|
Welcome to the COE Discussion Forum!
|
|
|
|
To participate in the discussion forum, you must be logged in to the website. If you forget your login information, please contact COE Headquarters at coe@coe.org or (800) 263-2255.
If you are new to the COE Discussion Forum and would like to participate, please register.
|
|
| Forum Highlight: CATIA V6 |
|
|
 |
|
 |
| You are not authorized to post a reply.
|
|
| Author |
Messages |
|
COE-FORUM-USER
 |
| 29 Jan 2003 05:11 PM |
|
[[[but the present PowerMac doesn't provide any clear hardware advantage here. That is why I believe the PowerPC 970/980 will play a role in increasing market share for the platform. It will conceivably draw scientific and 3D users over at a higher rate than present.]]]
See, this is where I don't agree with you here... I was hoping you read my earlier post and understood what I was getting at.... In short a 4 GHz Pentium will probably get you nothing in the way of a completely usable desktop from a perception stand point. And there comes a time when no matter how many MHz. you throw at it, the enduser will not be able to discern performance UNLESS ... it comes down to really high-end 3D rendering. If you're going to be working with it during a design and modeling session then a person sitting behind a 500 or so MHz. PC isn't going to tell the difference if he was suddenly placed behind a 2 GHz. PC. I've taken the blind taste-test myself... There is no noticed difference... Remove the term Mac and G4 from your argument and replace it with "Pentium 500 MHZ" or something of that ilk... Suddenly you would think that all those low end PCs would be FLYING off the shelves and companies would be upgrading more often because the price is right... the sad truth is that it isn't happening... The PS sector is a slump and not even those cheepie PCs are selling. And let's be honest here.... the types of workstations that people buy to run this stuff isn't the "under $1500.00" PC machines ... companies spend a lot more than that. At work my cousin apparently ordered 10 new Dell monster desktops (this was about a year ago I think) the added "must have" option seemed to be this $700.00+ graphics card that amounted to nothing more than a waste of money since no one can figure out when all this supposed performance was going to arrive... They were sorely disappointed... Again, these companies aren't buying the cheap PCs.
--
Ed M.
Author: Ed M. |
|
|
|
|
COE-FORUM-USER
|
| 29 Jan 2003 05:23 PM |
|
Jim, it seems to me that your entire argument boils down to something like this... "there is never enough speed". I disagree...
[[[While the IBM Marketing claims that V5 will run on a 350mhz Pentium 2 w/ 128MB of RAM, don't believe them. ]]]
When CATIA shipped for Windoze are you telling me that it wasn't a usable product because the Wintel crates at the time weren't 2GHz. Xeon processor based?
[[[While you may be able to do simple drafting on a Pentium 2, I wouldn't attempt a Digital Mockup session of a portion of an aircraft with anything less than a 2GHZ Zeon ]]]
But when CATIA V5 came out it was for Win98/2000 and those machines weren't as fast as the systems you state, yet CATIA was ported over -- to Windows (Yikes!) A an engineer, you must also be aware of the tradeoffs ... you say that you're team will be *that much more productive* given raw speed... I'll remain skeptical. What about the costs to power consumption and heat, not to mention the support costs associated with a Wintel environment. We've been through this before, you can't selectively exclude these costs. Anyway let's continue on with some of the things you mention...
[[[Years ago, when I was sharing a MicroVAX 2 with 2 other user's, I was happy to get a couple dozen parts on the screen. When I moved to the Mainframe, I could get up to 100 Part on the screen. In today's environment, my designer's are pulling up hundreds and even a thousand parts on the screen on a regular basis. ]]]
OK, I see where you are coming from and where you are going, but to be quite honest, I doubt that any of you actually focus on 100 parts at any one time let alone *thousands* What I'm saying is that a human can only do so much at any given time. However, it begs the question of why you'd want or even *need* *thousands* of modeled parts up on the screen simultaneously. You'll have to explain that to me.
[[[Give me a workstation (Intel, PowerPC, or whatever) with 10 times the power of today's machines, my users will pull up 10 times the information. And be more productive because of it. ]]]
Yeah, right That's a bit overstated don't you think? As I said before we're only human and there is only so much we can do at any given time.
