU (which is not that big a university, ~6000 undergrads) we run a massive distributed filesytem, afs, which is commercially available and was developed in part here (if we reimplemented the "andrew" system today it would probably be around CODA, which is Free iirc). The distributed fileservers have always been Solaris (and also a bit of HP-UX), and I'm not aware you could replace them with NT if you wanted to. AFS is supported by a a wide variety of clients: NT, Linux (and the other Unices), Mac.
The individual departments are not likely to give up their own special types of computers -- Design and Art want their SGIs, many of the professors (and students) use Macs, and the geeks all use Linux and Solaris. What solution based on NT can serve all those clients, for a system with tens of thousands of total users?
I just don't see any other way to run a computing environment the size of a university other than Kerberos and heavy-duty distributed fs stuff. Perhaps I'm missing something?
I certainly don't believe MS could provide anything like CMU's reliability. I've been here over a year now. Once or twice the routing has broken for a few minutes, and once the university blew a power feed and everything on the other side of the street shut down for a day (actually a lot of it was running on backup
General Processor Info.
Compare the SPECfp scores of high-end Intel and Alpha offerings. Take a look at a 600MHz PIII Xeon and a 667MHz Alpha 21264.
The reason to choose Alpha should be obvious.
So yeah, there has to be something made clear here: part of the reason Debian works so beautifully is that the distribution itself consists of ~3000 packages maintained by various people, all held to the same standards.
:)
:), the first thing I do is a
... I do :) in like 4 months, since there hasn't been a working version of gtkglarea. Unlike RedHat, there's no rufus repository with tons of builds of different versions of everything. On the other hand, in my experience (ran RedHat for a long time) a lot of stuff on Rufus didn't work because it was impossible to meet its dependency requirements, since there was no central control over the type of machine it was built on. In Debian, the latest version of each package in the Stable or Unstable (woohoo unstable!) trees is always built to work with the latest of everything else. It's only because of this that the auto-installed dependency thing works at all.
... let me set up a masquerading ppp line machine in about 3 minutes (of course less 45 minutes toying with isapnp to get the bloody pnp modem to work :). I just answered the config questions, gave it the network info, phone number, account, and password, and BANG! it worked. Hehe, way easier than doing the same thing years ago under Slack 3.2 :)
:) It's pretty ugly, and not to be approached if you mind rtfm'ing ...
The result is that for 99.5% (at least that's the number on my system) of stuff you want, you use the distro's official version/build, which is guaranteed to work perfectly. On my system, I have ~530 packages installed. There are only 3 that are not native Debian packages -- the only 3 things I've found that I wanted that weren't available. So I downloaded the rpm and installed 'em with alien
It's kind of a similar philosophy to BSD, in that the OS consists not just of a kernel and a set of basic stuff, but of everything. In a BSD this is true because everyone's running THE official BSD -- no distros. In Debian, there's the same kind of mentality. By comparison, RedHat's distro is sparse (granted most of the difference is in packages that very very very few people are interested in).
Whenever I find out about a new program I find I need (happens all the time, this is Linux after all
dpkg -l "*progname*"
to see if there's already a Debian package. Exactly 3 times, there hasn't been one. If Debian wasn't behind on processing developer applications (they're focusing efforts on catching up with bug reports) I'd have already built debs for these apps.
Sometimes the system doesn't work so well: as an example, there hasn't been a working version of XTraceroute (maybe you don't consider that essential
The result is impressive.
I've run a lot of both distros. I came to the conclusion Debian was superior because it has everything in a compatible, working version. That makes my life a lot simpler. In addition, Debian packages have Config scripts that make setting up a lot of stuff stuff easy even if you're not familiar with the specific config file formats
Man, that was way offtopic. I do intend to try Corel's stuff tho. I hope the license on their config stuff is good, cause Debian's idea of a "user-friendly" install is, the "voice" of the person who asks you various questions and tells you what's going on (in text mode) during the install, is pretty chatty and likable
I've used 3Coms in Linux boxes for years -- from terrible old 3c501s to 3c509s and now a 3c905. Don Becker's drivers have always worked perfectly for me.
I obviously shoulda checked my references rather than do this off the top of my head....
:)
... that would be interesting. I wouldn't mind an Ultra based subnotebook :)
I'm only saying: I don't see what advantage microII has in embedded applications -- I don't see a particular incentive for embedded chip designers to pick it up.
