workstation quadro p4000 vs gtx 1080 ti

[silent_sight]

So, I'm looking into which graphics card would end up performing better for my workstation to run Maya for 3d animation work... this computer is not used for gaming.  I model, rig, and animate (mostly rigging).  I'm mainly trying to determine what will allow my maya viewport to run the smoothest, but sifting through all the data online is very difficult since most (if not all) tests are run with gaming fps in mind.  I've seen some data comparing the quadros to each other, but not against any gtx cards.  Maya officially supports quadros (predominately), but I know gtx cards can work too.  I saw that the titan X is also supported, but I can't seem to find any info about where it ranks among the quadros, and I don't know if the 1080 ti will be better for maya in the same way that it's more powerful for gaming.

 

I found this on youtube

At first it looks pretty amazing but it seems a little deceptive because the quadro card is running NVidia architecture that is a version or two older than the 1080 (specifically keplar, vs maxwell).  I've also watched every linus tech tips video where quadro cards are mentioned, but they mainly just seem to be telling gamers not to bother buying them or just explaining what the difference is.  There aren't really any fair comparisons for speed using maya anywhere - at least, non that I could find.

 

I am specifically looking at the new quadro p4000 and the new gtx 1080 ti.  I may consider the p5000, but I think it's a little bit too expensive for the minimal performance gain over the p4000 (based on this NVidia info page http://www.nvidia.com/object/autodesk-maya.html). These cards are all on the pascal architecture, so I'm hoping there is someone out there who may be able to crack open maya and check the fps (tumbling the camera around a model, checking wireframe, shaded, textured, playing animation in the maya window, maybe running some dynamics and/or particles, etc) to compare the two cards (p4000 and 1080 ti).

 

I would greatly appreciate actual data to people simply saying to get one or the other - I've read posts on several forums of people simply stating what to get without much of anything to support it, to the point that it's actually not really helpful... I know linus tech tips mostly makes videos with gaming in mind, but I'm kind of hoping they may choose to do a video with data like I'm requesting... not sure the best place to ask them for that, but I'm also content to view feedback here even if it's not a youtube video.

 

I'm really only concerned about performance in the maya viewport - not gpu rendering... It sounds like I can always install a few gtx cards for the sake of gpu rendering (if I ever even need to do that...)

 

Thanks for taking the time to read this and respond!

[themctipers]

1080ti (quadro has less cuda cores, mhz, vram)

[LienusLateTips]

Quadro! Higher binned cards are better for workstation loads because of their reliabilty.

[Zodiark1593]

Unless you are into AUTOCAD or some extensive compute where a lot of money rides on a correct result, (or your scenes demand obscene quantities of VRAM) go for the 1080 TI. 

[matrix07012]

Get the 1080ti, unless you need ECC VRAM.

[Terryv]

2 minutes ago, Zodiark1593 said:

Unless you are into AUTOCAD or some extensive compute where a lot of money rides on a correct result, (or your scenes demand obscene quantities of VRAM) go for the 1080 TI. 

Pretty much this. 

[Coaxialgamer]

I'd go for the 1080ti

[silent_sight]

Thank you for the responses.  I don't want to come across as ungrateful, but I'm hoping for something a little more in depth with some sort of data behind it...

 

themctipers - yes, all of that is true, but I'm wondering if the drivers would make enough of a difference for performance in maya

 

JDE - reliability is a concern, but I've not come across anything in my research showing that gtx cards are not reliable... Could you perhaps expand and clarify?  Are you talking about the longevity of the card, gpu rendering, or maya viewport performance?

 

Zodiark1593 - while I won't be using auto cad, faster performance within maya will allow me to work faster/efficiently.  Between the quadro p4000 and the gtx 1080 ti, the quadro actually has less vram (and actually half the cuda cores); however, I'd like to know if the different drivers offset this disparity.

 

matrix07012 - based on this page http://www.nvidia.com/object/compare-quadro-gpus.html the p4000 doesn't have ECC (although it's not exactly clear if they're talking about the vram or not...)