Hey, I hear IBM is taking orders for their ASCI White.... I'm sure one of those would keep you guys from EVER needing to upgrade. What's more you could add engineers in droves and the system would hardly break a sweat... The cost savings would be HUGE up front, but would be peanuts over the long haul if what you say is true... Think of all that work you'll be getting done. (just a bit of sarcasm to make my point so no disrespect intended).
[[[While CATIA isn't designed for the likes of Pixar and Lucas Films, it does have a photo-realistic product that is capable of creating truely photo-realistic marketing type images.]]]
Well, for the money, you could get a superior renderer in Lightwave from NewTek. I'm not sure I understand your argument. Besides, if you need higher detail, there is always PR-Renderman. So what gap is CATIA trying to cover? the high-end rendering environment? The midrange? For the money there are vastly better products more suited to the task. And consider the fact that you're not going to be doing any of this rendering on a single machine.. certainly not a Windows box... It will be done by render-farms that consist of machines solely designed for that purpose and used for nothing else but rendering -- probably running Linux. I'm not sure what you are stating... However, there's a few questions I'd like to ask you...
- Would a dual-1.42 GHz. Mac runnig OS X be enough to run CATIA? Simple question.
- If not, what do you think it would take as far as Mac hardware is concerned, to effectively run CATIA? And don't even tell me that you guys are running those sub $1000.00 2GHz. pentium boxes that are for sale at the local Radio Shack and Best Buy...
--
Ed M.
Author: Ed M. |
|
|
|
|
COE-FORUM-USER
|
| 29 Jan 2003 06:26 PM |
|
Ed
>> If you're going to be working with it during a design and modeling session then a person sitting behind a 500 or so MHz. PC isn't going to tell the difference if he was suddenly placed behind a 2 GHz. PC. I've taken the blind taste-test myself... There is no noticed difference. <<
I have to side with Jim here - I can max my CPU out during a lot of intensive tasks in Cobalt and spend time waiting as the nice little wheel turns around.....
For more mundane tasks, of course speed is now at a level where you can't notice any difference between machines.
>> Remove the term Mac and G4 from your argument and replace it with "Pentium 500 MHZ" or something of that ilk... Suddenly you would think that all those low end PCs would be FLYING off the shelves and companies would be upgrading more often because the price is right... the sad truth is that it isn't happening... The PS sector is a slump and not even those cheepie PCs are selling. <<
Perhaps I didn't clarify things in my first post. I am only talking about this in relation to the possibilities in the higher-end workstation market, which is growing at present. I think the desktop market will continue to shrink, and I guess Apple is betting that it will - that's why they are investing so much in the laptops this year.
Anyway, this sort of sidetracks the original reason I popped over. I know several design studios who are looking at packages like I-DEAS and Alias here in Denmark, and these studios are running Macs now. They don't want to move onto Wintel, but the high-end programs aren't there on OS X. I think Dassault underestimates the size of the market. I have absolutely no idea how much it would take to compile CATIA to run in X-11, so it's impossible for me to judge how feasible this would be. I just know that there is a market in my corner of the world.
Cheers,
Brett
Author: Brett Patching |
|
|
|
|
COE-FORUM-USER
|
| 29 Jan 2003 08:31 PM |
|
Hey guys, I'm dropping back in here. Alot has changed with OSX-Jaguar since I first dropped in. Alias has come over to OSX which is just wonderful. A colleague of mine, a very well respected industrial designer in plumbing products has stated Alias to be "one of the greatest pleasures of life", which I might agree with. This is just great for us Mac users who want the best surface modeling program money can buy.
Perhaps I should explain why I am commenting in the first place: the new 17" Powerbook is just exquisite. I have a 550mhz Ti-book and it is feeling a bit slow anymore. I have a desktop in the office which makes it feel slow at least. So Ed, yeah - speed is really, really important when you are on a plane or in a hotel room trying to cram in some fillets on your latest model and the battery indicator is in the 00:59 and less range. And I agree with Brett - alot of the time I am sitting there and waiting for rebuilds and assembly file openings. Hey - I render during the design phase!!! I don't have a team of designers behind me rendering all of my stuff. If I am unsure of something or just want to look at my shape, I render it. And I work in Gouraud/HLS most of the time. And, I travel ALOT. I am a high-level designer in a plumbing company and I spend time on the road talking to customers and working/visiting trade shows and so on. So, a 17" 6lb. laptop is really exciting to me.
