Question about a Mac Pro and 3D Animation

[Tutor]

Originally Posted by KBS756

I currently own a 2009 Mac Pro with 16GB of RAM and Dual 2.93Ghz Quad Core Xeons and an ATI Radeon 5870.

I am going to be taking some 3D modeling and animation classes next year and intend to do some Rendering and Animation for work purposes as well as some personal & school projects so I was wondering for 3D animation what would be a more effective upgrade a Graphics Card or a CPU Swap to Hex-core Xeons to get the most out of my Pro especially in when rendering.

Autodesk Maya and Newtek Lightwave are the 2 programs I will be learning and using if that makes any difference. Also use After Effects a decent bit.

Thanks

For what its worth, here are my recommendations in order of my perceived importance to you (No. 1 will render 3d stuff {and some CS6 functions such as PhotoShop effects and Premiere} as well as certain other apps { http://www.nvidia.com/object/gpu-applications.html download and review the PDF} many times faster than No. 2 for a little more than 1/2 the price of No. 2, but No. 2 will make almost everything else noticeably faster):

(1a) Get this video card [ http://www.ebay.com/itm/Nvidia-GTX-5.../261112969627? Don't forget that the seller recommends adding an additional power source] because it's fast, has 3 gig of video ram, already modified, has CUDA capability (download CUDA software for Mac here: [ https://developer.nvidia.com/cuda-downloads ] for rendering - it costs $650 (w/o shipping); and (1b) go to Octane's site and read here [ http://render.otoy.com ] and watch here [ http://render.otoy.com/videos.php ] to get some information about what CUDA can do for you using Octane render; after you get the video card, download the Octane render for 64-bit Mac demo here [ http://render.otoy.com/downloads.php ] and if you like it purchase the Octane render Standalone/Maya plug-in combo for about $364 (USD - price fluctuates based on currency exchange rate) here [ http://render.otoy.com/shop/maya_plugin.php ]. And if that still doesn't make your day read on.

(2) Drop about $1.8k to get a pair of OEM x5680's from here: https://www.eoptionsonline.com/p-2049-594880-001.aspx . There are many threads on this forum about doing the CPU swap on the dual 2009 Mac Pro. Before the swap, you need to do the EFI hack [ http://forum.netkas.org/index.php/topic,852.0.html ] required to get the dual hexs recognized, both of which can be done in short order (I'll even help you). I started doing the swaps in June 2009.

Many have spent more than the combined cost of Nos. 1 and 2 just to do the CPU swap.

[Jduncan0392]

All comes down to what you consider worth it. Is it worth spending a **** ton of money to install 2 hex-core processors with a possible minimal performance increase? If anything get a nvidia gtx 580.

Also, to the guy recommending that gtx 580, do you realize how rediculously overpriced that is? Is that your own ebay ad? Idk why you recommend a gtx 580 that's overpriced over a gtx 680, but a better gtx 580 would be this:

http://www.newegg.com/Product/Produc...82E16814162092

idk the gtx 680's performance when it comes to 3d modelling n stuff like that, but for the most part the 680 stomps on the gtx 580. I recommend to the thread creator looking up a gtx 680's performance and 580's performance based on what you'd use it for, and pick one of the 2 cards.

Gtx 680 :

http://www.newegg.com/Product/Produc...82E16814130768

[Tutor]

Originally Posted by Jduncan0392

All comes down to what you consider worth it. Is it worth spending a **** ton of money to install 2 hex-core processors with a possible minimal performance increase? If anything get a nvidia gtx 580.

If the performance increase from upgrading CPUs in this instance was minimal, I would have completely advised against it, but it's not: 12 real and 12 virtual cores running at 3.33 GHz is a lot faster than 8 real and 8 virtual cores running at 2.93 GHz [ 8 * 2.93 = 23.44; 12 * 3.33 = 39.96; 39.96 / 23.44 = 1.70477815699659 or about 1.7 times faster for 3d rendering ]. Also, if you review the posts on the first page of this thread - http://forums.macrumors.com/showthread.php?t=1504715 - it'll give a better idea of the performance delta that can be expected. However, I recommended that the thread starter get the GTX 580 for starters. Thus, I agree with your ultimate suggestion here.

