Folding Forum

I was previously folding on a GTX 780, and just installed a GTX Titan X and tried to start folding. However, no GPU slot is configured and when I try to add it, no slot is configured. Do I need to wait for an update to FAH to see the Titan X?

Reinstall the fah client so it can detect the change in hardware. Titan X has been supported a long time already. No updated software needed.

I reinstalled the FAH client but no change. The Titan X was just released the other day, and is a GM200 GPU.

Edit: Just reinstalled again, trying to add a GPU slot returns "No available GPUs"

First, be sure you're installed the latest drivers. Then report the hardware codes which can be found using lspci or GPUz or several other possible sources, including the utility from the hardware manufacturer.

I'm running 347.88 drivers which are the newest to support the Titan X and here is the output of GPUz:

7im wrote:Reinstall the fah client so it can detect the change in hardware. Titan X has been supported a long time already. No updated software needed.

Titan X is a new GPU, release date - March 17.2015.

Sorry, confused with Titan Z added a year ago.

Delete the GPUs.txt file so the client downloads the new one I just updated and published.
Or grab a copy from here.

Would be cool to see some benchmarks using FAHBench.

FAHBench info.

Excellent, the new GPUs.txt got it working!

I ran FAHBench under stock clocks (boost at 1189MHz) using OpenCL and got the following:

Explicit Single Precision: 77.8121 ns/day
Implicit Single Precision: 336.337 ns/day
Explicit Double Precision: 6.96 ns/day

PPD on each WU?

I'm getting about 375k PPD with the Titan X right now. Project 9201 with 1:29 TPF. PPD is fluctuating a little so I'll run it for a day or so and see what it averages.

Edit: TPF down to 1:27.

Running two Titan X's @ 1.24 GHz GPU clock on an Gigabyte Z97X, i7-4790K @ 4GHz, Win7 x64.

On Project 9411 with 3:32 TPF produces 402,407 PPD on each via HFM v0.9.3, 18% through both work units.

Afterburner shows GPU utilization between 86% and 88%. Temps 63 to 64 Deg C

Which of the FAHBench tests correlates best with PPD?
For some reason I have in the back of my head that the single precison implicit is the best but I can't remember why.
anyone?

single precision explicit ...

But past performance is no guarantee..

FAHBENCH has not been updated for some time.

It does not predict how future cores will utilize an architecture.

Is Core 18 and core 19 predominately single precision ? IDK.

You could ask on reddit...but responses have been coming at the rate of every other month.