Another reason why I am here - I love the Catia/Solidworks system. I do alot of assemblies (unlike my colleague who loves Alias to death, I don't do as much surface modeling). It's not more than a few dozen models in the assembly, but it can get slow nonetheless. I just got into the software package in the past 12 months, but since my corp. "approved" the system people have been frothing at the mouth. We are heavy Pro/E. Alot of file conversions happening in design for sure, not as much in the engineering area as they just stick with Pro/E and the Dell $3500 "approved" workstation. But the other designers won't touch Pro/E and have no patience with the Wildfire noise and just wanna get to work.
But I travel. And I want that laptop. So please Dassault, bring us Catia on Jaguar!
Here is my other insight, which is very, very interesting:
One of my best friends is a senior networking engineer at one of the biggest dot.coms. He is a computer scientist/geek of the highest Gen X order (which means there are plenty of kids who can program stuff faster than him, but then again they don't sleep - yet). He poo-pooed the mac for as long as I can remember. I don't know, System 6? He laughed at me throughout college (we're now in our early 30's). Last week, over a cocktail at a nice NYC sushi bar, he said, "you know, the Mac is gaining alot of attention in the programming crowd. Geeks are moving to it in droves. Dudes with Titaniums are freaking about it's UNIX capability and are hiding away in closets with a fervor that I have not seen in awhile. Can you do WiFi on that thing?"
Yes John, I can. It's called Airport. I wander around my loft or about town and I am online. oh, sweetness.
OK, enough for now. In summary, I *do* need speed. I love that Apple is concentrating on laptops. And I want my software!
M
Author: Mick |
|
|
|
|
COE-FORUM-USER
|
| 30 Jan 2003 04:54 AM |
|
Hi Mick
>>Hey guys, I'm dropping back in here. Alot has changed with
OSX-Jaguar since I first dropped in. Alias has come over to OSX which is just wonderful.<<
I can't find an announcement about Alias Studio on OS X anywhere on the Alias|Wavefront site. All the news is about Maya 4.5. Tell us more!
/Brett
Author: Brett Patching |
|
|
|
|
COE-FORUM-USER
|
| 30 Jan 2003 06:40 AM |
|
It's quite a habit throwing the name Alias around - but what I really should have said was Maya, by Alias.
M
Author: Mick |
|
|
|
|
COE-FORUM-USER
|
| 30 Jan 2003 02:54 PM |
|
[[When CATIA shipped for Windoze are you telling me that it wasn't a usable product because the Wintel crates at the time weren't 2GHz. Xeon processor based? ]]
Welcome to my world. Actually, when DS released V5 R1, it was more of a "proof of concept" and "see how it's going to work" product. V5 really wasn't useable by most customers until they released V5 R7 or so. As for building aircraft, it really took until R9.
How much processor required really depends on what your design task is. If you are designing a new mouse or some kind of similar consumer product, and you only have 6-12 parts, then a 400mhz processor is probably sufficient.
If you are building aircraft with 10's of thousands of parts, then no, it wasn't really ready until we had both the R9 release, and at least 1ghz processors.
While not all designers need a thousand parts on the screen to design their little piece of the aircraft, many disciplines do require this. Just try routing flight control cables, hydraulic and pneumatic tubes, and electrical wire bundles thru the tailcone of an aircraft and you will better understand.
On aircraft and in automotive, many times single parts also will require a massive processor. Look at the design of a Landing Gear Trunnion, or of an automotive Cylinder Head. These parts are generally so complex that simply updating the solid can take anywhere from 15 minutes to 2+ hours on current hardware, and that's pure CPU.
As for the rendering, we do an awful lot of internal images to sell concepts to management, and we even generate these for marketing as well. We are currently using ALIAS, but we have been evaluating CATIA as well. One of the primary advantages of CATIA is that there is absolutely no data translation. If you have ever tried to translate data from one graphics system to another, you will soon realize that there is a tremendous amount of manpower involved in cleaning up the data after the translation.
You are correct, we are not buying $1000 machines at the local Best Buy. But our machines are not that much more. We are running on IBM Intellestation M Pro's, 2.2GHZ, 1GB Rambus Ram, SCSI Hard Drives, and 128MB FireGL2 Graphics cards.
