Folding Forum

If F@H could exist as a distributed computing project in 1992, what would it be like?
I was thinking about this and made a few conclusions:

- A math coprocessor (for CPUs without a built-in FPU) would've been a must-have since folding is mathematically intensive.
- Work units would have to be much simpler and smaller in file size.
- F@H would've been programmed in 16-bits for MS-DOS. There might have been a high-performance client with some 32-bit code (for Windows 3.1 enhanced mode users ). Don't know for the non-Microsoft OSes back then. Question: would the 640 KB conventional memory limit have been a stumbling block to F@H?
- There would have been fewer participants, because computers and Internet access were less common (and more expensive) back then.

The above points are just my 2 cents. Feel free to imagine and add what you think F@H would have been like if it could exist 20 years ago.

Just think of how long it would have taken to send and receive even a very small WU back then I can remember it taking a couple of min. just to download a small web page.

FAH is about 12 years old, so 20 years is stretching reality, but as little as 10 years ago, a completed WU was worth one point.

@Grandpa_01:
I remember downloading and uploading on 56kb dial-up.

As a child I remember my dad plunking down hundreds of dollars, when that was actually a lot of money, to get his Tandy 300 baud modem. It would spit out the characters like you see in old movies.

keithw1975 wrote:As a child I remember my dad plunking down hundreds of dollars, when that was actually a lot of money, to get his Tandy 300 baud modem. It would spit out the characters like you see in old movies.

I still remember having an 28.8k modem as a kid in the mid 90s. The sounds it made when connecting are unmistakable and unforgettable. There were no flat rates (to my knowledge) in those days either - you paid on a traffic basis, and I had to ask my parents for permission whenever I wanted to go online.

I remember how exhilarating it was to get ISDN and all of a sudden be able to surf with speeds of up to 12KB a second...it was pure awesomeness. Not to mention what a story it was when CPUs breached the 500Mhz barrier for the first time, and then 1Ghz...it was unbelievable.

These days when I can get a 100Mbit connection for lunch money, and an extraordinarily powerful CPU for very handy prices, it is like the old charm has just dissapeared. It is just business.

keithw1975 wrote:As a child I remember my dad plunking down hundreds of dollars, when that was actually a lot of money, to get his Tandy 300 baud modem. It would spit out the characters like you see in old movies.

I guess it's showing my age, but I got my first 300 baud modem for our Franklin Ace 1000 (Apple ][ Clone) in 1984 or so. When we got our first Apple IIgs around 1986 or so, the modem was a "must have" device. We had one gs at home, and our oldest son took another to UC Berkeley when he started there in 1988 - it got him through an Honors EECS degree. We went from 300 baud to 1200 to 2400 to 28.8 to 56K, before we finally got a cable modem in around 2000.

It's hard to realize how relatively fast the technology moved, and how connected we all are now.

keithw1975 wrote:As a child I remember my dad plunking down hundreds of dollars, when that was actually a lot of money, to get his Tandy 300 baud modem. It would spit out the characters like you see in old movies.

Ya, I can remember stuff like that in old movies, like "The Net" and "Hackers"

I doubt F@h would have existed in 1992. Distributed.net was launched in 1997 and it seems to me that any earlier wouldn't have really been practical. It may not have been worth it to utilize weak home computers, and may have been easier and more cost effective to use a supercomputer. Remember that our WUs are time sensitive and have a serial nature about them. F@h's first couple of papers explain why 2000 was an appropriate year to launch.

Speaking of distributed dot net, I was doing dnet in high school in around 2002, and I was following the mailing list, and one person asked if dnet could use the GPU, and somebody replied that it would be impossible, that a GPU is only for graphics, that the calculations are completely different, and so on, as if it were a ridiculous idea.

Haplo wrote:Speaking of distributed dot net, I was doing dnet in high school in around 2002, and I was following the mailing list, and one person asked if dnet could use the GPU, and somebody replied that it would be impossible, that a GPU is only for graphics, that the calculations are completely different, and so on, as if it were a ridiculous idea.

And they were correct, until 2006 when nvidia came up with the programmable shaders on the G80 gpu's, nowadays more and more parrallel processing is being ported to the GPU, especially for scientific research

And FAH was the first distributed computing project to use GPUs, but not on G80s.

7im wrote:And FAH was the first distributed computing project to use GPUs, but not on G80s.

True, but i wasn't around for GPU1 and i'm not sure how succesful it was. Besides, i did say programmable shaders

My first modem was a 300 bauds acoustic coupler on a Tandy CoCo 1 How has times changed with my 30 MegaBit Cable connection LOL!

It's a bit scary to think how powerful computers will be in 20 years if Moore's law can hold true. We are getting to a point were shrinking CPU sizes is getting harder but a breakthrough might send us into an other direction to increase CPU power.

Do not fear the future - any CPU advancements will be negated by some sloppy programming. What ever will happen you find your computer to work sluggish enough to get a shiny new model (every couple of years at least)...

Thanks for the informative replies.
It seems that it's not just computing power, but the linking of computers over the Internet that's necessary for a distributed computing project.