NVIDIA GeForce RTX 2080 Ti To GTX 980 Ti TensorFlow Benchmarks With ResNet-50, AlexNet, GoogLeNet, Inception, VGG-16

[coder]

Question about the RTX 2080Ti which may be a wee bit off-topic here:
There have been recent reports about the cards degrading rapidly (see link). I have a machine with dual RTX 2080Ti on order to use for machine learning. Now I'm worried my cards will be RIP before I know it.

phoronix : Have you been hitting your RTX 2080Ti sufficiently hard since you got it to assess whether there is an issue with your sample, Michael?

I know that this would be rather anecdotal evidence with a sample size of 1, just trying to be reassured here.

How long is your warranty? I'd not hesitate to run it as hard as you need, in hopes that it'll die in time to get a free replacement. I wouldn't do any extra burn-in, in case does outlast the warranty. I guess a 3rd option would be to sell it and buy a newer one, but now you're paying a known price to mitigate an unknown risk by an unknown amount.

It's always a risk of adopting new tech. In general, I like to buy products after they've been on the market long enough for these problems to surface. That said, we don't all have that luxury. Sometimes, you need a replacement and it's too tempting to go with the latest, just-released model.

[JanW]

Warranty is two years. We've run the cards hard, and there are no issues so far. Bleeding edge hardware definitely has its drawbacks and getting the machine stable was a hassle, but not a fault of the GPUs it appears. In this case, and for our application, the bang for the buck was decisive. Since the initial failures of the NVidia cards seem to be understood and we have a third-party card which largely seem to have been unaffected, I'm rather reassured. Anyway, after spending big bucks on such a machine, not running it hard is not an option. For us this is a tool, not a gadget.

[coder]

For us this is a tool, not a gadget.

My team has a Titan V card. It's not exactly sexy, when it's just sitting there, training models.

Luckily, I got a chance to run some Unigene demos/benchmarks on it, before it was put into service. I hope I get to use it for VR, someday.