But I am running pretty well on an IBM ThinkPad A31, 1GB Ram, 1.8GHZ Pentium 4, and a 16mb Radeo Mobile Graphics card. I'm not pulling up 1000's of parts, but it is working fine for the process development work that I am doing. IBM (and others) do have Lap tops that are Certified, and have mobile versions of the FireGL2 card.
Jim Strawn
Cessna Aircraft Co.
316-517-5851
Author: Jim Strawn |
|
|
|
|
COE-FORUM-USER
|
| 30 Jan 2003 06:13 PM |
|
[[[You are correct, we are not buying $1000 machines at the local Best Buy. But our machines are not that much more. We are running on IBM Intellestation M Pro's, 2.2GHZ, 1GB Rambus Ram, SCSI Hard Drives, and 128MB FireGL2 Graphics cards. ]]]
I didn't think you were running those bargain-basement crates... Still, you didn't answer my question... How do you suppose a dual-1.42 GHz. Mac running OS X Jag would perform?
[[[But I am running pretty well on an IBM ThinkPad A31, 1GB Ram, 1.8GHZ Pentium 4, and a 16mb Radeo Mobile Graphics card. I'm not pulling up 1000's of parts, but it is working fine for the process development work that I am doing.]]]
Hmmm... Are you saying that the current PowerBooks from Apple would fair quite good with CATIA too? Apple puts out some *sweet* portable Unix workstations ya know ;-)
Have you looked at them yet?
--
Ed M.
Author: Ed M. |
|
|
|
|
COE-FORUM-USER
|
| 04 Feb 2003 07:41 AM |
|
Since very little of CATIA is multi-threaded, a dual processor box is of not much additional value. There are some specific tasks that will make use of the second processor, but they are pretty specifically for Digital Mockup, not for day in and day out design work.
While it is nice to speculate how well CATIA might run on an Apple, the cold fact is that it doesn't. Perhaps DS will listen, and develop their next version on a more open platform.
But today, we have a limited set of choices. My main focus today is getting the existing software to work. And I don't mean functionally. The functionality is there today in V5. Our current focus is working on the business practices required to adapt to and make use of the new functionality in V5.
Jim Strawn
Cessna Aircraft Co.
316-517-5851
Author: Jim Strawn |
|
|
|
|
COE-FORUM-USER
|
| 04 Feb 2003 06:46 PM |
|
[[[Perhaps DS will listen, and develop their next version on a more open platform. ]]]
You would think that they would have listened to you (i.e., their customers), no? I mean did you *request* that CATIA be made available for Windows or was that a "Dassault" decision?
[[[But today, we have a limited set of choices. My main focus today is getting the existing software to work. And I don't mean functionally. The functionality is there today in V5. Our current focus is working on the business practices required to adapt to and make use of the new functionality in V5.]]]
Hmmmm... not sure how I read this ... does this mean that it would only complicate things if another platform was supported? A touch of sarcasm perhaps ;-) I'm suspecting that you are referring to training... If that's the case then this is a *criticle* point since 90% of people stick with things that are weaned on. This course helps to promote the use of Windows-centric systems and totally ignores why such a monoculture mentality might be bad. In any event, it would be interesting to see what pans out in a couple of months. Apparently Apple is in talks with all the high-end CAD developers. I'm assuming that Dassault is one of them. And you seem to keep forgetting about those new IBM workstations that are just around the corner... PowerPC 970-based. 4-way configurations initially (as hinted by IBM). These systems were NOT DESIGNED TO RUN WINDOWS. They run LINUX. Wouldn't you think IBM would want CATIA on these boxes? I think they would... And as you know it's likely that if IBM requests it, it's gonna happen -- simply because it's IBM. A LINUX version also indirectly implies an OS X version (see my other posts). I've provided the pieces of the puzzle, I'm sure you can figure it out.
--
Ed M.
Author: Ed M. |
|
|
|
|
COE-FORUM-USER
|
| 05 Feb 2003 07:50 AM |
|
[[[I mean did you *request* that CATIA be made available for Windows or was that a "Dassault" decision?]]]
