Folding Forum

New FAH viral ad campaign. We're also working on a new landing page for FAH and a new video to advertise FAH. This new web/video campaign is coordinated with new client/installer changes to make FAH easier to install and run, especially for those new to Folding@home.

OK, I'm going to come out and say this is a bad idea.

The user experience for a noob has actually decreased with the new version.

I wouldn't say this was ready for promoting until key issues are resolved.

1. driver issues with AMD graphics card
2. optimized cores for Keepler
3. Ability to throttle GPU usage for the average user who feels the screen is unresponsive with the default 100% gpu usage.

Bad initial experience will drive away new users faster than delay.

JMHO...anyone else?

I agree that launching it now seems like a bad idea, but Dr. Pande didn't say it was coming really soon, and his comment about client/installer changes suggests he's talking about an upcoming performance level option that appears to have GPU throttling included (also finally defaulting to smp:n-1 if AMD GPUs are present). If I had to guess, the client will default to medium, or at least recommend it for new users. I expect the new client version will get here much sooner than the GPU core rewrite, in which case I'd still agree that waiting for that to hit and increased AMD/Kepler productivity (or even basic functionality with AMD cards on the latest drivers) would be the time to start advertising. But I'm not remotely a marketing expert.

This thread seems to be falling back on the old topic of "What's holding back FAH?" And 2 of the big items there were CPU throttling for SMP, and GPU throttling.

Just my opinion, but those performance levels are way too course and don't seem to fix those 2 big items. Folding on (total cores/2) appeals to no one. GPU throttling at 80% might, if it works. Will have to wait and see...

Like mdk777 said, I agree that a lot more ducks need to be lined up before taking anything viral. But I disagree with him about V7.2.9 being less friendly than the previous version. The new version supports Kepler GPUs by default, and auto updates the GPUs.txt. And improved the Linux and OSX clients. Bad AMD drivers are not related to a client version. Neither is Kepler performance, but at least Kepler works in this version. Yes, bad initial experiences will drive away new users, but the client is improving and moving forward, not falling backwards.

but the client is improving and moving forward, not falling backwards.

My point is that from an insider/experienced user this might be the case.

However a totally new user, as of today, might have a lower chance of an initial good experience.

While before, you needed to pursue advanced installations to break things...now you have a greater than 50% chance of the default instal not really working(to the users satisfaction).

You really want that to be much higher positive experience before promoting widely.

I agree, we all know what can be improved. I just think the negative backlash from poor initial experience is high with the issues at this time.
I would hold out for at least some of these issues being resolved before making a launch push.

I agree that there are some issues that need to be fixed so that the new user does not have a bad experience before a big marketing campaign should be initiated. A bad experience can turn a new folder away permanently.

Specifically:

1. Kepler's need to work out of the box. I mean general release level (not advanced or beta).
2. AMD GPU's need to work properly with SMP without requiring modification of the number of cores used in the SMP slot.
3. Current release AMD drivers work for folding.
4. Single user installs work for non-administrators.
5. %GPU usage slider that can be used to minimize lag. Preferably automated so that installation sets the slider to an appropriate setting.

A new user expects folding to work without significant issues or requiring user intervention. I would argue that Stanford needs to meet those expectations before a large marketing campaign should be instituted.

1 is done.

Before considering a significant investment in advertising, it would be nice if the broken python(abi) dependency that keeps FaHControl from running or even installing on Fedora Linux could be fixed. All that really needs to be done is to change the package dependency from strict 2.6 to 2.6+. This has been broken for over a year now. FaHControl runs just fine under 2.7.3 after the necessary hacks have been applied, but new users are very likely to be put off by this.

csvanefalk wrote:...It would be nice if the broken python(abi) dependency that keeps FaHControl from running or even installing on Fedora Linux could be fixed....

Thanks. See Ticket #858.

bruce wrote:

csvanefalk wrote:...It would be nice if the broken python(abi) dependency that keeps FaHControl from running or even installing on Fedora Linux could be fixed....

Thanks. See Ticket #858.

Seen it before, that ticket was opened 8 months ago, it appears somebody updated it just now though, which is nice. Hopefully we will see a fix soon.