Originally Posted by Jduncan0392

Also, to the guy recommending that gtx 580, do you realize how rediculously overpriced that is? Is that your own ebay ad? Idk why you recommend a gtx 580 that's overpriced over a gtx 680, but a better gtx 580 would be this:

http://www.newegg.com/Product/Produc...82E16814162092

The GTX card that I recommended has twice the amount of ram of the one that you reference. Video cards with greater amounts of ram sell for a premium since Nvidia prices them to avoid cutting into the profits from its ultra pricey TESLA card sales. Moreover, having sufficient ram is extremely important when rendering 3d animations because animations are usually created on systems with many gigs of ram and the fastest GPU-based rendering solution relies solely on video card ram. And no - I do not own the ebay ad. It's owned by someone who many forum members prize for his video card performance enhancing skills. You'll pay more for a card that he has tweaked and made more Mac compatible so that you don't have to spend your own time doing so. So, all of these factors justify paying a premium.

Originally Posted by Jduncan0392

idk the gtx 680's performance when it comes to 3d modelling n stuff like that, but for the most part the 680 stomps on the gtx 580. I recommend to the thread creator looking up a gtx 680's performance and 580's performance based on what you'd use it for, and pick one of the 2 cards.

Gtx 680 :

http://www.newegg.com/Product/Produc...82E16814130768

I use my systems for creating animations and videos for business and do know how the GTX 680 performs for 3d and how it performs for video. I have video and animation CUDA and OCL render farms with 1 - GTX 690, 4 - GTX 680s, 6 - GTX 580s, 3 - GTX 480s, 3 - GTX 295s, 7 - GT 640s, 9 - ATI 4890s and 3 - ATI 5970s running OSX, Windows and Linux. I agree with you when it comes to buying a card for gaming or video output - there the GTX 6xx series excel vs. the GTX 5xx series. But if you google "GTX 680 vs. GTX 580," you'll find that the consensus is that the GTX 580 excels over the GTX 680 for CUDA based functions like rendering. I consider that significant since the thread starter stated, in relevant part: "I am going to be taking some 3D modeling and animation classes next year and intend to do some Rendering and Animation for work purposes as well as some personal & school projects so I was wondering for 3D animation what would be a more effective upgrade a Graphics Card or a CPU Swap to Hex-core Xeons to get the most out of my Pro especially in when rendering." However, I also agree with you that it all comes down to what the thread starter considers worth it. But since he asked, I added my one cent.

[Jduncan0392]

It's a theoretical 1.7x faster yes, but it depends on program optimization. Also, what has he done to tweak the card itself? Anyone could just overclock a card and increase the performance themselves. I also heard nvidia came out with cuda5, so shouldn't that make use of the gtx 680's insane amount of cuda cores? But, you seem to know more about modeling n stuff than I do, so I'll take your word on it. I'm just good with hardware when it comes to building pcs meant to max games out, not when it comes to professional software.

[Tutor]

Originally Posted by Jduncan0392

It's a theoretical 1.7x faster yes, but it depends on program optimization.

True. It's just that in the drive to have 3d app superiority, 3d app developers have had incentive and need to design their apps to take advantage of multi-core systems. So it is in the 3d arena where that "theoretical 1.7x" advantage is likely to the realized. A 3d app called Cinema4d, made by a company named Maxon, takes advantage of up to 64 cores. My WolfPackPrime0 (see my signature below) has 4 E5-4650 processors, each with 8 physical cores, with each core being dual threaded, for having a 64 core total in Cinema4d. It renders like lightning. In an app like Cinema 4d theory is reality.

Originally Posted by Jduncan0392

Also, what has he done to tweak the card itself? Anyone could just overclock a card and increase the performance themselves.

That's also true. I'm in the process of trying to assess that impact and how to do so optimally.

Originally Posted by Jduncan0392

I also heard nvidia came out with cuda5, so shouldn't that make use of the gtx 680's insane amount of cuda cores?

CUDA 5 had been out for a few weeks now. It allows developers to take a little more advantage of the greater numbers of CUDA cores in the 6xx series. The problem however is that the 6xx's cores, tho they be many, don't have the same ability as the cores in the 5xx series to handle floating point operations, especially high precision ones, as quickly.*/ So to really take advantage of the 6xx's superiority in integer operations vs. it's weaknesses where there are many decimal places, it's back to the drawing board for some developers to tackle the floating point issue differently. Octane developers do appear to have a step up on some other developers in this endeavor.