Yes, the user community asked for this, back when DS decided to start developing V5. Unix was expensive and had little other software applications developed for it. Linux didn't exist at that time, and the MAC was foundering, except in some niche markets (desktop publishing). The Open Software movement wasn't even a blip on the radar. The business side of most companies was pretty deeply entrenched in MS Office, and the engineering departments were tired of supplying their engineers with 2 computers to perform their work. PTC was eating their lunch, and they realized that a simple port to Windows wouldn't do it. They knew that they needed a completely new generation of CAD.
[[[does this mean that it would only complicate things if another platform was supported? ]]]
Absolutely. If it is just a port to a new platform, it's not worth our time - and if they do it as more of an emulation (like they are doing with Unix right now), then it is a complete waist of time. If it provided significantly improved functionality and performance, however, then it would be worth it. CATIA V4 is currently supported on AIX, HPUX, IRIX, Sun-OS. We stick with AIX and fight anyone who want's anything else, as it just complicates things immensely. Applications have to be developed, compiled, and tested on all platforms. Bugs have to be checked and verified on all platforms. Different environment settings and installation packages need to be maintained for all platforms. Etc, etc, etc. It's not worth it.
As for being stuck with the system you were weaned on, that's a bogus arguement. I started out on Interpreted Basic on a CDC Mainframe, moved to MS DOS 2.15, Windows 2 (yes 2), Windows 3.1, 98, NT, & XP. I have used HPUX and AIX. My first CAD system was on VAX/VMS (which still has functions superior to anything available today), moved to Mainframe CADAM/CATIA on MVS, moved to Workstation CATIA on AIX, and am now using V5 Catia on both AIX and Windows. All of the engineers here have been thru this same cycle (although most aren't old enough to remember VMS and DOS, let alone interpreted basic and Contol Data).
Jim Strawn
Cessna Aircraft Co.
316-517-5851
Author: Jim Strawn |
|
|
|
|
COE-FORUM-USER
|
| 07 Feb 2003 09:00 AM |
|
I think we can all assume how Ed will vote, but any one else interested in Mac OS X port of Pro/E should vote "yes" on the Mac OS X Pro/E Port Survey @ Pro/E Digital Digest below. Note than I'm aware that this is Dassault site, but a port to Pro/E could spur a port to Catia and/or SolidWorks...
http://www.d-digest.com/proedigitaldigest/V3I6/
David Chaney
Mechanical Design Engineer
Author: David Chaney |
|
|
|
|
COE-FORUM-USER
|
| 07 Feb 2003 09:02 AM |
|
I think we can all assume how Ed will vote, but any one else interested in Mac OS X port of Pro/E should vote "yes" on the Mac OS X Pro/E Port Survey @ Pro/E Digital Digest below. Note than I'm aware that this is Dassault site, but a port to Pro/E could spur a port to Catia and/or SolidWorks...
http://www.d-digest.com/proedigitaldigest/V3I6/
David Chaney
Mechanical Design Engineer
Author: David Chaney |
|
|
|
|
COE-FORUM-USER
|
| 07 Feb 2003 09:03 AM |
|
I think we can all assume how Ed will vote, but any one else interested in Mac OS X port of Pro/E should vote "yes" on the Mac OS X Pro/E Port Survey @ Pro/E Digital Digest below. Note than I'm aware that this is Dassault site, but a port to Pro/E could spur a port to Catia and/or SolidWorks...
http://www.d-digest.com/proedigitaldigest/V3I6/
David Chaney
Mechanical Design Engineer
Author: David Chaney |
|
|
|
|
COE-FORUM-USER
|
| 25 Feb 2003 10:19 AM |
|
You are underestimating Intel and Microsoft. I have no great love of either of these companies, and I would very much like to see more competition, but history has shown that they both are incredible competitors (that sometimes resort to illegal tactics). My prediction is that within 3 years there will be a fully 64 bit implementation of windows out there.
Intel is saying that there is no market for 64-bit desktop computers primarily because a) they don't have a 64-bit x86 chip ready to ship, and b) they are starting the FUD campaign against AMD. Rumor has it that there is in existance an Intel chip (Yamhill) that is a 64 bit implementation of x86. Intel is hesitant to release this chip, most likely so that their "cheaper" chips don't cut the margins of their new Itanium line.
