ZhengHe
03-27-2008, 09:22 PM
About a month ago I was greeted with the announcement that 1,000,000 PS3s (http://blog.wired.com/games/2008/02/foldinghome-rea.html) had officially joined the Folding@Home project. An impressive milestone when you consider that 2 years ago the number stood at 50,000. They now comprise 75% of all folding machines and have allowed the project to process at upwards of 1 Petaflop (http://www.theinquirer.net/en/inquirer/news/2007/09/20/ps3s-boost-foldinghome-to-a-petaflop). While the addition of so many new members is certainly a boon to the projects processing power and continued growth there was something that puzzled me about that much growth. We had never seen anything like this with regular computers. Even the SMP client couldn’t claim such a tremendous change and the GPU client was a small blip in comparison. So I decided to do some digging.
Folding@Home is a project almost obsessed with its own statistics. You can see it from the amount of traffic heading toward its own stats page along with the high number of 3rd part sites doing exactly the same thing. At last count the number stands somewhere around 25 worldwide and it is still growing. Even more incredible, the PS3 has only sold 10.5 million units (http://online.wsj.com/article/SB120526246285327923.html?mod=googlenews_wsj). That means that almost 10 percent of all consoles, which were bought in the last few years now fold. How could they have achieved such a sizeable following in such a short span of time? Was it easier access to an embedded Folding@Home console? Perhaps Sony was just more committed to distributed computing in general? Did the cell processor simply present a superior folding machine, which translated to greater usage?
To the issue of the Folding@Home embedded client it is a relatively recent addition. This is important, since there did exist a less direct Linux based beforehand along with a few other lesser known alternatives. So some people were folding on their PS3’s before Sony added the client with firmware update 1.60 (http://www.beyond3d.com/content/news/128). This is important, because it outlines a tangible base for the company’s future push into distributed computing. A group of enterprising users, even then realized the obvious advantages of using a PS3 as a folding client. Through doing this they illuminated this option to Sony who would eventually take hold of it as a legitimate marketing strategy and eventually collaborate with Stanford on their own “official” client (http://www.gamasutra.com/php-bin/news_index.php?story=10612).
It was that collaboration, which eventually lead to the unprecedented growth of the console as a folding platform. This also lent Sony a much needed PR boost. Suddenly the company was cast as a friend to distributed computing. A real message of philanthropic virtue in the eyes of many already associated with the project. While their motivations may have been less then charitable the community took hold of the client very quickly. This can be compared to Microsoft initial stab at distributed computing given the console’s obvious shortcomings and late arrival in the dialogue. Sony and F@H’s initial collaboration eventually gave way too many other similar efforts. This includes the similar efforts by Seti@Home along with the Guarav Khanna’s creation (http://setiathome.berkeley.edu/forum_thread.php?id=39910.) of a client making use of PPE for Einstein@Home (http://einstein.phys.uwm.edu/). This is the same man that created a network of 16 PS3s (http://www.reghardware.co.uk/2008/02/28/ps3s_put_to_use_simulating_blackholes/) to simulate black hole collisions. His efforts gave way to a similar ones by SIMAP (http://boinc.bio.wzw.tum.de/boincsimap/) along with a joint effort through Berkeley’s BOINC (http://boinc.berkeley.edu/) project to make a client available to almost all the major projects. This has the potential for a second explosion in use on a wider scale.
As to the contention that the PS3 was simply a superior folding platform, which made exponential growth inevitable, that has a few holes. Although at present the PS3 is arguably the best folding platform for the money it is doubtful that will be the case for long. With exponential growth in the multi-core processor market cheaper prices should soon outpace any advantages the PS3 may have. Still as a console the PS3 is more widely accessible to the general public, which decreases any advantages a price drop may have for a time. It is also relatively compact and easy to configure so that it may be a more reasonable option for those with less expertise.
What becomes patently obvious as to why the PS3 has become so popular, it was simply in the right place at the right time. It chose to embrace a market of philanthropic, technology minded consumers who were willing to spend considerable sums for a good cause. When their expertise was needed they were willing to put form time and expertise toward the creation of a vastly improved client. Most importantly, their product was up to the task. By spending considerable sums on making a more stable piece of hardware they set the stage for its use in mission critical activities, like distributed computing.
As any major trend evolves competitors eventually want in on the action. Distributed computing is no exception. With the exponential growth in its usage and application suddenly the term has finally come into its own. Microsoft and Intel just recently promised $20 million (http://www.informationweek.com/windows/showArticle.jhtml?articleID=206904517) toward the creation of 2 parallel computing facilities in Berkeley and the University of Illinois-Urbana. With these facilities they intend to test and refine their own distributed computing projects while finding research into ever more powerful processors more equipped to the task. Sony has also looked into the commercial implications (http://www.dmwmedia.com/news/2007/04/11/sony-exploring-commercial-use-of-ps3-distributed-computing) of distributed computing. Other universities and companies have also gotten on the bandwagon. Many have started their own teams and some have even developed projects of their own.