Originally Posted by Jduncan0392

... . I'm just good with hardware when it comes to building pcs meant to max games out, not when it comes to professional software.

I disagree with you here. "Professional software" is a broad category, covering many needs. Being able to max video game performance relies on some of the same talents/techniques it takes to max professional video playback performance. The 6xx series is superior at gaming and video playback. I bought mine for the latter reason.


Another important consideration in Octane's approach to handling rendering completely on the GPU, is that you can achieve a better design experience by having another video card (that's where I mainly use my modest 4 gig GT 640 video cards or my ATI 4980s - not relying at all on them to compute in the renders) to handle interactivity while the other more powerful video cards handle the CUDA rendering chores. Moreover, since Octane relies solely on the GPU to do the heavy lifting, while performing octane renders, I still have almost 100% of my CPU(s)'s ability to do as I need, which can include doing other 3d rendering and video chores. It's like having, and this isn't theoretical only, or should I say, "It's having multiple powerful computing systems in the same box." This is how the Otoy, the owner of Octane, describes the advantages, in part: "... . With current GPU technology, Octane Render can produce final images 10 to 50 times faster than CPU unbiased render engines, or even more with multiple GPUs (depending on the GPU(s) used). .... Octane scales linearly with the number of CUDA cores within a given GPU architecture (Tesla, Fermi, Kepler). For example, there are much more CUDA cores on Kepler GPUs but they are individually simpler and less performant than Fermi CUDA cores and can therefor not be compared directly in number and performance."[Emphasis added] So adding a strong CUDA card can be like adding, at least, 10 CPUs per CUDA card to the mix.

*/ I, like some others, believe that the Nvidia did so, at least in part, so as not to cripple their TESLA card sales.

[theSeb]

Originally Posted by Jduncan0392

All comes down to what you consider worth it. Is it worth spending a **** ton of money to install 2 hex-core processors with a possible minimal performance increase? If anything get a nvidia gtx 580.

Minimal performance increase? What do you base this on? Your extensive experience in playing games?

Also, to the guy recommending that gtx 580, do you realize how rediculously overpriced that is? Is that your own ebay ad? Idk why you recommend a gtx 580 that's overpriced over a gtx 680, but a better gtx 580 would be this:

http://www.newegg.com/Product/Produc...82E16814162092

Always stop to think for a couple of minutes before accusing someone of being a shill in public.

idk the gtx 680's performance when it comes to 3d modelling n stuff like that, but for the most part the 680 stomps on the gtx 580. I recommend to the thread creator looking up a gtx 680's performance and 580's performance based on what you'd use it for, and pick one of the 2 cards.

Gtx 680 :

http://www.newegg.com/Product/Produc...82E16814130768

If you don't know, then perhaps stick to threads where people are asking about games performance? When it comes to "3d modelling n stuff like that" the 580 "stomps" all over the 680.

[thekev]

Originally Posted by Tutor

The GTX card that I recommended has twice the amount of ram of the one that you reference. Video cards with greater amounts of ram sell for a premium since Nvidia prices them to avoid cutting into the profits from its ultra pricey TESLA card sales. Moreover, having sufficient ram is extremely important when rendering 3d animations because animations are usually created on systems with many gigs of ram and the fastest GPU-based rendering solution relies solely on video card ram.

Agreed although I'm curious what you're using in terms of a mature GPU driven renderer. Texture maps seem like they would be painful on GPU based rendering solutions. Also I wouldn't be surprised if they tried to protect tesla sales, especially given NVidia's troubles overcoming fabrication costs.

Originally Posted by Tutor

CUDA 5 had been out for a few weeks now. It allows developers to take a little more advantage of the greater numbers of CUDA cores in the 6xxx series. The problem however is that the 6xxx's cores, tho they be many, don't have the same ability as the cores in the 5xxx series to handle floating point operations as quickly.*/ So to really take advantage of the 6xxx's superiority in integer operations vs. it's weaknesses where there are decimals, it's back to the drawing board for some developers to tackle the floating point issue differently. Octane developers do appear to have a step up on some other developers in this endeavor.

I'm curious how they're going to work around this.