Secondly, it is also rumored that Microsoft has a 64 bit version of windows already running on AMD's x86-64 hardware. Windows NT was originally written on MIPS chips, so I imagine it is quite a bit more portable than you think.
Thirdly, nobody except computer science professors and people that own RISC chips really care what instruction set a chip uses. Everyone bashes x86 as if it has been the same since 1979. It isn't: MMX, SSE, SSE2, etc. have all been added on. The segmented architecture of the 8086 and 286 sucked, I will grant you that, but the 80386 and up all have 32 bit modes with flat address spaces. The instruction set may not be as "elegant" as the Alpha, MIPS, or PPC, but so what? The x86 code size is smaller than just about every RISC architecture, and RISC chips have had to resort to an instruction decode step in their pipelines now because their instruction sets are no longer "reduced" (for example: Altivec has been tacked on to the PPC). x86-64's 64 bit mode is a clean 64 bit implementation with a flat memory model. It doesn't get much better than that. You can argue about which instruction set is more "elegant," but at the end of the day the person sitting in front of the terminal just does not care.
If AMD has any success with x86-64, Intel will release a competing 64 bit x86 product within a year. This does present a problem for people who need more than 2GB of application memory right now, I will admit. Then again, most of the CAD packages that are available right now, even on UNIX, are compiled as 32 bit apps, so the 64 bit question is moot until Dassault et al release 64 bit apps.
Finally, Windows may have not been the best choice for Dassault, but the question you need to answer is "Why is OSX better?"
Author: Julian Laxton |
|
|
|
|
COE-FORUM-USER
|
| 25 Feb 2003 11:17 AM |
|
[[[Intel is saying that there is no market for 64-bit desktop computers primarily because a) they don't have a 64-bit x86 chip ready to ship]]]
Yep, that's exactly right.. It isn't ready when others will be ready FAR (years?) sooner.
[[[Rumor has it that there is in existence an Intel chip (Yamhill) that is a 64 bit implementation of x86]]]
Which is still a little different from AMD's X86-64 implementation; enough so that in can probably be considered a different platform.
[[[Intel is hesitant to release this chip, most likely so that their "cheaper" chips don't cut the margins of their new Itanium line.]]]
Yep, now you see the bind they were in... Oh, and you forgot to mention the MHz. thy've been preaching for so long will likely come around to bite them ;-)
[[[Secondly, it is also rumored that Microsoft has a 64 bit version of windows already running on AMD's x86-64 hardware. ]]]
SERVER hardware. This is different from a Windows *desktop* Microsoft has no plans for a 64-bit version of the Windows desktop primarily because there is no 64-bit Intel chip ready to ship. What will they test it on? lol Be serious.
[[[If AMD has any success with x86-64, Intel will release a competing 64 bit x86 product within a year.]]]
For SERVERS.
Anyway, to clear up a few points...
It's my understanding that Windows running on those non-Intel systems, especially Alpha, ran dog-slow (on an Alpha!?) Furthermore, applications had to be recompiled to run on it. How many will be recompiled this time? How long will it take for developers to release robust apps? Oh probably not many since it's a completely NEW platform that will need to run older apps through emulation, which is really out of the question.
So, will developers be asked *yet again* (and so soon) to throw significant $$ into application development for these new and separate platforms running the SAME OS? Can you answer that? It's also my understanding that anyone who was interested in shelling out for Alpha systems were running OSF/1 - Tru64 UNIX or OpenVMS. It would appear that the Itanium is running into the same debacle that the Alpha did with respect to Windows. So you keep right on believing that Intel and Microsoft will provide you with the rosiest of futures and 64-bit desktops within 3 years with all the current apps to go along with it... You've been buying it up to this point. And you still missed the point completely.
By MS only developing Server 2003 for IA-64 tells us that developers and endusers alike will now have to worry about 3 to 4 completely different platforms running the same OS: x86-64 and IA-64 and x86-32. AMD's 64-bit isn't exactly what an Intel x86-64 implementation would be so that also has to be factored into the equation as perhaps a half-platform. You act like Windows ran like a champ on those other platforms when in fact it was a dog and thus, not nearly as portable as Microsoft marketed, which was my point.
