I've been running DC projects in my spare cycles for years, starting with the SETI project, the precursor to BOINC.
Quote:
I found this recent comment on-line:
Quote:
People think they are doing a great thing by using BOINC, however:
1) They are brute force projects. Instead of understanding the underlying processes of what they are attempting to solve, instead they generally just understand the top level, and test against it for candidates that might work.. Messy work.
2) People don't realize that it actually costs quite a lot to run Boinc. If they donated the money they spent on electricity and the lower life span caused by running their computers 24/7, things would probably move a lot quicker. Its more efficient to create a cluster of servers in a central location which are dedicated towards the tasks.
3) CPU's speed to power consumption ratio is increasing rapidly. Wait 3 or 4 years and it will take only a handful of desktop computers
Despite the good intentions of BOINC, I think resources would be better spent elsewhere. Some examples may be:
1) On efficient renewable energy technologies. Then we can run all the research we want with less cost, and less problems.
2) Funding low-level research instead! Brute forcing a bunch of molecules to see if it solves Alzheimer's is one thing. But understanding it completely may prevent it occurring in the first place. Furthermore, you are less likely to end up with thousands of drug candidates which aren't targeted at only Alzheimer's and so cause other (possibly unknown) side effects.
|
|
This sounds like somebody whining because they got a smaller piece of the pie. I don't think these critics understand how distributed computing works.
Brute force? Umm, isn't that why you toss problems at the world's most powerful computers? Only instead of using VERY expensive and limited mainframe and super computers, they rely on millions of smaller computers running parallel processing that cost them nothing. In the case of BOINC, note the B stands for Berkeley - a tax payer funded university. How is saving tax payer money bad?
Quote:
|
instead they generally just understand the top level
|
This is simply an ignorant opinion about ALL the researchers, the people, and not about the data being collected and analyzed. So apparently, this guy does not like how a project is being run, so therefore he assumes all DC projects are run by idiots. Not the case.
Quote:
|
lower life span caused by running their computers 24/7
|
This is simply WRONG! There is no evidence to suggest running a system 24/7 shortens the life span. In fact, many long time hardware technicians (including yours truly) personally feel that running electronics in a constant operating environment MINIMIZES damage caused by the different expansion/contraction rates of matter in mated dissimilar materials (as used virtually all semi-conductive devices) during frequent heat up and cool down cycles. Over time, these stresses can and often do form microfractures, fatigue, premature aging of the device, and eventual failure. I note this is one of the primary reasons sensitive electronics equipment is maintained and operated in environmentally controlled facilities - so there are no wide swings in ambient temperature and humidity conditions. A system that runs 24/7 stays at a fairly constant temperature. In environments where the ambient temperatures of the facility are low - meaning a wide swing from cold to fully heated - condensation can actually become an issue in some cases - not good.
Quote:
|
Its more efficient to create a cluster of servers in a central location which are dedicated towards the tasks.
|
This is ignorance talking. Who is going to pay for this cluster of high-end servers, the facility, the power to run these computers? Who is going to pay for the IT specialist to maintain these servers, the OS and software running on them? Total efficiency of a project is measured by MUCH more than just electricity consumed.
Quote:
|
Wait 3 or 4 years and it will take only a handful of desktop computers
|
Yeah, right.
For the regular home computer user, there is no monetary incentive to run any DC project. But it can be fun, there is often a active community of crunchers that talk and share experiences and have friendly rivalries. It makes good use of older hardware that would otherwise sit in a closet or be tossed in a landfill.
For me, I keep my systems on 24/4 for a couple reasons - with security being one of the top. At night, while I am asleep, my systems scan themselves for malware.
For me, searching for cures for cancer by donating idle time on a computer gives me a warm fuzzy, while not hurting anyone.