Does anyone know what its power needs etc. are? Have there been advanced, low-power versions of this chip produced? The microII 85MHz in my SB3 is pretty power hungry
Now an embedded Ultra
The MicroSPARCII is not a particularly exciting processor. It's the chip that was used in the SS5, and it ran up to 170MHz in the last SS5 offered. It's not designed for SMP (Turbo- and HyperSPARCs did that in SS10 and SS20). It's a 32bit chip, a bit faster at floating point than an equivalent Pentium. (it's also the chip in my SPARCBook 3 :).
:) But it does on the ARM, Motorola and MIPS chips too.
So don't expect cheap high powered crazy Suns floating around soon. Sun wants people using MSII's where they're now using R4xxx's and ARMcores and m68k's in PDAs.
An advantage is of course Linux already runs fine on it
SGI has to redefine who it is. Of course the logo thing was a mistake -- long live the old SGI logo!
:)
:)
...
They have a serious problem: you don't need a crazy SGI station to do graphics anymore. You'll do decently well with a high-end PC, 3DSMax, and good consumer graphics hardware, and for a fraction of the price. There was a reason so many people fell in love with equipment like the original Indigo (mmmm....purple box): it got the job done in a much more primitive age. It was the only way.
I recently heard a story about a little graphics design company a couple of years back (not all that many years, either, maybe 2 or 3). They didn't have money for SGIs, so they invested in PCs and MAX back when MAX was the first product for PC that didn't suck. They would get clients interested with samples, then give a tour of their office. When the clients saw they used PCs, they ran away. So then they got the clients interested, but made excuses why they couldn't visit the office ("We're painting...","We just moved...","There was a water leak, it's a mess..."). After that they started landing jobs and making money
But nowadays you can equip a 3D artist with a $2k PC and a $2500 copy of MAX and be in good shape, rather than SGI equipment and software totaling 4x as much.
So that's their problem. I don't know what the solution is. Linux is great, but it runs on cheap PCs too
I guess if I knew what the solution was I'd go be the CEO of SGI
a) Unlikely. It's not necessarily that well suited -- xfs is useful on servers, where fscking for a long time (say over several hundred gigs of RAID diskspace) can be a real turnoff. But xfs is a big-iron solution -- it has overhead that will make it inappropriate for workstations with limited diskspace. In addition, there are efforts (which may benefit from xfs) such as Tweedie's ext3, which will journal and may also be more reasonable for lighter weight applications (you'll have to ask him about that tho ... ).
... you might really want it in there. I think there's a good chance we'll see xfs pretty much straight dumped into the kernel in the near future (perhaps in the 2.5 series). Later we'll see (hopefully) more elegant solutions based on previous work, including parts of xfs.
...
Put it this way: if you haven't lost a day of productivity for a whole team of people due to a server fscking its drives after an unplanned/unclean shutdown, you don't care about xfs. On the other hand, if that's you
Anyone care to comment on more precise schedules for these projects?
b) Probably none. SGI's commitment to Linux is too strong. Better yet, as a strong point of Free software, all they have to do is release the source and let the Linux community do the work.
An interesting property of the OS: If MS goes out of biz or decides to drop some OS, you're fsck'd. But while a key programmer on a Linux component losing a job and being forced to go elsewhere to work on unrelated stuff to make ends meet can hurt a Free project, it can't kill it, the way Amiga abandoning its userbase will (IMHO) kill the Amiga community, which can no longer expect any progress on a completely closed system.
Offtopic-bait hint to Amiga: Free the source, if you want it to live
I know a lot of double majors (and plenty of double degrees too) which makes the pool pretty mixed.
... don't know why. But it's not that many.
I might agree there are more (in my observation) socially inept ones in Physics
as well as of massive overdiagnosis of things like ADHD/ADD.
... ) because it's comfortable and they honestly have more important things to do than care. When they do choose to dress well they have no problems doing so...
I suspect this psychologist is reading things a bit far into these results. While is possible there are a few highly intelligent technical people who are mildly autistic, I think it is very rare.
I think it's much more likely that parents who were themselves clumsy and socially inept bring up distant, socially uncomfortable children. This may be part of the "nurture" makeup of autism that these researchers are considering, but I have trouble seeing it as a disease.
Speaking from pretty decent samples -- TJHSST and CMU SCS -- I think it's far more likely for nerds to
a) Do ballroom. Nerds dance as well as anyone and we're always looking to meet girls.
b) Do S'n'S or have other interest/background in acting and theater, including film.
c) Be perfectly socially graceful, except for lapses of lust over hardware, and a deep enjoyment of conversations no one else can understand.