[Tutor]

Originally Posted by

Agreed although I'm curious what you're using in terms of a mature GPU driven renderer. Texture maps seem like they would be painful on GPU based rendering solutions. Also I wouldn't be surprised if they tried to protect tesla sales, especially given NVidia's troubles overcoming fabrication costs. ... . I'm curious how they're going to work around this.

I mainly use Maya and Cinema 4d for 3d tasks. I'm learning to use Octane. Otoy describes Octane as follows: "Octane Render is a true GPU based engine, not a hybrid CPU/GPU, nor a CPU based engine that was modified with GPU acceleration. Since a CPU adds a negligible speedup to the total rendering speed, having your CPU free for other tasks while rendering with Octane Render allows you to use your system for other tasks while rendering."

Here's what Otoy (Octane) has posted in the FAQ section of its website (it was posted prior to late November of this year when Otoy released Version 1.0 - Final ) and it is still there: "If you are interested in purchasing a new graphics card to use with Octane Render, the Geforce GTX570 or GTX 580 currently have the best Performance to Price ratio. The latest generation of Nvidia GPUs (Kepler) is supported, but currently works slower than their Fermi equivalents. We are still optimizing the performance of Octane on the Kepler GPUs. The GeForce line is higher clocked and renders faster than Quadro and Tesla GPUs, but the latter GPUs often have more memory. A powerful multi-core CPU is not required as Octane does not use the CPU for rendering, but a faster CPU will improve the scene voxelizing speed." The main piece of software had been in various beta stages for at least a couple of years.


Here's what the user manual says about GPU selection: "OctaneRenderTM runs best on Fermi (e.g. GTX 480, GTX 580, GTX 590) and Kepler (e.g. GTX 680, GTX 690) GPUs, but also supports older CUDA enabled GPU models. GeForce cards are fast and cost effective, but have less VRAM than Quadro and Tesla cards. OctaneRender scales perfectly in a multi GPU configuration and can use different types of Nvidia cards at once e.g. a GeForce GTX 260 combined with a Quadro 6000. .... .Looking to buy a new GPU for OctaneRenderTM? There are several things to consider when purchasing a new GPU. You'll want to purchase a video card with the largest amount of RAM (we recommend a minimum of 1.5 GB video RAM), with the most amount of CUDA Cores for your budget. Make sure your Power Supply can handle the new card as well. If you're using a Mac, make sure that you purchase an Apple approved GPU." I cannot locate it now, but I remember having read something by Otoy that recently gave a surprising glowing assessment for users of the GTX 680 and 690, but that may have been driven, at least in part, by the dwindling supply of the GTX 5xx series.

Additionally, here's two reviews of Octane: (1) http://www.evermotion.org/tutorials/...-3dsmax-review
(2) http://www.3d-sphere.com/renderers/o...nder-10-review

And checkout raytracey's blog: http://raytracey.blogspot.co.nz

You can download the Octane manual (in PDF format) from here: http://render.otoy.com/downloads/Oct...UserManual.pdf

[KBS756]

Thanks to everyone for providing your insight on this topic. I think I am going to try to see if I can budget in both a CPU and a GPU upgrade if possible. I have a couple questions though before I do it.

On the CPU side I am a little cautious about it due to the fact I have seen things like the article on Anandtech where he managed to fry the CPU board along with the processor and I have a 09 Mac Pro as he did so I am a little worried about the same thing happening to my Mac Pro so if someone can link me to a detailed tutorial with either video of pictures it would be really helpful. I have built computers before, but they had clips holding down the processors into their sockets.

Also would a Memory upgrade be of any benefit to me? I currently have 16 GB of 1066 MHz Memory in my Mac Pro 8*2GB.

In terms of the GPU the first one to reply to my post noted that the Quadro line of Nvidia cards are more stable than the GeForce line so wouldnt the 580 suffer the same issues he noted? Also you noted the requirement for outside power on a 580/680, in relation to that would I be able to get that additional power without installing something outside the case? or would solutions like that lead to increased heat and noise which I should avoid? Also my displays are currently 2 Apple Cinema Displays a 24 inch and a 27 inch how would I run these 2 through a card like a 580 which seems to only have 1 display port. I remember seeing adaptors from dual dvi but remember them being fairly expensive. Also since the 680 is available with 1 more GB of memory and seems to have a higher clock speed for memory shouldnt it be faster in relation to even 3D applications? Also I noticed before someone had posted on these forms that OSX requires some sort of hack to see more than 2 GB of memory on a video card? Is that still the case and if so how do I get around that.

