I while ago, I bought this Optiplex GX260. Not a bad, general computing machine from a generation ago. But, for Pro/e, it was encumbered with an onboard graphics processor that used system memory for its GPU; adequate but slow and memory restricted (with 2, 3 windows open, it started to bog down). So, I looked around for a good nVidia card that would softquadro to a 3000. Found it, installed it and now, I'm puzzled. What's to softquadro! It's the most flamingly great video I've ever seen for Pro/e: hundreds of files in an assembly spin like a part; open 10, 15, 20 parts, modify and regen these parts and no slow down; nothing I try can slow this thing down and it's FASTER than the dual quad doofus I use at work (both Dells, no heavy networking, no Intralink, on mine.) OTHERWISE, mine ought to be inferior but seems to perform better!!! Must be the overhead on the professional system. Anyone else notice such discrepancies in their setups? We get inexplicably slow here, how about inexplicably fast!?! BTW, my card was based on the geforce 6800 gs gpu, can't remember who manufactured the card, but I got a good price.
The Quadro FX 3000 is basically a Geforce FX 5900, so you need to find a gfx card with this GPU on it. Since both the Quadro FX 3000 and the Geforce FX 5900 are old and outdated you will have a hard time finding any of them.
I don't know what gfx cards you use at work but your experience has probably nothing to do with Softquadro. Of course an Optiplex (which is Dells line of office computers) has an abysmal performance on Pro/E with the integrated GMA shared memory gfx. For Pro/E even the slowest gfx card you can buy today would have provided you with *much* better performance. I did run Pro/E for a while on a Geforce 6600 non-GT which was a low end card and even without softquadro (which btw. only tricks the driver to think the gfx card is a Quadro and thus enabling some Quadro-only features like AA lines) it was fast like hell. No comparison to running that on an integrated chipset gfx with shared memory.
What "overhead"? It's probably more likely that you simply have a faster gfx card than the machines you have at work.
You didn't notice any discrepancies but just the logical increase in performance due to the addition of faster hardware.
Never ever had that case, there always was a reason why a system was inexplicably slow or fast. Sometimes you just have to dig deep enough to find the reason.
A 6800GS would probably equal to a Quadro FX 1300. Of course it's faster than the old Quadro FX 3000/Geforce FX 5900.
I have no idea what softquadro is, but while we're off topic..
I recently upgraded from a four year old duel xeon 2.4 with 1 gig rdram and a quadro4 750 xgl to a duel xeon 3.73 with 8 gig ram and a quadro fx 4500. Fairly high end 4 year old, to fairly high end current.
The performance increase doesn't exactly blow me away. I figured four years of improvements would yield a huge increase. It is about twice as fast, but what happened to Moore's law? It should be about 5 times faster, no? Feh...
My current WF system doesn't feel as fast as the old system did when it was on 2001. The software bloat grew faster than the hardware.
Even with a relatively powerful system, very large part files become cumbersome with the new object/action way of doing things. Selecting a single surface on a 100 mb part can be kinda slow.
Whenever I hear someone say "computers are more powerful than anybody needs, who needs more power", I just want to shake them.
Selecting a single surface on a 100 mb part can be kinda slow.
Maybe something like a part simp rep with only the surfaces you need for particular feature functions? Or maybe just turn off 'preselect', which, by virtue of having to calculate EVERYTHING below the cursor before it displays ANYTHING, slows the system WAY DOWN (especially with big, complicated, multilayered surface models with many patches AND quilts and multiple merges AND accuracy problems.)
Their necks are ALREADY snapped, thus the brain-deadedness. Couldn't program robots to be stupider. It is uniquely human; any other species, so burdened, would have died out a long time ago.
Honestly, if you really upgraded recently to a dual XEON 3.73 (XEON
5080) then you probably just wasted your money. The XEON 5080 was outdated already a year ago and is nothing more than a frying pan, even intel didn't recommend anyone to use them. The XEON 5000 series are based on the old, very inefficient and very hot running Netburst architecture. You should have choosen the newer and much better XEON
5100 (dual core) or XEON 5300 (quad core) which are based on the Core2 architecture, run much cooler and despite the lower clock rate are much faster than your XEON 5080. A cheap XEON 5130 with only 2GHz will run circles around your 5080.