I strongly believe that physicists have a reputation (well deserved IMHO) for being poorly dressed (atrocious prevalence of khakies/sneakers
I've played with Dreamcast -- though I only played 3D Sonic. I loved it. I don't know if I would buy one (haven't seen the other games) but I wasted a lot of time at E3 on it :). The graphics were way ahead of the other consoles, but that's not saying much. I don't think the graphics were up to par with a high end nVidia or 3dfx.
... ) and 3dfx and nVidia will be out with new cards in the spring. PSII will realistically be available for XMas'00 (i.e. will definitely be ready). If it's ready this spring, it'll be pretty cutting edge for about 6 months. If it's ready in the fall (I think this is more likely) it'll already be outdated.
:) but I'd rather pay $100-150 every year for a new 3d card for my PC that will keep me up to date, and since I already need a new PC every ~18months, than pay $450 for a console that will be cool for a few months, then old for years to come while I can't buy the competitor's console and still play the same games... but then I'm kinda PC-biased...
That said, I don't expect to be impressed with PSII. Sure the numbers sound good now (even accounting for the blatant lies that they put out -- don't believe anything like those poly rates, or any other numbers from any 3d hardware company), but the PC industry moves damn quick. GeForce sounds pretty bloody slick (I want
Not that I do spend money
Cache is not something you get to control directly in code.
When you read a dword from main memory (i.e. not already in the cache) and then do operations on it, the cache controller takes advantage of a free memory bus to go ahead and read the rest of the cache line. Assuming your code is well optimized for cache performance, the next things you read should already be in the cache.
If you're doing a lot of kernel stuff, large chunks of the kernel will be in the cache, as you would want. And if you're running Quake 3, Quake 3 will be in the cache. It's exactly what you want.
I believe (get me if I'm wrong) Rob has reduced the number of moderators from when he originally introduced it. He's said before that he feels that moderation should be a rare thing and only a few comments to an article should ever be moderated in either direction.
I didn't agree initially, but now I do. I try to read everything at least down to 2, and I think this means you get all the intelligent comments, not just the ones lucky enough to make it up to 5.
Of course, there are still a few rare ones that make it up to 5, and some of them don't really deserve it. IMHO moderators should focus on looking for diamonds in the rough rather than further moderating up 3's and 4's.
Perhaps moderating up an article should get harder (requires more points) the higher it starts, so it would cost a moderator a lot of points (or alternatively take several moderators) to bring a 4 up to a 5, but only one to take a 1 to a 2.
Obviously ripping (assuming a good ripper eg. cdParanoia :) should be perfect. You should end up with the same bits on your HD as on the CD. Of course the encoder has something to do with this too (since mp3 is inherently lossy).
You've got it right when you say the real problem is the sound card -- even an SBLive! gold is pretty noisy in analog mode, and the inside of your computer might as well be a freaking radio station. But with the Gold (and I believe certain versions of Diamond's MonsterSound, tho I'm not sure) you can output in digital, which in theory could be hooked up directly to your stereo.
At least one of the digital out's on the SBLive! Gold is a proprietary (i believe, does anyone know the name of the format/standard it uses?) plug that works with the FPS2000 digital speaker set. It's good sound for a computer but not exactly audiophile stuff. I'm not sure if there are any standard (eg. Dolby digital? SPDIF? you tell me...) outs on the SBLive! Gold.
There are however definitely Dolby digital outs on many DVD players, and you can hook those up in digital to a good component system.
It's perfectly possible to do 3D in software. It's even somewhat more interesting, since you're not bound to and 3D card's paradigm. Software is where you get cool stuff like true voxels etc...
Glide probably wasn't out before DX3, but DX3 was pretty much useless (only a very minimum 3D API) so Glide may have beaten anything useful, tho not by much...
------------------------
Replying to another comment:
No, an API cannot be faster. But an implementation can. And an API's design can affect an implementation. In any case, Glide (the implementation) is a hell of a lot faster than D3D (the implementation), as per my original comment.
Microsoft writes each version of D3D by asking the manufacturers "What features do you guys need?" and then writing them in.
:)
...
There were no 3D games bfore D3D. No one had cards.
Just 'cos a card supports D3D doesn't mean you can assume your program will work right. You still have to test and debug each individual card. This is the voice of experience
Matrox's bump will be in D3D i'm pretty sure...