MacVideoCards seems to be down for the holidays so I will look him up after the holidays since I am trying to teach myself hopefully wont hit issues related to performance for a couple weeks.

Lastly you mention Octane as a renderer, are there any other renderers that benefit from a fast GPU, or for most of those would a fast GPU be ideal? Thanks again for the great help on this topic. I just finished up finals so I should be able to start diving into this soon. Thanks again.

[Tutor]

Originally Posted by KBS756

... . On the CPU side I am a little cautious about it due to the fact I have seen things like the article on Anandtech where he managed to fry the CPU board along with the processor and I have a 09 Mac Pro as he did so I am a little worried about the same thing happening to my Mac Pro so if someone can link me to a detailed tutorial with either video of pictures it would be really helpful. ... .

I have 4 Mac Pros. One is 2010 look and act alike. When I acquired it, it was a refurb 2.23 GHz dual 2009. While Anand was waiting for his well done CPU board to be replaced, I installed 2 - x5580s successfully in mine. Since then I have installed all manner of 56xx pairs and singles in it (after applying the EFI hack) for various reasons. I don't have pics or video. I don't remember where you can find pics or videos by others on this forum, but I do recall having seen pics posted by others who later did the CPU swap and discussed it on this forum. In 2009, I helped gugucom, among others, with the CPU swap and as I recall he did post some pics on this forum.

Originally Posted by KBS756

Also would a Memory upgrade be of any benefit to me? I currently have 16 GB of 1066 MHz Memory in my Mac Pro 8*2GB.

Yes, especially for 3d and video apps and even more so if you go the Octane route because Octane taxes only the GPU; so while the GTX 580 GPU is taxed to the max rendering a 3d project, and especially if you still use your GT 120 [which you can also use for increased interactivity while setting up your 3d scenes - see Octane videos to get a better idea of what I'm referring to] , you could be rendering video simultaneously with, e.g., Premiere CS6, using the CPU. For such work you'll never regret having maximum memory. Should you also go with upgrading the CPU to the 56xx series (after applying the EFI hack), your system can also take advantage of the faster 1333 MHz memory.

Originally Posted by KBS756

In terms of the GPU the first one to reply to my post noted that the Quadro line of Nvidia cards are more stable than the GeForce line so wouldnt the 580 suffer the same issues he noted?

Since I disagree with the first premise, i.e., that the Quadro line of Nvidia cards is more stable than the GTX, you should be easily able to discern that my answer to the projection regarding the GTX 580. What can make any video card unstable is overtaxing it, denying it the power it craves and overclocking it to the point that thermal issues arise.

Originally Posted by KBS756

Also you noted the requirement for outside power on a 580/680, in relation to that would I be able to get that additional power without installing something outside the case?

That was in reference to one specific cards that had lots of ram as far as GTXs go and it already had been "touched" by MacVideoCards. That a card is a GTX 580, 680 or whatever says nothing really about what you need to fully hook it up. There are sooo many varieties that it can make one's head spin. You should be able to find one with just 2 - 6 pin power connectors, but it'll probably not be one of the faster ones - those require more power; so they may require 1 - 6 pin and 1 - 8 pin, or as in the case of mine 1 - 6 and 2 - 8 pin connectors. But keep in mind that I now build my own systems so that doesn't pose a problem for me.

Originally Posted by KBS756

or would solutions like that lead to increased heat and noise which I should avoid?

One with common sense are you. You got it. As a general rule, more speed -> more heat -> more cooling needed -> more noise, unless you're building your own and pick the quietest fastest fans or go with H20 or something more exotic. In the end, this is a personal/subjective matter however.

Originally Posted by KBS756

Also my displays are currently 2 Apple Cinema Displays a 24 inch and a 27 inch how would I run these 2 through a card like a 580 which seems to only have 1 display port. I remember seeing adaptors from dual dvi but remember them being fairly expensive.

Yes, adapters can be expensive; but keep in mind that you may be able to find a GTX 580, either already hacked, with 2 display ports, or you may chose to use another or additional video card for interactivity while setting up the 3d scenes and/or to use for CPU based rendering of 3d or video. I use GT 640s and ATI 4890s in this way.