Just more confusion, chaos and compatibility issues for developers and endusers to consider. I mean what about all those nice brandy-new desktops that were sold to all you guys this year? Those legacy systems aren't going to go away any time soon. Are people going to ask to upgrade *again* ? Will OEM suddenly drop support for those systems? It might have considerable blow-back.. people might indeed update -- to a platform that's much more forward looking.
--
Ed M.
Author: Ed M. |
|
|
|
|
COE-FORUM-USER
|
| 25 Feb 2003 12:01 PM |
|
I love talking to the macintosh faithful. It reminds me of when I was a fan of OS/2.
Anyway, to answer your points...
1. There is still no 64 bit macintosh. Will there be one with a 64 bit operating system before a 64 bit x86 implementation with a 64 bit operating system? Nobody knows. I'd say the odds are even right now.
2. I have no idea if Intel's 64 bit x86 implementation is the same as AMD's. In it's current state, likely not. Will they release a x86-64 implementation if AMD's chips take off? I don't know. AMD didn't patent the instruction set (as far as I can remember), so Intel could clone it. Even if AMD has patented it, I'm sure they would be willing to cross-license it with Intel in exchange for SSE3 and upcoming Intel extensions.
3. Intel is having to back off the Mhz marketing already - their new "Centrino" or whatever the mobile chip is called runs at low Mhz but apps run about the same speed as a P4 with a much higher clock. It's all marketing - what else can you say? Just about every chip company has resorted to other performance measures besides clock speed, and intel will have to now as well. At the end of the day, if the chip is fast, people will buy it.
4. AMD provided a x86-64 simulator to microsoft and the linux community long before there was any x86-64 silicon. Intel will do (or has already done) the same thing. Microsoft can test to their heart's content on the simulator.
5. NT on Alpha and MIPS was actually quite fast, but it had to be running natively compiled applications. NT/Alpha could run x86 binaries; the problem was that running x86 binaries via the FX32! emulation package was not as fast. Unfortunately, changing CPU architectures is not easy (as I am sure the Mac faithful well know). Recompiling to get native binaries is a fact of life that will not go away.
NT was developed to be portable primarily because Microsoft and Intel were not the best of friends at that time in their corporate histories. This was a move by Microsoft to put some pressure on Intel. At the time, the other RISC architectures had a sizeable performance advantage over Intel. This performance lead has mostly evaporated, but Intel chips remain cheaper than their RISC counterparts. This is why NT only runs on Intel these days: there's no cost or performance benefit to running it on any other platform. I am not a big fan of Intel, but they can turn out huge numbers of chips for low cost, something the RISC vendors (and AMD) can only dream about.
6. I really don't understand the difference you are trying to define between making people port apps from a 32 to a 64 bit version of windows, and a 32 and 64 bit OSX. If you want 64 bit functionality, the OS is going to need at best a recompile, and at worst a re-write. Is OSX 64 bit clean code? Then the apps must be recompiled, and you face the same difficulties. There is certainly uncertainty in the Intel/AMD side of the world as to which 64 bit implementation will win, but in the end I expect only one to survive.
Let me close this note by making one final point. Microsoft and Intel have been presenting "confusion, chaos, and compatability issues", as you said, for the last decade and a half. Yet, together they hold over 90 percent of the market. I can only conclude that they must be doing something right. What it is, I don't know, but I don't see that Apple is presenting anything new in the way of an alternative. I may very well be wrong, but I still don't see what is going to compel the marketplace to move to OSX.
Author: Julian Laxton |
|
|
|
|
COE-FORUM-USER
|
| 15 May 2003 09:09 AM |
|
Comparing AutoCAD to CATIA is like comparing a Ford Focus to a Freightliner Truck. Both of them will help you haul your goods, but one of them will haul a lot more.
While it may be nice to wish and dream that Dassault would provide a Linux Based (or OS X, or generic UNIX, or any other OS out there) CAD/CAM/CAE System, that will not happen until some time in the future, and most of us here are dealing with the present. CATIA V5 on Windows is the most functional and powerful Fully Integrated CAD/CAM/CAE System in existance today.
Jim Strawn
Cessna Aircraft Co.
316-517-5851
Author: Jim Strawn |
|
|
|
|
COE-FORUM-USER
|
| 15 May 2003 10:06 AM |
|
[[[Comparing AutoCAD to CATIA is like comparing a Ford Focus to a Freightliner Truck. Both of them will help you haul your goods, but one of them will haul a lot more.]]]
