Folding Forum

Kamicrit wrote:One little detail I love to spit out when talking about #7 is . Not only is F@H 3x more powerful than any super computer, but it is the only super computer that is self upgrading. You never hear about Stanford buying new IBM racks with every new Intel processor.

That's not quite 100% true, though it's a pretty accurate statement.

You never hear much about it, but servers do need to be upgraded. Stanford does upgrade their servers every few years, though like you say, it's not a big deal. Fortunately it's handled smoothly and efficiently with very little publicity.

As the number of clients increases, they have to increase the number of servers. The total amount of RAID storage may be a big number but it gets bigger every year. The ability of an individual server to maintain adequate responsiveness when dealing with a large number of simultaneous connections does degrade as the number of connections increases, demanding more powerful servers as well as more servers. It's almost insignificant compared to the self-upgrading process that you're talking about, but it does happen.

The software is also improved periodically, both by adding new features to the client software that you install on your computer and by adding new features to the code that runs on the servers. Such upgrades can be pretty tricky, of course, since they have to happen to a real-time system that keeps running. Upgraded server code has to work with a variety of client versions and upgraded client code has to work with a variety of server versions with minimal disruption to everyone. (The only time FAH has actually gone down is when there has been a massive power failure at Stanford, but even that isn't 100% because servers are located in a variety of places on campus and more and more places around the world as time goes on. Redundancy continues to improve on a regular basis.)

I'm not suggesting that you need to cover any of this in your presentation, but it's factual background material that you should know in case you're asked. The actual details are not particularly important to your audience, but you also don't want to tell them something that isn't accurate.

Kamicrit wrote:One little detail I love to spit out when talking about #7 is . Not only is F@H 3x more powerful than any super computer, but it is the only super computer that is self upgrading. You never hear about Stanford buying new IBM racks with every new Intel processor.

I found out that the japs developed in 2009 a supercomp. called Kei which currently has 10 petaFLOPS while FAH has 6. So FAH was only the best (and by a great margin) up until 2 years ago.

Nantes wrote:

Kamicrit wrote:One little detail I love to spit out when talking about #7 is . Not only is F@H 3x more powerful than any super computer, but it is the only super computer that is self upgrading. You never hear about Stanford buying new IBM racks with every new Intel processor.

I found out that the japs developed in 2009 a supercomp. called Kei which currently has 10 petaFLOPS while FAH has 6. So FAH was only the best (and by a great margin) up until 2 years ago.

Actually, we have just under 8.1 x86 FLOPS, which are there because x86 is a standard architecture. Based on FLOP numbers this may be true. But I've heard a number of statements about LINPACK's measurements vs real world tasks. See:
viewtopic.php?p=67416#p67416
viewtopic.php?f=16&t=19118#p191157
http://www.technologyreview.com/blog/mimssbits/25981/

I had the presentation a few days ago, I did my best to advertise it, even going to each class and giving a summary of what FAH was and asking them to attend my presentation. Despite my best efforts less than 30 people came, and out of those only 8 decided to donate to the project. I sent them e-mails asking them what processors they have (and giving instructions on how to find that out) and an installation guide in Portuguese I made, but none of them have replied yet. Sigh.

The vice-principal sent me an e-mail half an hour prior to the presentation saying a sudden appointment came up and apologizing that she'd be unable to come. However the Biochemistry teacher I invited and some of his doctorate students did come.

Many people praised me on the presentation, and that they could tell I am really excited about FAH and I believe in it. However I should have focused less on the technical aspects, I didn't know most people there had no clue at all about processors and didn't even know what GHz means on a processor -.-. They also said I should have started some philosophical discussions such as why it is important to donate. I didn't judge that necessary because "helping scientists find a cure to diseases" isn't in my mind something you can have many logical arguments against :/ I didn't have enough time for that anyway, I only had 1 hour.

Thanks for the support guys

How big is your family, how many brothers and sisters?

Two parents, one sister. My parents both have an i5 2500k processor, but I don't run FAH on their computers because they like silence and FAH would make their computers too noisy.

You asked how to get people involved... Ask people how big their family is... Then ask them if they know how likely a person is to get cancer. 1 in 100? 1 in 10? It is actually about 1 in 3. It is fatal in 1 of 4.

Now look back at your family of 4 people. Yes, a little disturbing to think about it... But you can actually run a simple program on your computer to look for a cancer treatment, even a cure is possible.

Then talk about folding at home.

This is one of many ways to engage people rather than just talk at them.