Originally Posted by KBS756

Also since the 680 is available with 1 more GB of memory and seems to have a higher clock speed for memory shouldnt it be faster in relation to even 3D applications?

If you chose to explore going the Octane route, contact Otoy and get the latest info about how they're fairing optimizing it for the GTX 6xx series (and if you don't mind, please update all of us).

Originally Posted by KBS756

Also I noticed before someone had posted on these forums that OSX requires some sort of hack to see more than 2 GB of memory on a video card? Is that still the case and if so how do I get around that.

Yes. But don't rush it. Get with (i.e., PM) MacVideoCards when time allows. He knows his stuff and can lead you in a proper direction. Also, there is a thread on this forum where he has walked others step-by-step through the process of making that hack, but I cannot find it at the moment.

Originally Posted by KBS756

MacVideoCards seems to be down for the holidays so I will look him up after the holidays since I am trying to teach myself hopefully wont hit issues related to performance for a couple weeks.

Obviously a wise choice for a wise one.

Originally Posted by KBS756

Lastly you mention Octane as a renderer, are there any other renderers that benefit from a fast GPU, or for most of those would a fast GPU be ideal? ... .

In addition to the native renderer usually found in and specific to a particular 3d app, there are, among others:

3delight,
AIR,
Arion Render,
FinalRender Stage 2,
Fryrender,
Indigo Renderer,
LuxRender,
Maxwell Render,
Mental Ray / Iray,
Octane Render,
Pixar's RenderMan and
VRay.

Here is where Google can really help you out.

Some may not take any advantage of the GPU. Some are hybrids which take advantage of both the GPU and the CPU simultaneously and then there's Octane that relies completely on the GPU. To date, it's my understanding that Octane is the fastest of them all for that one reason. Moreover, with the hybrids some give you the advantage of being able to render much larger scenes than your video card may usually allow on its own, but come with the downside that since they're using the CPU also, you cannot do as much multitaking because the render is now also taxing the CPU.

An interesting free multi-platform 3d app - Blender [ http://www.blender.org ], has a built in renderer (Cycles) that takes advantage of CUDA (and to a much lesser extent ATI Stream cores) [ http://wiki.blender.org/index.php/Do.../GPU_Rendering ]. But it isn't nowhere near as fast as Octane; but then the same apples to all of the others I've listed above.

N.B. Keep in mind that you can use multiple GPUs with multiple renderers.

[KBS756]

Btw I noticed on here that someone got a HD 7970 working in a Mac Pro would this be of any use in terms of viewport performance if I were to use a CPU based renderer. It has 2 mini display ports and by most benchmarks it seems to blow everything away in computing performance. I know ATI doesnt have Cuda so it wont be useful in terms of CUDA accelerated tasks.

http://www.amazon.com/XFX-Double-105...70+ghz+edition


Also is there anyway to channel extra power to make for that 8 pin by using power that would have gone to the 2nd DVD burner? or with a 6 and an 8 pin is it always just external extra power supply ? Also dont have any experiences with external power supplies, how would it be hooked up so it checks if the computer is on and such? would it be 1 6 pin to the board and the 8 pin to the external? And if I were to go down the ATI 7970 route which is the best manufacturer for the cards?

Thanks again

[KBS756]

Also In addition to the ATI Graphics Question above I was looking up options for where to buy X5680 Processors for upgrade and I came across an eBay site

http://www.ebay.com/itm/INTEL-XEON-6...item2ec492157e

Selling what I believe is the same processor as the one someone else got from here for $945 each. Only it is listed as "NewOther" which the seller told me meant "New other means that these processors were pulled from new servers. These are fully functioning chips, the only reason they were pulled is so that they can be sold They will be packing in our packaging"

Also the seller informed me that they provided a 3 year warranty on the processors so if the processors were to fail or have issues they would replace them for me.
http://www.itcreations.com/view_prod...oduct_id=20316

The eBay account that these are listed through seems to have a very good feedback percentage.