People use Glide rather than D3D 'cos it's way faster. Speed really is all that matters
I read another article (the one on rivaextreme ... pretty good article), I can add a few more comments.
... imho this is kind of limiting ... we'll see.
... I again expect GeForce to better that ...).
GeForce has environment mapping (iirc so does Permidia3) but not bump mapping.
It can do 8 free hardware accel'd lights
A Voodoo3 on a fast machine under Glide can handle about a 10-12kpoly scene lighted textured w/effects and phsyics running about 20-25 fps on a 450a. I'll be very impressed if GeForce can do twice that -- 25kpolys at 25fps, or about 500k real polys/sec (BTW a Voodoo3 under ideal conditions w/out features can do 500k "fake" polys/sec
But 15million polys/sec is the kind of bloated number that usually comes out of graphics shops. Don't believe it for a second.
As for 100kpoly models lighted w/fx running smoothly, i'll believe it when I see it.
if DaveS or DaveR wants to correct me on any of this stuff, go right ahead guys...
So basically nVidia chose to make a high fill-rate card with hardware lighting and transforms (geometry acceleration). These aren't innovative directions -- they were the obvious ones. None the less, the other major player, 3dfx, has pulled back from these choices. I'll explain why:
... everything is (we once demonstrated that it's physically impossible under DX6 to be faster than a Voodoo3 under Glide). There are some downsides: if you want to do crazy weird stuff with your lighting (eg. wrong faster stuff, funky effects) you may not be able to get it to work. Similarly with geometry -- special fast cases will become normal cases. So there may be a 50%-100% gain in triangle rate, but it's unlikely geometry acceleration will ever be able to provide much more than that.
... 3dfx talked about putting a geometry accelerator on V4 but I believe they backed off from it. Voodoo4 is however still an SST and therefore still a true descendent of the original Voodoo chipset conceived as a flexible, long-term solution for both PCs and arcade games.
... it may be finally time to kill some very old paradigms in 3d hardware...
nVidia has a card which can do supported operations fast. It obviously has a lot of fill. It'll be a good board. Of course it'll still be slow in D3D
nVidia seems to have chosen not to support the hardware bump mapping of the Matrox G400, an extremely high fill (runs beautifully bump mapped in a window in 1600x1200x32bpp) card without geom accel. 3DLabs' long awaited Permidia3 will also have some kind of hardware bump. IMHO this is a relatively flexible feature -- you could do a lot with it. It remains to be seen how flexible nVidia's lighting and geom turn out to be.
I'll be impressed if D3D ever delivers real hardware geometry benafits. We have yet to see a single benefit of DX6 over DX5 (not screwing with the fp control word especially) actually work. I'm highly suspect of anything MS sez.
So what about the remaining behemoth, 3dfx? Their Voodoo4 is supposed to be an extremely high fill card (fill has always been their hallmark). It may not support any more hardware features (eg. bump, lighting, geom accel), but it will fill like crazy. It's supposed to do full screen anti-aliasing
I'm eagerly awaiting the new generation. But I expect the real crazy stuff to start happening in the following generation
I've been running Q3A reasonably happily on my TNT 2 Ultra for a couple of weeks now. The machine is a Celeron 450a.
The GLX renderer isn't amazing -- I have to turn off lightmaps, and I run in 640x480 (haven't really messed w/higher res modes, the card may have plenty of fill to do them...). I'm still waiting for DRI to get full performance out of the board.
Nonetheless, q3 ought to run acceptably on even a TNT with a few more of the settings tuned down...
Okay, there's not a Linux hacker on the face of the planet who wouldn't kill to have your job.
- Paid by RHADL.
- Wake up when you want, work when you want.
- Go to all the big trade shows.
- Work with the likes of Linus and all the other regulars.
- Get free toys (and I mean good toys) like PA-RISC systems from HP and Athlons from AMD.
But it wasn't always that way. Back before even RedHat paid you, back when you hacked on your aging spare equipment, what drew you to it? How did you know this was what you wanted to do before you knew about all the perqs involved, or that there would ever be perqs?
A language without excellent, maintainable, preferrably native C/C++ bindings is useless.
I can't walk into a contract job, inform them that I'll code their solution in Obj-C, and then do it and split. I'd be leaving them with something they have no prayer of maintaining or extending.
Java is unacceptable -- too slow, no particular reason for existence.
