GeForce GTX750ti video card

Jerry K wrote on 7/8/2014, 9:26 PM

I'm looking for information on NVIDIA's GeForce GTX750ti video card. I'm trying to find out if this gpu works with Vegas Pro-13 with out any problems and utilize most or all of the 640 cuda cores.

If anyone here is using this card please give some feedback.

Here's the link with information on the GeForce GTX750ti card.
Once on the link there are 31 pages of information and gpu comparison charts.

http://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_750_Ti/

Here are some highlights of the GTX750ti card:

* Idle wattage 4 watts.
* Maximum wattage card uses 57 watts.
* Card is power by the PCI express slot, no power cable needed, no large power supply needed, a great card if you have a small power supply.
* Core Clock 1020MHz
* Boost Clock 1085Mhz
* Cuda cores 640
* Memory 2GB
* Memory Interface 128-bit
* Price here in the US under $150 Newegg had a 4th of July special for $119.99

Any feedback would be greatly appreciated.

Jerry K

Comments

ritsmer wrote on 7/9/2014, 1:29 AM

As my Hex-core Haswell CPU seems to be severely crippled by my GTX650Ti in terms of previewing and rendering I have started looking for an alternative.

Just saw this comparison: https://compubench.com/result.jsp?benchmark=compu20

In the upper drop-down menu you can chose the test to compare.

The comparison shows, among others, the GTX760.
Might give some ideas...

OldSmoke wrote on 7/9/2014, 7:28 AM

ritsmer

How does your link show how well a card would perform in Vegas? Also the 760 and 750Ti have different GPU architecture, Keppler and Maxwell.

DavidK wrote on 7/9/2014, 8:36 PM

Jerry don't bother with any of the GTX 700 series video cards with Sony Vegas 11, 12 or 13. Due to the new architure in the Kepler and Maxwell cards the are slower than using CPU only. I had a GTX 760 in my system and if I rendered anything using the GPU, it would take almost twice as long than if I rendered using only the CPU.

You are better off with an AMD card or one of the older GTX 500 series cards.

This is also true with Movie Studio.

Dave

Jerry K wrote on 7/9/2014, 9:54 PM

Thanks for all the information. Being my GTX470 is working fine I'll just stay with it for now and see how it works out. Just in case anyone is interested I found a Sony Vegas Pro 12 gpu render test on 10 different video cards at this link below. Just scroll down to the third test for vegas.

http://www.anandtech.com/show/7481/the-amd-radeon-r9-290-review/14

Jerry K

ritsmer wrote on 8/8/2014, 4:39 PM

OldSmoke wrote:

ritsmer

You are perfectly right: I bought and installed the newest and firebreathing Radeon R9 290.
Result: rendering with GPU On is still significantly slower than CPU only in my daily video-format-and FX's environment.
MSI afterburner shows that the R9 290 works at up to nearly 100% - while the CPU then does next to nothing (below 10 %).

So the R9 290 is returned.

No problem, however, as I always do my many daily test renderings to 720p 25 Fps which renders at- or above real time - and which are well suited for viewing on the computer screen.
Final renderings to Full HD 50i format are done in the background while I'm doing something else - so the time for this is not an issue at all.

Only reason for buying the R9 290: The benchmark numbers looked so nice :- ))))

BruceUSA wrote on 8/8/2014, 6:06 PM

My R9 290X crossfired working awesome here in Vegas.

OldSmoke wrote on 8/8/2014, 6:44 PM

ritsmer

Did you select OpenCL in the render template? And don't leave it in Auto either. Which driver are you using?

ritsmer wrote on 8/9/2014, 12:40 AM

Sorry - it is some 2 weeks back - but I think I did what was described here in the forum.
When I saw that the bottleneck just changed so significantly from the CPU to the R9 290's GPU I gave up and returned the card.

Anyway I always use MainConcept mpeg2 rendering and I think that the codec used in Vegas 12 maybe is not so well tuned to balance the load between a superfast Haswell 6-core and a superfast R9 290 GPU - resulting in max-loading the one of them leaving the other running below 10%.

OldSmoke wrote on 8/9/2014, 5:25 AM

The difference between your 4930 Ivy Bridge and my 3930 Sandy Bridge is marginal. From my system I can tell you that the load between CPU and GPUs, 2x GTX580, is well balanced and I see a vast improvement with GPUs on. I would rather suspect there was something else wrong when you tested the R290.

BruceUSA wrote on 8/9/2014, 9:27 AM

here the MainConcept mpeg2 rendering test on mine system. Render .m2t in 23s and .mv2 in 33s. CPU usage 67-84% GPU both usage ranging from 19%-88%

ritsmer wrote on 8/9/2014, 9:38 AM

EDIT: Ooops, sorry pressed send 5 minutes too late :- ))

Agree, Oldsmoke.
Sounds interesting, however:

Have you tried to watch the CPU/GPU usage balance when rendering, say, some AVCHD Full HD (or the like) to Main Concept Blu-ray Full HD 50i or 60i at 25 Mbps (it's a standard template) ? If both the GPU and the CPU are well over 50 busy at the same time that would be the optimum.

ritsmer wrote on 8/9/2014, 9:43 AM

Bruce: Thank you - maybe I should re-order that R9 290 and do some further experiments - it looks as if there might be something to gain there.

BruceUSA wrote on 8/9/2014, 9:44 AM

You will never get a 50/50 CPU/GPU usage. in my test cpu/gpu are kicking butt and rendering speed is very good and I am please with the results.

ritsmer wrote on 8/10/2014, 3:19 PM

What I would ultimately hope for would be a total synergy like 100% / 100% - but that is, for many reasons, not possible.

What I did not like in the past was either something like 10% / 90% or 90% / 10% where either the CPU or the GPU was just idling waiting for the other one.

Right now here is much editing work to do - but in a month or so there might come time for further tech experiments.