The Quadro FX 4500 is a good card even when being already two generations old. But it still requires a system infrastructure fast enough to feed it properly.
On Jun 26, 4:25 am, Benjamin Gawert (WET BLANKET) wrote: blah blah blah
The XEON 5080 was (blah blah blah)
You should have choosen (blah blah blah)
Damn Ben, it's just a computer man. Got it about six months ago I guess. I said fairly high end, not bleeding edge. My point being it was a similar level of "state of the art" compared to the computer it replaced, but four years younger, yet it wasn't as great an improvement as four years of computer evolution used to be worth. My other point being that application bloat has exceeded the improvements in computational power these last few years, making Pro/E (the topic) feel slower than it once did.
"You should have choosen the newer and much better XEON
Dude, do you go home at night and touch your computer in inappropriate ways?
The Quadro FX 3000 is basically a Geforce FX 5900, so you need to find a gfx card with this GPU on it. Since both the Quadro FX 3000 and the Geforce FX 5900 are old and outdated you will have a hard time finding any of them.
I don't know what gfx cards you use at work but your experience has probably nothing to do with Softquadro. Of course an Optiplex (which is Dells line of office computers) has an abysmal performance on Pro/E with the integrated GMA shared memory gfx. For Pro/E even the slowest gfx card you can buy today would have provided you with *much* better performance. I did run Pro/E for a while on a Geforce 6600 non-GT which was a low end card and even without softquadro (which btw. only tricks the driver to think the gfx card is a Quadro and thus enabling some Quadro-only features like AA lines) it was fast like hell. No comparison to running that on an integrated chipset gfx with shared memory.
Yes, I acknowledge and maybe even understand the perceptual issue you bring up: comparatively speaking, it just SEEMS a whole lot faster. And, as engineers, we really need, love, eat up the numbers (such as those available from Corten's proesite.com) which, these days, is actually managing to compare apples for apples which, in the past, I criticized it for not doing, for presenting a menagerie of systems, setups and software versions. And then saying, here, you figure it out.
However, I'm not so entrance by the issue of speed (which I more or less expected) as by the issue of the number of models open. Before, with whatever non-quadro setup, I've been told I'm naturally limited to 4-5 windows open and then the system bogs down (OpenGL rules apply) and you need to Softquadro to get around this BIG limitation. Seems now, since I've opened a dozen files at a time, that this is OGL mythology. I suffer no performance degredation when opening between 10-20 files with this slick gaming card (512 M DDR3 memory, AGP 8 [of which I can use only AGP 4] and [of no use to Pro/e] DirectX 9.) Is the memory that big a factor?
David Janes
What "overhead"? It's probably more likely that you simply have a faster gfx card than the machines you have at work.
You didn't notice any discrepancies but just the logical increase in performance due to the addition of faster hardware.
Never ever had that case, there always was a reason why a system was inexplicably slow or fast. Sometimes you just have to dig deep enough to find the reason.
A 6800GS would probably equal to a Quadro FX 1300. Of course it's faster than the old Quadro FX 3000/Geforce FX 5900.
nothing more than a frying pan, even intel didn't recommend anyone to use them.
Freakin' amazing what you can do with a goddamn fryin' pan, man! And, just imagine, they engineered, designed and built the Golden Gate Bridge (and every freakin' thing before about 1970) without a computer, yeah, with lots of paper and pencils.
Fine. I just wanted to point out that the XEON 5000 was never ever a good buy. Even intel and most workstation manufacturers recommend their customers not to go for the XEON 5000 but for the XEON 5100 which was available at the same time, faster, cooler, and besides all that, even cheaper! I understand that this doesn't sound good but it's a fact.
Which is clear since you choosed an outdated and inefficient technology when buying this system. Of course the resulting performance increase is not as if you would have compared your four year old system to (at the time you bought the new system) current (not highend!) technology. The Core2-based XEONs are out for almost a year now, the XEON 5000 predates the 5100 by about a month or so.
It doesn't.
Nope. But I usually get informed about what I buy before I buy instead of buying old technology and then complain about the lack of improvement.
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.