 

It seems like there's just too many variables to merely look at the card specs and determine which is going to be better in actual real world use.  The difference between the quadro and gtx drivers, the way gtxs redraw the whole image for each screen while quadros redraw smaller parts, whether or not the larger number of cuda cores and vram actually allows the 1080 ti to work faster in maya than the p4000, etc, etc, etc - these are the things that really need some sort of benchmarking to help show which one would be better to go with - it may be surprising to see one truly outperform the other, but hard data is really the only way to be sure.  The quadro p4000 is about $160 more than the gtx 1080 ti; if it performs better in maya, then it's very much worth it to me, but if it's actually weaker for my needs, then it's a complete waste of that money.

 

Is there a place on this forum for linus tech tip video suggestions, or do they read through the forums too?  Of course... since most of what they do is in reference to gaming, they may not deem this to be a cost effective test for them to do... you never know without asking though, right?

[Qwweb]

19 hours ago, matrix07012 said:

Get the 1080ti, unless you need ECC VRAM.

Or double precision FLOPs or 32 bit color depth.

10 minutes ago, silent_sight said:

The quadro p4000 is about $160 more than the gtx 1080 ti; if it performs better in maya

If reliability will effect the overall render times I would go for the quadro as it will likely last longer than the 1080 ti.

[Princess Luna]

second hand Titan X Maxwell, it'll beat the 1080ti for this task while being cheaper, I think you can get one for 400 dollars, and if it's EVGA superclocked you can even overclock it to be as good as a 1080 stock in gaming.

[Princess Luna]

20 hours ago, silent_sight said:

I saw that the titan X is also supported, but I can't seem to find any info about where it ranks among the quadros,

 

[silent_sight]

Qwweb - I'm actually not going to be using the card for rendering... I'm just concerned about the speed in which the card will allow interaction within maya - specifically in regards to modeling, rigging, and animating.  I've seen several youtube videos of people creating gpu render machines by utilizing many gtx cards (not in sli), so I would probably head that route if I need rendering capacity.

 

Princess Cadence - I could be wrong, but I highly doubt a graphics card with Maxwell architecture would outperform a card with Pascal architecture (at least between the titan X and 1080 ti) especially since the 1080 ti has been shown to be more powerful than the titan X with pascal architecture.  Also, since I'm making a workstation, I'll not be doing any overclocking to ensure stability and longevity of the machine.  As for your second post - I apologize, perhaps I wasn't clear enough in my post... I meant that I had not seen any information that showed where the titan X ranked among quadro cards for use within Maya.  There are some charts showing how the quadro cards perform against each other, but I have no idea where the titan X fits in among those (at the bottom? somewhere in the middle? no idea...).  I have watched that video before, but went ahead and watched it again to be sure.  Linus basically tests the M6000 against the titan X for gaming.  Afterward, he does test them on some type of rendering software and also a video editing program.  I'm not rendering, so I don't care about the rendering performance of the cards; I need to know about maya interactivity, and he doesn't take a look at that - aside from that, it's not even the cards I'm looking into, so it wouldn't be exceptionally helpful unless I find something showing the performance of the p4000 against the m6000 in maya.

 

Just to help clarify further - in case some people aren't aware - rendering and maya interactivity are two completely separate things that have different requirements.  Interactivity within maya is all about how quickly you can spin a camera around objects/characters, manipulate geometry, manipulate a rig and set keyframes, etc - all while working interactively with maya in one of the viewports.  Rendering is the process of determining the color required for each pixel of the final image after geometry, lighting, and texturing are taken into account.  Rendering requires no interactivity - you start a render and pretty much walk away while your computer crunches numbers...

[scottyseng]

10 minutes ago, silent_sight said:

-snip-

Well, in certain cases, some older architectures were better because Nvidia didn't cripple the double precision performance (They have to sell the Tesla cards somehow).

