Folding Forum

And the bots are back ...

In total there are 3351 users online :: 5 registered, 1 hidden and 3345 guests (based on users active over the past 5 minutes)

I've been using a 1070 for years and it is producing between 1.2 and 1.5 million PPD. Your 1080 shouldn't be far from these numbers.

If you see values lower than that, something else is wrong.

Post some logs from the client.

And don't have high expectations for such low power things, if it's less powerful than a RPi 5, it won't make the deadlines and will be useless.

GROMACS on GPUs sucks It's still too dependent on the CPU.

And it will already run on ARM CPUs (although big.LITTLE shit is always a pain).

Yes, v8.5.5 seems to be doing fine.

Writing the date at each line would add too much data to the log ...

The date is printed at the beginning of the log and also every time a core is started. Also, the log in v8 rotates everyday at midnight, so you should only have current day logs in current log.

Something is still wrong with your OpenCL/CUDA installation.

Or something is preventing the client fro getting the accesses it needs ... do you have some security software/policy that could be limiting your actions ?

Where are the time per frame and PPD values from each test cases, taken from the same projects ? These are the only values we trust to measure FAH performance.

Base credit has been increased to 28800 for p19229. Does it look more in line with your expected PPD ?

p19229 is indeed below typical PPD on my CPU ...

And don't look at Run 0 / Clone 0 / Gen 0 reports, there's a bug in the work servers that counts all dumped WUs in the project as Run 0 / Clone 0 / Gen 0.

Be careful with PPD display from the client :
- it is not accurate at the beginning of the WU, it will settle after a while
- it is completely delirious if stop and resume FAH, until the WU completes and a new one starts.

And of course, it also varies from project to project on a given GPU.

marchidaniele wrote: ↑Wed Jan 21, 2026 9:32 am What do you mean, normal?

1 hour and 16 minutes wasted.

It's not wasted, it should have restarted from checkpoint.

But yes, this project has a higher than usual NaNs rate. See talks in the beta thread : viewtopic.php?p=372233#p372233

I know two prebuilt FAH v8 templates for Vast.ai : This one is from Firedfly : https://cloud.vast.ai/?ref_id=27485&template_id=353f795455392eb766983b88d13222b1 This one is from JWhy : https://cloud.vast.ai/?ref_id=260637&template_id=f9d0a2158f66836798fe89580760e7fd They are quite simple to u...

You might want to try the latest beta, I reported a few weird behaviours with intermittent connections (I sometimes use my phone as a modem that is not always connected to the laptop) and this one has been fixed : - Delay after upload fail to WS and all CS. If you still have troubles, feel free to o...

The latest beta fixes the issue of config.xml errors preventing the client from starting :

v8.5.6 changedlog wrote:- Failing ``config.xml`` load logs error but is now non-fatal.