Some FluidMark hybrid Physx results (GT 545, GTX 650, GTX 650 Ti)

[aznative]

For a brief time this weekend I had a GT 545, GTX 650, and GTX 650 Ti on hand. I ran FluidMark 1.5 on all three of them using setting similar at which I game. 2560x1440 with 4x anti-aliasing.

I cranked up the particle count to 120,000 to put the burden on the Nvidia cards. My primary GPU is an ATI Radeon 7970. My CPU is a i7-2700K at 4.1GHz.

Here's how the PhysX scores broke down:
CPU: 7 SPS
GT 545: 19 SPS
GTX 650: 27 SPS
GTX 650 Ti: 38 SPS

I also followed the same benchmark setup in Test 3 found at Geeks3D to compare against the GTX 480, 580, and 680. In this scenario the GTX 650 Ti ran at 39 SPS, or about 89% of the GTX 680.

I went with a hybrid setup specifically for Borderlands 2. Subjectively all of these cards provide the same gaming experience at the graphics settings I use (2560x1440, SMAA injector, all settings maxed out).

The GT 545 is the real bargain in this scenario. I got a refurb for $70 after shipping on Ebay. It's single slot, doesn't need a power connector, and the Zotac version I got is very quiet.

However, I am going to keep the GTX 650 Ti for my hybrid setup and use the GT 545 in my wife's computer. Both the GTX 650 and Ti are very quiet also, but they do need an extra power connector and are dual slot.



Screenshots of FluidMark results:
CPU: http://i.imgur.com/aC8yf.png
GT 545: http://i.imgur.com/VvQhU.png
GTX 650: http://i.imgur.com/h3iSl.png
GTX 650 Ti: http://i.imgur.com/nXscR.png

 

[unreal89]

Interesting results.

It's sad that proprietary technology such as this is being used in games. Makes AMD less competitive and forces even AMD owners to fork over money to nVidia for full game features.

That being said I'm probably only buying Nvidia because of physx.

 

[Parja]

Both the GTX 650 and Ti are very quiet also, but they do need an extra power connector and are dual slot.

Which, for the GTX 650, is bizarre considering it has a TDP of only 64W.