Also, for my Maya / 3DS Max use, the 980 Ti beats out my 1070 (Yes, both are overclocked). I originally had a FirePro V7900 but I upgraded because it simply could not keep up anymore. The view ports are faster on the 980 Ti than the 1070 (While the 980 Ti sits at a lower GPU usage).

 

That being said, the Quadros / FirePros do have optimizations to handle large amount of geometry better than a Geforce card. However, the brute force speed of the gaming equivalents usually beats out this gap. The main advantage of the workstation GPUs aside from 24/7 reliability is ECC (Though only the higher end cards have it), the higher amount of VRAM (Though the newer Nvidia cards kind of win out on this), or the need to stick these in servers (These have rear power connectors so they fit in tighter spaces). Also 10-bit color output (Though some high end Nvidia cards / AMD have 10-bit output).

 

Tom's Hardware had great reviews on Workstation GPUs, but they haven't reviewed any since then (Well, it's hard to get these workstation GPUs to be honest).

http://www.tomshardware.com/reviews/firepro-w9100-performance,3810.html

http://www.tomshardware.com/reviews/best-workstation-graphics-card,3493.html

 

The reason you don't see many workstation GPU tests are because the major tech youtubers don't have the skills to properly review these cards (aka, you get stuff like Linus saying they aren't great for gaming / one use case scenario (video editing), and this results in Nvidia / AMD being weary to send out these cards for reviews. There are some programs that really shine with the Quadros / FirePros (Well, now Radeon Pro).

 

In my opinion though, the brute force power of the 1080 Ti should give better interactivity (At least versus the P4000). Though I wonder how 1070s in SLI would compare. Also, note that some Autodesk Software is heavily CPU bound for some stuff (Revit / Inventor use primarily CPU for geometry tasks). I wish I had a quadro to compare myself, but they cost so much...I haven't had any issues with viewport performance on my 980 Ti / 1070 though.

[silent_sight]

6 hours ago, scottyseng said:

Well, in certain cases, some older architectures were better because Nvidia didn't cripple the double precision performance (They have to sell the Tesla cards somehow).

Also, for my Maya / 3DS Max use, the 980 Ti beats out my 1070 (Yes, both are overclocked). I originally had a FirePro V7900 but I upgraded because it simply could not keep up anymore. The view ports are faster on the 980 Ti than the 1070 (While the 980 Ti sits at a lower GPU usage).

 

That being said, the Quadros / FirePros do have optimizations to handle large amount of geometry better than a Geforce card. However, the brute force speed of the gaming equivalents usually beats out this gap. The main advantage of the workstation GPUs aside from 24/7 reliability is ECC (Though only the higher end cards have it), the higher amount of VRAM (Though the newer Nvidia cards kind of win out on this), or the need to stick these in servers (These have rear power connectors so they fit in tighter spaces). Also 10-bit color output (Though some high end Nvidia cards / AMD have 10-bit output).

 

Tom's Hardware had great reviews on Workstation GPUs, but they haven't reviewed any since then (Well, it's hard to get these workstation GPUs to be honest).

http://www.tomshardware.com/reviews/firepro-w9100-performance,3810.html

http://www.tomshardware.com/reviews/best-workstation-graphics-card,3493.html

 

The reason you don't see many workstation GPU tests are because the major tech youtubers don't have the skills to properly review these cards (aka, you get stuff like Linus saying they aren't great for gaming / one use case scenario (video editing), and this results in Nvidia / AMD being weary to send out these cards for reviews. There are some programs that really shine with the Quadros / FirePros (Well, now Radeon Pro).

 

In my opinion though, the brute force power of the 1080 Ti should give better interactivity (At least versus the P4000). Though I wonder how 1070s in SLI would compare. Also, note that some Autodesk Software is heavily CPU bound for some stuff (Revit / Inventor use primarily CPU for geometry tasks). I wish I had a quadro to compare myself, but they cost so much...I haven't had any issues with viewport performance on my 980 Ti / 1070 though.

thank you so much for taking the time to reply with so much information!