So what is the paradox of the PS3? The paradox is how a company who originally designed their platform for the gaming elite would eventually open up distributed computing to the masses and ostensibly help to jumpstart a revolution in it’s use. The effects of which will be felt for years to come.
Folding@Home is a project almost obsessed with its own statistics. You can see it from the amount of traffic heading toward its own stats page along with the high number of 3rd part sites doing exactly the same thing. At last count the number stands somewhere around 25 worldwide and it is still growing. Even more incredible, the PS3 has only sold 10.5 million units (http://online.wsj.com/article/SB120526246285327923.html?mod=googlenews_wsj). That means that almost 10 percent of all consoles, which were bought in the last few years now fold. How could they have achieved such a sizeable following in such a short span of time? Was it easier access to an embedded Folding@Home console? Perhaps Sony was just more committed to distributed computing in general? Did the cell processor simply present a superior folding machine, which translated to greater usage?
To the issue of the Folding@Home embedded client it is a relatively recent addition. This is important, since there did exist a less direct Linux based beforehand along with a few other lesser known alternatives. So some people were folding on their PS3’s before Sony added the client with firmware update 1.60 (http://www.beyond3d.com/content/news/128). This is important, because it outlines a tangible base for the company’s future push into distributed computing. A group of enterprising users, even then realized the obvious advantages of using a PS3 as a folding client. Through doing this they illuminated this option to Sony who would eventually take hold of it as a legitimate marketing strategy and eventually collaborate with Stanford on their own “official” client (http://www.gamasutra.com/php-bin/news_index.php?story=10612).
It was that collaboration, which eventually lead to the unprecedented growth of the console as a folding platform. This also lent Sony a much needed PR boost. Suddenly the company was cast as a friend to distributed computing. A real message of philanthropic virtue in the eyes of many already associated with the project. While their motivations may have been less then charitable the community took hold of the client very quickly. This can be compared to Microsoft initial stab at distributed computing given the console’s obvious shortcomings and late arrival in the dialogue. Sony and F@H’s initial collaboration eventually gave way too many other similar efforts. This includes the similar efforts by Seti@Home along with the Guarav Khanna’s creation (http://setiathome.berkeley.edu/forum_thread.php?id=39910.) of a client making use of PPE for Einstein@Home (http://einstein.phys.uwm.edu/). This is the same man that created a network of 16 PS3s (http://www.reghardware.co.uk/2008/02/28/ps3s_put_to_use_simulating_blackholes/) to simulate black hole collisions. His efforts gave way to a similar ones by SIMAP (http://boinc.bio.wzw.tum.de/boincsimap/) along with a joint effort through Berkeley’s BOINC (http://boinc.berkeley.edu/) project to make a client available to almost all the major projects. This has the potential for a second explosion in use on a wider scale.
As to the contention that the PS3 was simply a superior folding platform, which made exponential growth inevitable, that has a few holes. Although at present the PS3 is arguably the best folding platform for the money it is doubtful that will be the case for long. With exponential growth in the multi-core processor market cheaper prices should soon outpace any advantages the PS3 may have. Still as a console the PS3 is more widely accessible to the general public, which decreases any advantages a price drop may have for a time. It is also relatively compact and easy to configure so that it may be a more reasonable option for those with less expertise.
What becomes patently obvious as to why the PS3 has become so popular, it was simply in the right place at the right time. It chose to embrace a market of philanthropic, technology minded consumers who were willing to spend considerable sums for a good cause. When their expertise was needed they were willing to put form time and expertise toward the creation of a vastly improved client. Most importantly, their product was up to the task. By spending considerable sums on making a more stable piece of hardware they set the stage for its use in mission critical activities, like distributed computing.
As any major trend evolves competitors eventually want in on the action. Distributed computing is no exception. With the exponential growth in its usage and application suddenly the term has finally come into its own. Microsoft and Intel just recently promised $20 million (http://www.informationweek.com/windows/showArticle.jhtml?articleID=206904517) toward the creation of 2 parallel computing facilities in Berkeley and the University of Illinois-Urbana. With these facilities they intend to test and refine their own distributed computing projects while finding research into ever more powerful processors more equipped to the task. Sony has also looked into the commercial implications (http://www.dmwmedia.com/news/2007/04/11/sony-exploring-commercial-use-of-ps3-distributed-computing) of distributed computing. Other universities and companies have also gotten on the bandwagon. Many have started their own teams and some have even developed projects of their own.
So what is the paradox of the PS3? The paradox is how a company who originally designed their platform for the gaming elite would eventually open up distributed computing to the masses and ostensibly help to jumpstart a revolution in it’s use. The effects of which will be felt for years to come.