geforce titan or quadro k4000?

I listed observations and testing results. The editing applications that have GPU acceleration and perform better with Geforce cards are professional applications used in all the media content creation industry. Davinci and Adobe are far more prevalent now than what few applications that benefit from the Quadro Open GL plugins. GPU acceleration in general has supplanted Open GL as the primary GPU processing. Since it has, the raw specs of the cards decide the performance. This has been tested and shown with results by far more people than myself. So these are not opinions. The 700 series cards are showing greater performance than the previous gen Nvidia cards so obviously the cards were not limited that much if at all. The applications caching models are currently limiting the performance of the current cards as are the CPU's available. The CPU's themselves have to decode the data first that get's GPU processed and create all of the buffers that transit the data down to the GPU for processing. That is where the current limitation is for the GPU processing load. GPU-Z is reflecting this when monitoring the GPU load during application processing. Only certain codecs with resolutions greater than 2K such as red are using enough frame data to really push the cards. This points to applications limiting performance and not the hardware. There are far more applications now using GPU acceleration that gain the benefit from the specs of the cards than there are applications that take advantage of the Quadro Open GL plugins

Nothing you have mentioned supports any professional stability argument for the Quadro cards. I actually listed a current problem with the Quadro cards that effects any processing unit and that is heat. That has been observed in testing and long term support here with failed video cards. The Quadro k4000 series due to the single slot profile, limited cooling, and poor fan control has the highest failure rate. Those cards are averaging well over 80+C with GPU acceleration applications especially since they are using lower end GPU chips that are often at 75% load or greater. Time, experience, and results show the higher end Geforce cards have a higher stability probability due to heat generation and constant long term load percentage. The only reason scientific, medical, and engineering GPU acceleration applications are used with the Quadro cards is the ECC ram option. When processing data sets that take weeks to complete or have extremely high cost involved in research, errors due to GPU ram is not something those R&D entities want to risk. In those scenarios the Quadro cards are the best choice. That is the only stability based advantage the Quadro cards bring to the table. That however no real impact on media content creation which we are discussing in this forum.

Results and observations don't require a resume of experience to validate. Simply perform the testing as other have done like myself and list them here. I am stating results and observations that have been reported by many others so the validation is the simply the number of results that reflect the same data. If you disagree then report your results and let others either duplicate or disprove. That is not blind faith but simply data. Listing a bunch of marketing material as facts does nothing to validate your argument and are not facts as you state. They are marketing points that can be interpreted any number of ways hence the lack of any real value to the actual editors or content industry in general.

Actually the Geforce cards have better cooling and run much cooler than the Quadro cards on average especially the higher end Geforce cards. The load put on the Geforce cards by games on average is higher than the load Adobe puts on the cards most of the time. The Geforce cards handle rendering long form fine with their temps averaging 40 to 65C including 24/7. Many of our clients have been doing that since CS5 released. Max like many other 3D applications are moving to Direct X so the Open GL acceleration on Quadro cards specific to their drivers is far less utilized. Geforce cards are better than Quadro cards for Direct X. Quadro cards do give you 10 bit color preview in Adobe but I/O such as Blackmagic and Aja are often better solutions for that coupled with a Geforce card. Open GL is an API and Open CL is GPU acceleration. They are completely different and pointless to compare. Cuda is what you compare to Open CL right now. GPU acceleration performance in general with regards to the video cards is tied to the raw specs. That means more cores and ram bandwidth decides the performance you will see from GPU acceleration. The higher end Geforce cards have far better specs than the Quadro's other than the K6000 which is extremely expensive. You could get 2 Titan Black edition cards for half the price which would perform far better. Many Red users use Titan cards and Blackmagic recommends them for Davinci because the performance is the best at lower cost. Geforce cards in general are a far better buy than the Quadro cards for GPU acceleration. We have also seen a lower failure rate with the higher end Geforce cards 760GTX and above than we have the Quadro K4000 series and above. The Quadro's just don't have correct fan profiles for the cooling which allows them to heat up too hot especially over long periods. Nvidia also has not artificially limited the Geforce cards due to the Quadro cards. The Geforce cards are vastly out performing the Quadro cards currently so if they had limited them then they failed badly. The GPU clock speed and memory clock speed likely were reduced because of the amount of cores added to the latest GPU's. That is the main benefit to parallel processing. More cores equals better performance even with far lower clock speed. X86 CPU processing is completely different which is why clock speed has far more impact. Right now the current CPU's are just not pushing the current higher end Geforce cards to their limit. It takes Dual Xeons to push a single Titan at this point due to how GPU acceleration works. The performance ceiling on the 700 series cards is far higher than the 500 series.

Eric

ADK

cc_merchant --

Thanks for your response. i ran across that page earlier today in my searches. it led me to my tag on question of "And given my setup listed above, would the Titan card be massive overkill/be bottlenecked by slower components?"

the article helps me in the abstract but not the concrete/case specific with my setup, unfortunately. Basically, would jumping up in a graphics card be only partially beneficial due to other parts that may restrict it? For example, will i only get maybe 30% of an improvement from a great new GPU due to my processors, which may not be able to fully utilize the card?