 

Do you think the 980 ti is beating out your 1070 due to the "ti" distinction?  From what I gather, NVidia slapping the ti on a card means it gets a pretty significant power boost over it's non-ti counterpart - although I haven't done enough research to see if they're all upgraded in the same method...

 

the Tom's Hardware articles were pretty interesting to read through.  It was a little unclear exactly what they were testing in the maya section (were they playing through an animation interactively and timing it?  There was a video on the page, but it was just some guy doing a review of a laptop and not a description of the test).  The quadro cards seemed to outperform the gtx cards in maya - for the most part... I don't know much about the difference between quadro and gtx cards at that time though - it would be interesting to see the results if they were to do the tests again with current cards.  I have been curious if the "brute force speed" would even out or even surpass the lower tier quadro cards even with the difference in drivers - especially since it sounds like the gap between quadro/gtx specs is less significant these days.  I wish you had a quadro to compare to as well, haha!

 

As to your question about how the 1070 in sli would compare, I have read in multiple places (forums and articles online) that maya does not make use of sli at all... So, I assume it would perform exactly the same as the 1070 alone.  At least for now, sli may just be a gaming thing (I've even read that some games don't support sli - I guess it has to be factored in to the programing of the software)...

 

Thanks again for the input!

[scottyseng]

8 minutes ago, silent_sight said:

-snip-

Yeah, the 980 Ti was the flagship of Maxwell cards. It also has a CUDA core advantage over the 1070. However, you shouldn't consider the TI to be a significant power boost. There is a boost for sure, but you can't say all of the Ti cards were that great. The 780 vs 780 Ti wasn't too big. However, the 980 Ti vs 980 and 1080 vs 1080 Ti were pretty big (They pretty much matched or out performed the equivalent Titan card for that gen).

 

Ah, the Maya Catia benchmark pretty much loads a high poly count object and rotates it in the scene. The more fps, the faster the rotation can be finished. Catia times this, and so the faster it finishes, the smoother the viewport is. But yeah, I agree, it's a shame that Tom's hardware haven't tested the Maxwell / Pascal Quadros (And AMD equals) against their gaming card equals. I guess it's because of how hard it is to get the workstation cards (If you think about it, that's about $10K+ of workstation GPUs that had in that lineup alone) when those articles aired. Yeah, I remember the FirePro having a optimized tessellation engine (at the time, it was wicked fast at 3D modeling, my old FirePro V7900). Same for the quadros. It is really is hard to tell if modern GPUs can outpace the Quadro version for 3D model work (Since both sides have grown a lot in power since those Tom's Hardware articles...heck, the Quadro P6000 is actually faster than the Titan XP (More CUDA cores)).

 

Yeah, I was on the fence of getting a Quadro P4000 myself, but I render as well, so I kind of had to go with a used 980 Ti ($350) and 1070 ($330) for my needs. I do plan to get another used 1070 when people start selling them off for 1080 Tis.

 

I think you're right on that, if the viewports used Direct X, then maybe it would...Well, that's informative.

 

Yeah, worst comes to worst, you can do the evil thing, buy a P4000, test it, and then return it. And buy a 1080 Ti and return it...and see which ones fares better...though that's a lot of work.

 

Oh, I'm only a hobbyist though. I'm still learning 3D modeling.

[Zodiark1593]

13 hours ago, scottyseng said:

Yeah, the 980 Ti was the flagship of Maxwell cards. It also has a CUDA core advantage over the 1070. However, you shouldn't consider the TI to be a significant power boost. There is a boost for sure, but you can't say all of the Ti cards were that great. The 780 vs 780 Ti wasn't too big. However, the 980 Ti vs 980 and 1080 vs 1080 Ti were pretty big (They pretty much matched or out performed the equivalent Titan card for that gen).