Viewed from another angle, it tells me that there a lot more things to worry about when things go wrong. I just think it's a poor analogy, but I get the gist. And I still think it's a flawed argument that we've already covered.
[[[While it may be nice to wish and dream that Dassault would provide a Linux Based (or OS X, or generic UNIX, or any other OS out there) CAD/CAM/CAE System, that will not happen until some time in the future, and most of us here are dealing with the present. ]]]
As the competition passes you by smiling and waving because you decided it was a "safe bet" to rest on your laurels. What about the future? Long term? Growth? Diversity? How about just being *forward looking*?
[[[CATIA V5 on Windows is the most functional and powerful Fully Integrated CAD/CAM/CAE System in existance today.]]]
And you must be aware of how quickly that could change, but still, having it on Windows doesn't say much. Why not diversify? Why place all your eggs in a single basket? Why sell your soul to an opportunistic company like Microsoft? I'm betting you'll come around sooner than you might think.
Still you have failed to address IBM's plans for the 970. IBM has already stated it's intentions. They will be releasing high-end desktop workstations and low to midrange servers based on LINUX-on-PowerPC-970. Do you mean to sit there and have me believe that if IBM wanted CATIA to be running on their new platform offering you would sit there and preach to them how great 32bit CATIA is on Pentium-based Windows systems?!? No, I think not. I'm willing to bet that IBM holds a LOT more clout than that and if IBM wanted CATIA running on LINUX-PPC970 it'd be offered and there would probably be very little argument. And suddenly CATIA on Windows wouldn't look that great. Let me ask you this... Do you think that offering CATIA *only* for Windows was a smart/safe move?
--
Ed
Author: Ed M. |
|
|
|
|
COE-FORUM-USER
|
| 15 May 2003 11:40 AM |
|
When it takes 3 to 5 years to implement a new Graphics System, I'm not concerned with the Johnny-Come-Lately systems. If a truly innovative and competitive system comes along, it will still be there the next time we start looking. In an industry that takes 3-5 years or more to develop a new product, we do not change CAD systems lightly. The amount of training alone is a significant enough investment that management will not even consider more frequent changes.
Nor am I concerned with my competitors beating me to the punch. They have the same budgetary, training, and legacy issues that we do. And we are not "resting on our laurels". We are in the the process of switching from CATIA V4 to V5. This implementation started last year, and will probably last another 2 to 3 years before we are fully migrated. If DS were to re-write CATIA and make it fully Linux, OS-X, or Unix operable, and if it were as significantly different as V5 is from V4, we would either have to start over, or wait a few years (and waiting would be good anyway, as the first few releases of any CAD system are pretty unusable).
Having run CATIA for many years, and many releases, on both Unix Workstations (AIX) and Windows, I can tell you that the vast majority of issues are with the application, not the OS. Most of these end up being functions that do not work the way they were supposed to. The only OS dependant issues that seem to come up are memory leaks, and those are not always OS dependant - sometimes the identical errors show up on all platforms.
Since our CAD system is integrated with a VPDM system for data storage and management, and that database resides on a Unix Server, data security is less of an issue than most Windows Only installations. Even our File-Based processes reside on Unix Servers. We do not rely on Windows for our security.
As for the 32 bit arguement, CATIA is still a 32 bit application, and until they re-write it and make it a 64 bit application, it really doesn't matter whether my OS is 32 or 64 bits (yes, I know that there is supposed to be some improvement even with 32 bit apps, but I do not believe that it is significant). I suspect that, when CATIA does become 64 bit, it will only be certain functions within CATIA (at first anyway), and those types of applications are already running on Unix Workstations and Servers.
As for IBM's new iron, I'm sure that, if IBM wants CATIA to run on these, they will provide AIX (probably AIX 5), and V5 will run on these workstations. But, just like V5 on AIX today, it will be running as an emulation, with a fair amount of missing functionality.
PS we are currently running our V5 users on a mix of AIX and Windows 2000, with some IRIX thrown in.
Jim Strawn
Cessna Aircraft Co.
316-517-5851
Author: Jim Strawn |
|
|
|
|
|
| You are not authorized to post a reply. |
|
|
|
ActiveForums 3.6
|
|
|
|