While I also noted that another user posted on these forums about buying 2 of these from amazon through a seller named eoptionsonline
http://www.amazon.com/Intel-X5680-Pr.../dp/B003ELYSJQ

It seems he sells the same processors on his site for $900 each. https://www.eoptionsonline.com/p-1285-slbv5.aspx
And this is what he has listed under warranty "(Your Satisfaction Guaranteed) (90-day replacement warranty on all parts) (Full 7 Day DOA warranty [Dead on Arrival], Full Refund Including Shipping)"

Either way I will use a purchase option such as credit card that protects me in case of issues but was wondering which way would be the least problematic route for me to take. I would like to save as much money as I can though so I can put it towards a Graphics card update as well.

So I would appreciate it if anyone has any opinions on either of the sellers listed and the way that these items are listed. Which would be the best option for me to use in this case in terms of upgrading the processors in my 09 Dual Processor Mac Pro. Thanks again everyone for all your help

[Untethered]

Xeon swap for sure, though my friend has a late 2010 Mac Pro and he uses Blender and it renders in 10-20 minutes for a 10 minute video, so i doubt it would be any different on a different program. Plus, yours is 2011, so it's bound to be faster. But I would do the Xeon swap.

Best,
Untethered

[KBS756]

Originally Posted by Untethered

Xeon swap for sure, though my friend has a late 2010 Mac Pro and he uses Blender and it renders in 10-20 minutes for a 10 minute video, so i doubt it would be any different on a different program. Plus, yours is 2011, so it's bound to be faster. But I would do the Xeon swap.

Best,
Untethered

Think I mentioned earlier mine is a 2009 that I'd like to upgrade to with 2 X5680's so was hoping someone could give me advice about my previous 2 posts on where to buy the processors and in Relation to ATI 7970 GHz Edition

Thanks Again

[KBS756]

Merry Christmas and Happy Holidays everyone.

I would appreciate it if someone could give me advice on where to buy the processors as I stated in an earlier post.

I was looking up options for where to buy X5680 Processors for upgrade and I came across an eBay site

http://www.ebay.com/itm/INTEL-XEON-6...item2ec492157e

Selling what I believe is the same processor as the one someone else got from here for $945 each. Only it is listed as "NewOther" which the seller told me meant "New other means that these processors were pulled from new servers. These are fully functioning chips, the only reason they were pulled is so that they can be sold They will be packing in our packaging"

Also the seller informed me that they provided a 3 year warranty on the processors so if the processors were to fail or have issues they would replace them for me.
http://www.itcreations.com/view_prod...oduct_id=20316

The eBay account that these are listed through seems to have a very good feedback percentage.

While I also noted that another user posted on these forums about buying 2 of these from amazon through a seller named eoptionsonline
http://www.amazon.com/Intel-X5680-Pr.../dp/B003ELYSJQ

It seems he sells the same processors on his site for $900 each. https://www.eoptionsonline.com/p-1285-slbv5.aspx
And this is what he has listed under warranty "(Your Satisfaction Guaranteed) (90-day replacement warranty on all parts) (Full 7 Day DOA warranty [Dead on Arrival], Full Refund Including Shipping)"

Either way I will use a purchase option such as credit card that protects me in case of issues but was wondering which way would be the least problematic route for me to take. I would like to save as much money as I can though so I can put it towards a Graphics card update as well.

So I would appreciate it if anyone has any opinions on either of the sellers listed and the way that these items are listed. Which would be the best option for me to use in this case in terms of upgrading the processors in my 09 Dual Processor Mac Pro. Thanks again everyone for all your help.

Also I would also appreciate advice on if an ATI 7970 (Since it has 2 Mini Display ports and I wouldn't need to worry about adaptors) would suffice for viewport performance in Maya, as well as which brand and amount of VRAM would be ideal for 3D Modeling.

Btw I noticed on here that someone got a HD 7970 working in a Mac Pro would this be of any use in terms of viewport performance if I were to use a CPU based renderer. It has 2 mini display ports and by most benchmarks it seems to blow everything away in computing performance. I know ATI doesnt have Cuda so it wont be useful in terms of CUDA accelerated tasks.

http://www.amazon.com/XFX-Double-105...70+ghz+edition

http://www.amazon.com/Sapphire-DL-DV.../dp/B0098HW0H2

And lastly I have never used an external power source for a video card so was just wondering if simply running both 6pin and 8 pin to the external is ideal or if one should go to the tracers on the board and one to the external (if I am running one high power video card)

I would like to do this upgrade soon when I have some free time so I would greatly appreciate some insight so I may get the ball rolling.