 

Ah, the Maya Catia benchmark pretty much loads a high poly count object and rotates it in the scene. The more fps, the faster the rotation can be finished. Catia times this, and so the faster it finishes, the smoother the viewport is. But yeah, I agree, it's a shame that Tom's hardware haven't tested the Maxwell / Pascal Quadros (And AMD equals) against their gaming card equals. I guess it's because of how hard it is to get the workstation cards (If you think about it, that's about $10K+ of workstation GPUs that had in that lineup alone) when those articles aired. Yeah, I remember the FirePro having a optimized tessellation engine (at the time, it was wicked fast at 3D modeling, my old FirePro V7900). Same for the quadros. It is really is hard to tell if modern GPUs can outpace the Quadro version for 3D model work (Since both sides have grown a lot in power since those Tom's Hardware articles...heck, the Quadro P6000 is actually faster than the Titan XP (More CUDA cores)).

 

Yeah, I was on the fence of getting a Quadro P4000 myself, but I render as well, so I kind of had to go with a used 980 Ti ($350) and 1070 ($330) for my needs. I do plan to get another used 1070 when people start selling them off for 1080 Tis.

 

I think you're right on that, if the viewports used Direct X, then maybe it would...Well, that's informative.

 

Yeah, worst comes to worst, you can do the evil thing, buy a P4000, test it, and then return it. And buy a 1080 Ti and return it...and see which ones fares better...though that's a lot of work.

 

Oh, I'm only a hobbyist though. I'm still learning 3D modeling.

From what I've read of Maxwell, it had received a very substantial improvement in polygon and tesselation performance over Kepler. This was important in the gaming lineup due to things like Hairworks being utilized in some games. Naturally, the Pascal lineup is going to have no problems with polygon count. 

 

From what little I've read of workstation cards, OpenGL performance tends to be better than the gaming equivalent, however, due to the vast difference in capability between the P4000 and the 1080 TI, this point is completely moot. 

[O_and_N]

On 25/3/2017 at 4:16 AM, silent_sight said:

thank you so much for taking the time to reply with so much information!

 

Do you think the 980 ti is beating out your 1070 due to the "ti" distinction?  From what I gather, NVidia slapping the ti on a card means it gets a pretty significant power boost over it's non-ti counterpart - although I haven't done enough research to see if they're all upgraded in the same method...

 

the Tom's Hardware articles were pretty interesting to read through.  It was a little unclear exactly what they were testing in the maya section (were they playing through an animation interactively and timing it?  There was a video on the page, but it was just some guy doing a review of a laptop and not a description of the test).  The quadro cards seemed to outperform the gtx cards in maya - for the most part... I don't know much about the difference between quadro and gtx cards at that time though - it would be interesting to see the results if they were to do the tests again with current cards.  I have been curious if the "brute force speed" would even out or even surpass the lower tier quadro cards even with the difference in drivers - especially since it sounds like the gap between quadro/gtx specs is less significant these days.  I wish you had a quadro to compare to as well, haha!

 

As to your question about how the 1070 in sli would compare, I have read in multiple places (forums and articles online) that maya does not make use of sli at all... So, I assume it would perform exactly the same as the 1070 alone.  At least for now, sli may just be a gaming thing (I've even read that some games don't support sli - I guess it has to be factored in to the programing of the software)...

 

Thanks again for the input!

Was browsing the internet in search of more info in relation to the same question when i saw this thread and i have some info to share.(Its funny to see another person looking at the same gpus p4000 vs 1080ti and maya at the same time as I am lol)

(sli doesnt work in maya but if you have more than one gpu with no sli you can do gpu rendering with a redshift/octane plug in)

 

 

So, i made a thread over at the autodesk forums with a heavy maya scene and a few people tested their gpus with it(including my k4200 and a k5000 from the user of the video you posted)to compare results.