Obj-C is unacceptable -- if it's as good as Smalltalk, it probably belongs in the same catagory (along w/ML and others) of academic languages of no practical value.
Tell me what Obj-C does so much better than C++ that it's worth the cost of conversion.
I think the best way to succinctly describe OpenStep is to say it is to software developers what the Macintosh is to normal users.
Considering I can't for the life of me figure out how to use a Mac (it doesn't have any nipples...it's not intuitive in my book and I couldn't find the man(1) command) I don't think I'd enjoy it.
Seriously, the Mac's a pretty limiting environment. It's a padded room. Is that really where you want to be coding? I sure don't.
Ian points out I didn't consider GNUStep. I'll do so now:
It's called "Objectionable C" for a reason.
Go code in C++ you weenies!
(warned ya I was a bigot)
but I'll attempt to justify myself anyhow. I'm choosing to skip the whole display driver architecture question and focus on application programming.
First of all, I don't believe XML/whateverML/Mozilla to be the answer. I don't want my application written in a scripting language. Running everything through your browser is a bad idea, and everyone except MicroSoft thinks so. Frankly, I haven't seen a browser that doesn't suck (except Lynx!) and I'm not convinced I ever will. I'm not that impressed (yet) by the Mozilla efforts, tho I support them.
I'm only going to consider compiled widget toolkits, no Java. I'm sorry. Java sucks. It involves way too much typing, and it's slow. I don't approve of anything that makes code slow. The whole point is to be fast.
What does that leave?
1. Motif I once sat thru a college lecture on Motif. In an hour, the lecturer had managed to cover use of the clipboard. No function name was less than 20 characters, and it took 15 calls to get the most basic thing done. No wonder Netscape under Motif was ~1million lines. 'Nuff said.
2. MFC MFC isn't really so bad as a design. It can be annoying, and it often lacks flexibility, but it's not completely unusable. However, I find that using MFC (specifically in VisualC++, which conveniently writes the hard/annoying parts for you) is a good way to kill a project. It forces you to use their class heirarchy and objects, and if the generated code gets screwed up you can be in real trouble. The usability is also heavily dependent on spotty MS documentation -- if the documentation is bad (eg. the MDI multimedia devices in VC5) figuring it out can be a chore. A mediocre choice at best. At least it has nice built in database drivers.
3. Visual Basic has got to be the quickest way of throwing together a UI I've ever seen. I really like it, especially for prototyping. I have no speed problems (since it's compiled now and officially in my OK book), but BASIC can be a dreadfully limiting language. Writing the interface in VB and all the hard code in a DLL isn't a half bad idea, tho. At the very least useful for trying out interface designs on people.
Some built in database drivers.
4. PowerSoft Power++, Borland CBuilder I've only ever used Power++. It's really just a wrapper around MFC classes, but it cleans a lot of stuff up and does all the hard stuff. It still pushes you to use MFC-type objects, but they're quite a bit less ugly to work with, and the docs are better. But if the rest of you're code's not in Watcom, Power++ might cause compiler issues (which is why I eventually dropped it). I've never use CBuilder, but it's more popular than Power++ so likely as good or better. Both of these are probably very good, fast ways of getting an application written. Mad build in DB drivers.
5. Gtk+ A very nicely written, madly flexible widget toolkit with the added benefit of almost no baggage. In other words, at no point do you subclass some dialog box to start writing code. It's C dammit, there's nothing to subclass! This has great advantages, even for a bigotted C++ coder like myself -- it means that my application's structure, instead of the widget toolkit's, determines the structure of the code. The result is a much more maintainably laid out project. I also find gtk+ quite fast to work with. You can pack the dialogs in code and know exactly how they will turn out, write your handlers and you're done. Unfortunately there is no built in DB support anywhere -- it's a pure widget set, and you're on your own. You could always use direct SQL, or pick up one of several DB libraries I've never used (any recommendations?). Also, the function names are kind of long...oh well. Another plus: Gtk+ has a bajillion language bindings.
I've never used Qt so I'm afraid I can't comment...
I'll choose Gtk+ as the preferred widgetset. Applications written with it (rather than around it like some others) end up far more maintainable IMHO, and that's the most important thing for app developers.
this is a cheap ploy to get attention by being posted to /.?
... features...
Seriously, they make games. Games never have bugs, take it from an old hand. Only features.
Ask our graphics programmer, Dave Rosenthal. Not a day goes by when his code isn't full of