950 FTW vs 1050ti SSC vs 1050ti FTW SHOWDOWN

[H0R53]

I have the liberty of getting new GPUs about once a month, either through purchase, sampling, or other means like Chinese auctions (most of the time someone at the factory is lifting cards).

 

I've had the GTX 950 FTW, the 1050TI SSC, and the 1050TI FTW.

I benchmark and overclock each card as soon as I can get my hands on it, so here are some results.

 

These are the best of 10 passes.

 

First up, the GTX 950 FTW

Better results in DirectX 9 than the other two, but this is within margin of error. I got this card as part of a trade off Craigslist - I traded a Quadro FX 3800 for the 950 and a 550ti for an older nForce 680i build. That was after my 1050ti SSC died for the last time.

 

Next up, the 1050ti SSC

Stupidly good performance, the extra 2GB of vRAM really help this card over the 2GB variants and the 2GB of the 950FTW. I overclocked this card too much and popped a capacitor twice, and after I fixed it the third time it refused to work.

And finally the 1050ti FTW

This card really blasted through the CUDA benchmarks, like Open CL, bitonic sort, and liquid physics.

 

I just got this card in the mail today.

 

Keep in mind final scores are not a direct representation of the card's overall performance. If there was a Quadro version of the 1050ti the GPU Compute score would be through the roof.

[Jurrunio]

Is there a way to put the 3 scores into 1 graph? Inconvenient to read

[H0R53]

1 minute ago, Jurrunio said:

Is there a way to put the 3 scores into 1 graph? Inconvenient to read

 

I'd rather have the scores for each test present as well - the graph shows how each card performs relative to loaded paselines.

As I don't pay for Passmark uploading these baselines is impossible after the trial is over. I have to wait for a new build of the program.