 

https://forums.autodesk.com/t5/maya-forum/are-quadros-and-firepros-still-important/td-p/6772440

 

I will be buying 2 gpus(2xp4000 or 2x1080ti) in a week and i will most probably make the final decision in the last minute.From what i see,this problem here is because in the past maya(and max) worked bad with gaming gpus.Long time users remember the pain and are nervious to switch as they dont want to get a nasty surprise.On the other hand,we are tired on wasting money...

 

Looking at the tests in my link i can see that gaming gpus ¨¨MOST¨¨ probably will work just fine in maya BUT using only the viewport 2.0.I will confess that i still use the old viewport,but i think its time to change as its called legacy for a reason now.

On the other hand using vieport 2.0 can be used in dx11 mode and open gl.The dx11 will probably be the best for gtx gpus but i have the feelong that the open gl coud not perform that good on a gtx.Things like paint effects dont work in vp 2.0 so legacy one is needed there.

 

Quadro

-less power

-probably beter silicone used and will survive high/long operating temp longer

-eliminates the moment of a nasty surprise in case you do something new in the workflow?

 

gtx

-more brute force

-less cost(its all about that)

-the feeling you are not being riped off?

 

Many things point me that the gtx woud be fine,its just that feeling in  my stomach that still warns me that with time i might regret it.(thanks to years of experience and as i always spend a lot on one main workstation and dont use another pc)

 

The p4000 looks good.No reviews yet,except some small words here and there showing that in 3d mark orange room is superior than a m5000 for vr.On paper it is a gtx 1070 but cripled.Still i suspect that in raw power it shoud be no lees than a stock gtx 980(game performance).As I do use the unreal engine and maya together.And yes,quadros play games(develop)just fine lol!

 

If you buy/both a p4000 or a gtx 1080 before me,please share the info dude.

Cheers

[silent_sight]

O_and_N - wow, thanks for all the info and the link to the thread in the autodesk forum!  I'm sure there are more people than us looking for comparisons between those 2 cards.  That was a lot of stuff to chew through... It looks like the gtx cards performed really well in viewport 2.0, and even some old quadro cards did pretty well there too.  It was pretty obvious that the legacy viewport kinda needs a quadro; but even then, VP2.0 was way faster anyway.

 

I get the impression that if I was to do a lot of gpu rendering (with redshift/octane), then a few 1080 ti cards would be best since it would provide way more cuda cores for the rendering.  It also seems like it wouldn't matter much which card I go with if using VP2.0 (unless I'm just not really reading everyone's feedback correctly on the thread you linked to); legacy viewport is where gtx cards are notably weaker than quadro cards.

 

Due to some of life's changes I may be moving soon, so I was holding off on purchasing/shipping anything until that is determined (don't want to ship something to a location I may move away from, heh).  I was thinking of going ahead with the 1080 ti, but I kinda feel back on the fence again, blast!!!  It really seems like getting both and testing each to return the weaker one will be the way to go...  I will definitely share the info if I happen to get both before you do - please do the same for me!

[O_and_N]

6 hours ago, silent_sight said:

O_and_N - wow, thanks for all the info and the link to the thread in the autodesk forum!  I'm sure there are more people than us looking for comparisons between those 2 cards.  That was a lot of stuff to chew through... It looks like the gtx cards performed really well in viewport 2.0, and even some old quadro cards did pretty well there too.  It was pretty obvious that the legacy viewport kinda needs a quadro; but even then, VP2.0 was way faster anyway.

 

I get the impression that if I was to do a lot of gpu rendering (with redshift/octane), then a few 1080 ti cards would be best since it would provide way more cuda cores for the rendering.  It also seems like it wouldn't matter much which card I go with if using VP2.0 (unless I'm just not really reading everyone's feedback correctly on the thread you linked to); legacy viewport is where gtx cards are notably weaker than quadro cards.

 

Due to some of life's changes I may be moving soon, so I was holding off on purchasing/shipping anything until that is determined (don't want to ship something to a location I may move away from, heh).  I was thinking of going ahead with the 1080 ti, but I kinda feel back on the fence again, blast!!!  It really seems like getting both and testing each to return the weaker one will be the way to go...  I will definitely share the info if I happen to get both before you do - please do the same for me!