Thanks again everyone

[Tutor]

Originally Posted by KBS756

... . I would appreciate it if someone could give me advice on where to buy the processors as I stated in an earlier post.

I do not have a recommendation for you because I've dealt with only one of those vendors - eoptionsonline - and that was positive from a cost point of view.

Originally Posted by KBS756

Also I would also appreciate advice on if an ATI 7970 (Since it has 2 Mini Display ports and I wouldn't need to worry about adaptors) would suffice for viewport performance in Maya, as well as which brand and amount of VRAM would be ideal for 3D Modeling.

I do not have a recommendation for you because I have not used an ATI 7970 with Maya, but have I used ATI cards which appear to have less performance than the 7970 and they were completely satisfactory for viewport performance in Maya. I'm not sure what you mean by your reference to brand of VRAM (aka Video RAM). Generally, you won't have a choice when it comes to the brand of video ram on the video card, but get a card with as much as you can afford.

Originally Posted by KBS756

And lastly I have never used an external power source for a video card so was just wondering if simply running both 6pin and 8 pin to the external is ideal or if one should go to the tracers on the board and one to the external (if I am running one high power video card)

You should get a supplemental power supply with sufficient power to run the high power video card with both power cables connected to the supplemental power supply.

[KBS756]

Originally Posted by Tutor

I do not have a recommendation for you because I've dealt with only one of those vendors - eoptionsonline - and that was positive from a cost point of view.

I do not have a recommendation for you because I have not used an ATI 7970 with Maya, but have I used ATI cards which appear to have less performance than the 7970 and they were completely satisfactory for viewport performance in Maya. I'm not sure what you mean by your reference to brand of VRAM (aka Video RAM). Generally, you won't have a choice when it comes to the brand of video ram on the video card, but get a card with as much as you can afford.



You should get a supplemental power supply with sufficient power to run the high power video card with both power cables connected to the supplemental power supply.

The Question in relation to the point about VRAM I meant it as two separate questions. What I meant was what brand would I purchase in terms of XFX, Sapphire, and brands such as those if anyone had any recommendations of the brands who sell the 7970 sorry I guess I wasnt as clear there as I should have been.

Thanks again Tutor your UserID is very well deserved I really appreciate all the help.

Would you know if theres anywhere I can find a video of how to do the CPU upgrade, still a little nervous and want to be able to double check with a source to make sure I do everything right.

[Tutor]

Originally Posted by KBS756

The Question in relation to the point about VRAM I meant it as two separate questions. What I meant was what brand would I purchase in terms of XFX, Sapphire, and brands such as those if anyone had any recommendations of the brands who sell the 7970 sorry I guess I wasnt as clear there as I should have been.

The brands that I like best are Power Color and H.I.S., but I haven't heard about any negatives for the ones you name specifically.

Originally Posted by KBS756

Would you know if theres anywhere I can find a video of how to do the CPU upgrade, still a little nervous and want to be able to double check with a source to make sure I do everything right.

I'm not aware of any, but you could search Utube.

[xcodeSyn]

Originally Posted by KBS756

...if theres anywhere I can find a video of how to do the CPU upgrade, still a little nervous and want to be able to double check with a source to make sure I do everything right.

I am thinking of upgrading a 2009 DP model, and not sure there is any video available although there has been plenty discussion with photos in this forum. You can probably start with this one, and there are two more detailed threads referenced in it. The OP apparently recorded his upgrade on video and let's hope he'd make it available soon.

[mjvlego]

I am a lightwave modeler. I also use Zbrush some, and sculptris. I was doing modeling and some animation on a 2007 iMac with 2 GB ram. I recently bought a 2012 21" iMac with 16GB Ram, 3.1 GHZ CPU and 1TB Fusion Drive. I rendered this image on my old iMac, my new iMac, my high end PC at work and had a friend who also uses lightwave render it on his 2009 MacPro.

On my old iMac it took 8.5 minutes to render. On my 2009 64 Bit PC with 64 Gig of ram and professional grade video card it took 4 minutes to render. On my friends 2009 MacPro it took 4.5 minutes to render. On my new iMac it took just over 2 minutes to render. Originally my plan was to buy the new iMac and use it until the MacPros come out and then give the 2012 iMac to my wife. Now I'll probably just keep the iMac and use it myself.