Hello,

I’m running Max 9 in Win XP Pro 64 bit with Nvidia Quadro 4500.  Is it worth running MAXTreme drivers? There isn’t even an updated user guide on Nvidia’s site.  Is it being phased out?  What are the advantages? Is it stable?

TIA

Dean

Max Design 2012 SP2 / Win 7 Pro
NVidia Quadro FX 4800 x 2 (non sli)
HP Z800 - 2 Xeon X5680 @ 3.33 GHz; 36 GB ram

Sorry, I only had disadvantages.  Never really worked for me, too slow as well as inverted faces.  Maybe that’s why its phased out . . . didn’t work.

Calvin Clayton, Jr.

Max 9 is optimized for DirectX display… it’s generally the fastest for Max9.  OpenGL or Maxtreme or whatever may be faster in some cases, under certain circumstances, but DirectX worked best for me as advertised.  Try that.

Maneswar Cheemalapati [FA]

Maneswar Cheemalapati [FA]

Yeah, Direct X is what I’m currently using. Just thought I would ask to see if I’m missing anything.

Max Design 2012 SP2 / Win 7 Pro
NVidia Quadro FX 4800 x 2 (non sli)
HP Z800 - 2 Xeon X5680 @ 3.33 GHz; 36 GB ram

I’m going to have to disagree for what we do. I work on appliance models all day. Sometimes many at once. I am able to shade engineering models (although the shading process can take a while), several at a time, and still navigate around the scene without using any sort of adaptive degradation.  that’s why I always laugh when Autodesk shows us those cool new adaptive degradation features. I don’t have any problems spinning around my crazy high poly scenes anyway.
I was going to give you an idea of how high they are, most are several million, but the file I have crashes every time I try and bring up a polycount. To give you a good idea though, the last big model we did was a dishwasher. The file was about 250mb and took about 2.4gigs of memory just to start rendering. I’m not sure how high it got, but that was the initial size when it was sent.
That model had no problem rotating around once it finally shaded the first time and I was getting a frame rate of at least 15fps, but that’s just a guess.

Anyway, we’ve had some great results with the maxtreme drivers. One person here doesn’t use them and has to constantly turn on adaptive deg. The only time I remember having huge problems was when I had a model originating from CAD data, as 99% of our data does, was when we had a model where none of the pieces were joined. Basically every small piece, almost down to the size of each polygon, was a separate object. There were over 100,000 objects in the scene. I could barely do anything. No part of Max is built for that though.

If you don’t have at QuadroFX card then I probably wouldn’t do it with just an NVS card or something, but the maxtreme drivers offer us quite a bit more speed.

Try it, its free, if you don’t like it you can always switch it back, simple as that.

I have used it on and off for a few years, it has gotten better, well for somethings it works great and others not so great. I have heard both sides of the story from quite a few folks and they either use or don’t use it.

I have had some issues when modeling, no issues when working with particles, don’t really do any serious texturing so I can’t comment there. Viewport speeds seem to be ok when animating.*shrug* Give them a try…

particle mechanic

Maxtreme has some bugs though that bothered me quite a bit… and I think similar bugs bothered me with my FireGL Maximum drivers as well… vertices would show up in the oddest of places when working with Poly’s.  I eventually worked around it with driver / display settings, but it frustrated me that cards for professionals still had problems that obvious… it was impossible to miss if you turned on certain display settings.  It’s been a while so I can’t recall the specifics or I’d state them.

DirectX is not nirvana, but it’s been pretty reliable for me. :) But sure, try all the driver options, they’re there to be used for various different tasks.

Maneswar Cheemalapati [FA]

Maxtreme use to be based off almost all OpenGL, but since they switch to be based more off the D3D drivers the bugs are much less.
I do remember the vertices problem though. It was Really bad. At the time I didn’t have to do much modeling. I switched drivers when I had to poly model. The vertices were still selectable in the correct location. There was just a viewport error. That was really strange.
They seem alright now though. Who knows what will happen with 2008 though. Of course, 2008 really doesn’t seem different.

has anyone tried maxtreme in max2008? i’ve noticed my objects shift slightly when moving them in viewport.

I can confirm the shifting experience in Max2008 with Maxtreme, this is very annoying. Also I find that I might have to create an object twice to actually see it in the viewport, annoying again.

Back to DX3 for me!

PS: yes Maxtreme had a great little bug that when you went into wireframe mode the verts shrunk off the wireframe, this made it hard to select the correct verts without going back to shaded mode.

I tried oGL, DX and maxtreeme on quadroFX 4000. Ogl was clearly the worst (performer) of the bunch. It didn’t take much to slow it to a crawl.

Maxtreeme and dx on the other hand… not so easy to say who’s a clear winner; so that fact alone makes dx a clear winner.

@ Andy Engelkemier, I had similar file of a huge 17th century fortress where each window had separate piece and threes had almost all leaves as separate objects- it was a crazy huge detailed scene. I first opened it at slower computer that had 9600pro ati card. That was a slide show of 1 frame per 10seconds, it was silly, really.

Nice test for a quadro, i thought. And interesting thing is that both dx and maxtreeme gave very smooth frame rates on this scne. Like I mentioned earlier- almost identical.

The fact that pi**ed me off and made me ignore quadro/firegl cards (for as long as i live, unless they change) is that with nvida 6800gt (which coincidently has same architecture as the quadro 4000) got exactly the same smooth framerates. In fact, no scene I tied was any slower on it. I could run millions of polies by it and it would shade them happily with 16x AA with no stutter. If i remember correctly, quadro could do it up to 24x AA, but honestly i couldn’t tell the difference. ~2k euros compared to ~300 euros. Worst investment ever. From now on whenever someone asks me- i tell them that i dislike quadro/firegl line at great intensity.

Now it’s even worse, with the introduction of unified shader architecture of g80, r600 (nVidia and ATi respectively) quadro/firegl lost even more value (not that it was realy high in my mind to begin with). Their ability to do 32x AA is nice on paper, but in RL I challenge anyone to tell me the difference to 16 AA in under 5sec. And that limitation on GF series is software imposed, not architecture.

Lucky for us users there are such programs such as nHancer that allow 4x4 supersampling of the image, and the new g80 and g92 cores are more than capable of carrying this out. 4x4 supersampled image is somewhat better than 32 AA (that can also be selected from within nHancer).

So i ask, is a five x $ ok for a product with somewhat higher on board memory? (note that you can get both g92(nv 8800gt/s) and rv670(ati 38x0) with up to 1.5gb of ram for very little additional cost over the mainstream product)

EoDEo

Ideas drift like petals on the wind. You have only to lift your face to the breeze.