Well we are both on the fence sadly.As  said i see many people saying gtx are perfoming just as good but i still have that gut that tells me they are not that experienced and that they probably dont use maya as we do.Here is one thread that does confirm some of my fears.

 

http://www.tomshardware.co.uk/answers/id-3331673/gtx-1070-performance-maya-buying-card.html

 

Doesnt matter if its a firepro,will will just treat it as a quadro.Looks like the gtx has the same vieport perfomance as the amd pro gpu,but when he selects stuff,animation playback etc,the gtx1070 lags.Im confused as others in my thread do say they dont notice any manipulation lag with a gtx.Im not sure what maya ver. is the guy with a w9100 using but its probably one of the latest.

 

To tell the truth,look at how xsi was.A vieport that didnt needed a pro gpu to go fast.Or modo,or cinema.I suspect autodesk is getting payed by nvidia and amd to make the perfomance that way,as if it wasnt,than they wont sell any overpriced gpus...

 

If i knew this years ago,i woud have probably changed software.Still i have a maya perpetual license and i do like maya so i will have to make the most of it.

 

Im waiting to find a full review on the p4000.If it at least in gaming power is equal to a gtx980(i suspect it is)i will probably go with that one most probably and 2x2k screens.

 

Tough chouce man,tough choice.If you decide on taking advantage of amazon with a 1080 or p4000,please let me know here.

 

Cheers!

 

 

 

 

 

[Immitem]

31 minutes ago, O_and_N said:

Well we are both on the fence sadly.As  said i see many people saying gtx are perfoming just as good but i still have that gut that tells me they are not that experienced and that they probably dont use maya as we do.Here is one thread that does confirm some of my fears.

 

http://www.tomshardware.co.uk/answers/id-3331673/gtx-1070-performance-maya-buying-card.html

 

Doesnt matter if its a firepro,will will just treat it as a quadro.Looks like the gtx has the same vieport perfomance as the amd pro gpu,but when he selects stuff,animation playback etc,the gtx1070 lags.Im confused as others in my thread do say they dont notice any manipulation lag with a gtx.Im not sure what maya ver. is the guy with a w9100 using but its probably one of the latest.

 

To tell the truth,look at how xsi was.A vieport that didnt needed a pro gpu to go fast.Or modo,or cinema.I suspect autodesk is getting payed by nvidia and amd to make the perfomance that way,as if it wasnt,than they wont sell any overpriced gpus...

 

If i knew this years ago,i woud have probably changed software.Still i have a maya perpetual license and i do like maya so i will have to make the most of it.

 

Im waiting to find a full review on the p4000.If it at least in gaming power is equal to a gtx980(i suspect it is)i will probably go with that one most probably and 2x2k screens.

 

Tough chouce man,tough choice.If you decide on taking advantage of amazon with a 1080 or p4000,please let me know here.

 

Cheers!

 

 

 

 

 

 Hi there. I was the one who created that thread (which seems to be thrown around a lot recently) and you are correct in that I am running a relatively recent version of Maya (2016 SP6).

 

 A little while ago I got my hands on a Titan XM for texture painting in SP2 before Firepro support was officially added and now I use it mainly for GPU rendering alongside the two 1070s but thought what the hell and tested it in Maya. To cut to the chase, the Titan XM is slightly faster than the 1070 and w9100 in viewport navigation but it lags behind the firepro everywhere else and in much heavier scenes than what I originally tested a couple of months ago the difference only grows! A new problem rears its head where the geforce cards can take up to a whole second or two before the view begins to pan/orbit/zoom whereas the firepro just does. I will be creating a comprehensive list of the 3 card's performance tomorrow after work.

 

 I will post them in the thread here: https://forums.autodesk.com/t5/maya-forum/are-quadros-and-firepros-still-important/td-p/6772440

 

 P.S. This is only inspiring me more to create a proper video review of the w9100!

[O_and_N]

1 hour ago, Immitem said:

 Hi there. I was the one who created that thread (which seems to be thrown around a lot recently) and you are correct in that I am running a relatively recent version of Maya (2016 SP6).

 

 A little while ago I got my hands on a Titan XM for texture painting in SP2 before Firepro support was officially added and now I use it mainly for GPU rendering alongside the two 1070s but thought what the hell and tested it in Maya. To cut to the chase, the Titan XM is slightly faster than the 1070 and w9100 in viewport navigation but it lags behind the firepro everywhere else and in much heavier scenes than what I originally tested a couple of months ago the difference only grows! A new problem rears its head where the geforce cards can take up to a whole second or two before the view begins to pan/orbit/zoom whereas the firepro just does. I will be creating a comprehensive list of the 3 card's performance tomorrow after work.

 

 I will post them in the thread here: https://forums.autodesk.com/t5/maya-forum/are-quadros-and-firepros-still-important/td-p/6772440

 

 P.S. This is only inspiring me more to create a proper video review of the w9100!

That is so cool!Lets start filling more missing pieces!I feel i know the aswer that pro gpus are still needed but if you have the time,when you start writing your review it will most probably be the best option to tell us the right way of installing different gpus/drivers and in what order.I was over at the octane forums and people do tend to use quadro/firepro with a few gtx for gpu rendering.Problem seems to be that at windows startup,gtx drivers tend to force themselves over the main quadro driver???

[Immitem]

I just finished the benchmarks and you can find them here: https://forums.autodesk.com/t5/maya-forum/are-quadros-and-firepros-still-important/m-p/7005008#M41216

 

 In these sets of tests the 1070 appears to really pull ahead of the XM, interesting!

[O_and_N]

2 hours ago, Immitem said:

I just finished the benchmarks and you can find them here: https://forums.autodesk.com/t5/maya-forum/are-quadros-and-firepros-still-important/m-p/7005008#M41216

 

 In these sets of tests the 1070 appears to really pull ahead of the XM, interesting!

Great contribution!So pro drivers come usable if you have a scene composed of many independent object and if you want to avoid those 1-2 seconds pauses,and maybe skeletal animation.About right?

Im not sure how bareble are those pauses if you spend 4-5 hours a day in maya.Coud it become nerve breaking?

 

As a final,which gpu do you prefer to keep on going in maya?

[scottyseng]

7 hours ago, Immitem said:

-snip-

Very interesting results. I could actually see the firepro / quadro pulling ahead in smoothness overall though. They are optimized for it, probably seen in high polygon models.

 

4 hours ago, O_and_N said:

-snip-

Not directly related, but I have noticed adding vray to 3DS Max makes it lag now (The UI side). You can tell just by messing with the render settings, there's constant pausing between clicking through render settings.

 

I think that's more vray than GPU related I think though...

[silent_sight]

immitem - thanks for taking the time to do your testing; hope you didn't pull your hair out with all that lagging with the gtx cards!  It's really interesting how much better the pro card worked than the gtx cards - seems like the only real exception was that the gtx cards had slightly higher fps in VP2.0, yet they lagged a lot.  I think I would rather have slightly slower fps with little-to-no lag.

 

It looks like your test is with a pretty huge scene and I'm curious if there's much of a difference with a smaller scene (like an average sized character, for example).  I'm mainly just curious... I wonder if a lot of people who do not have any issues with gtx cards simple haven't pushed it to the point where the quadro really shine?  If that's the case, it would be helpful info for people when choosing a graphics card based on their maya usage (if you're not likely to work with such huge files, then maybe a gtx card really will be enough for you).

 

Does anyone know how the AMD w9100 compares to the quadro p4000?  From very briefly looking into it, it seems the w9100 is really old, so I'm guessing the p4000 would be a lot better, right?