Message boards :
Number crunching :
New Twist For "Aborted by project" Issue ?
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · Next
Author | Message |
---|---|
zombie67 [MM] Send message Joined: 22 Apr 04 Posts: 758 Credit: 27,771,894 RAC: 0 |
Here's a thought: If a DC project cannot keep up with the machines of their contributors, then perhaps they shouldn't be DCs. At that point, their projects just don't require the crunching power offered by DC. Just go buy a few machines and do the work in-house. Dublin, California Team: SETI.USA |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13769 Credit: 208,696,464 RAC: 304 |
The figure of $538,000 was mentioned for this year, Keep in mind that figure was to keep things going as they are now & help with Astropulse etc development. More would be required to increase the system's capabilities over what they are now. Also notice the "this year" part. Even with the "one off" expense of new hardware, the ongoing costs would be very significant. Grant Darwin NT |
kittyman Send message Joined: 9 Jul 00 Posts: 51469 Credit: 1,018,363,574 RAC: 1,004 |
Here's a thought: But the whole point of Boinc is to see how much computing power can be harnessed to work on a single project (or many projects, as Boinc has evolved). "Freedom is just Chaos, with better lighting." Alan Dean Foster |
zombie67 [MM] Send message Joined: 22 Apr 04 Posts: 758 Credit: 27,771,894 RAC: 0 |
What are you talking about? "reasonable expectation" was a threshold you created. Perhaps I should have put it in quotes to make it clear. I am just trying to get you to define it. Dublin, California Team: SETI.USA |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13769 Credit: 208,696,464 RAC: 304 |
Here's a thought: Why? Their projects still require the crunching power offered by DC, they just don't have the funds to increase their sytems to meet the demand (and probably have even less than enough funds to do the crunching required with their own hardware if they were to attempt to buy it), people get bored etc & go elsewhere. Good old supply & demand at work. Grant Darwin NT |
zombie67 [MM] Send message Joined: 22 Apr 04 Posts: 758 Credit: 27,771,894 RAC: 0 |
But the whole point of Boinc is to see how much computing power can be harnessed to work on a single project (or many projects, as Boinc has evolved). I thought it was to harness idle computer cycles.... Dublin, California Team: SETI.USA |
Brian Silvers Send message Joined: 11 Jun 99 Posts: 1681 Credit: 492,052 RAC: 0 |
My quaddy averages about 52m or so on a full flavored Wu...with Chicken. That comes out to about 108 WUs/24 hours. With a quota of 400 for the rig, I'll never reach the WU quota limit. I'm even willing to be generous and say that you could probably hit 200/day fairly consistently. Even then, it is still half of the limit. All increasing the quota does is let you fill faster. You're still limited by the RDCF and the number of seconds of work it takes to come up to the cache level (assuming no shenanigans). The comparison of filling a gallon jug full of water via a faucet-end water filter (quota limit) or straight from the tap (higher quota) is a good example. The water filter method will take longer to fill, but it will fill. |
kittyman Send message Joined: 9 Jul 00 Posts: 51469 Credit: 1,018,363,574 RAC: 1,004 |
Here's a thought: But I think the concept of DC is not to see if you can get people to contribute computing power to a project, it's to see if it's possible to harness enough computers together to work on a task impossible to accomplish without investing millions in a supercomputer to do the same work. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Brian Silvers Send message Joined: 11 Jun 99 Posts: 1681 Credit: 492,052 RAC: 0 |
I never brought up the reasonable expectations angle until what I said there. Your reply gives the appearance that you thought you were replying to Grant. You were not. You were replying to me. The question I posed is still a legit question, btw... |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13769 Credit: 208,696,464 RAC: 304 |
I thought it was to harness idle computer cycles.... That was a selling point, but some people get carried away & buy/build systems just for crunching. The idea of Distributed Computing was as mentioned in another post- to use lots of computers to do work that only a (or several) supercomputers could otherwise do. The idea of BOINC was to give DC projects a common interface & control for users so they could easily run more than one project at a time. Grant Darwin NT |
Brian Silvers Send message Joined: 11 Jun 99 Posts: 1681 Credit: 492,052 RAC: 0 |
Here's a thought: Grant, you remember the change from Classic 2.x to 3.x, right? That was the response of the project to harness the crunching power that had become available, by increasing the resolution of the search. The move to the BOINC "SETI-Enhanced" was a similar move. @Zombie - I said it before and I'll say it again, I understand what you're saying. From looking at your graph on BOINCStats, it is clear that the majority of your crunching time here came after BOINC SETI-Enhanced was released, thus you've never seen the intentional braking (yes, as in applying brakes, not as in BREAKING) that is done to better utilize the horsepower that has become available since the app was first deployed... |
zombie67 [MM] Send message Joined: 22 Apr 04 Posts: 758 Credit: 27,771,894 RAC: 0 |
The real question is, do you consider all your expectations to be "reasonable by default" simply because they are your expectations and disregarding any and all monetary contributions? Sorry. You're right. I confused you two there. The question I posed is still a legit question, btw... I consider my expectations to be reasonable because the are. And they are not disregarding any and all monetary contributions. Dublin, California Team: SETI.USA |
kittyman Send message Joined: 9 Jul 00 Posts: 51469 Credit: 1,018,363,574 RAC: 1,004 |
I thought it was to harness idle computer cycles.... Interesting question...which came first, the chicken or the egg, or the Seti or the Chicken? Was the initial concept of Boinc to make it easier for users to contribute their computer time to a project, or to see how much computer power a project could amass toward a single goal? "Freedom is just Chaos, with better lighting." Alan Dean Foster |
zombie67 [MM] Send message Joined: 22 Apr 04 Posts: 758 Credit: 27,771,894 RAC: 0 |
@Zombie - I said it before and I'll say it again, I understand what you're saying. From looking at your graph on BOINCStats, it is clear that the majority of your crunching time here came after BOINC SETI-Enhanced was released, thus you've never seen the intentional braking (yes, as in applying brakes, not as in BREAKING) that is done to better utilize the horsepower that has become available since the app was first deployed... You keep referencing SETI Classic. It sounds like you have some luggage you're bringing along. I'm not seeing the relevance (for this discussion). Dublin, California Team: SETI.USA |
zombie67 [MM] Send message Joined: 22 Apr 04 Posts: 758 Credit: 27,771,894 RAC: 0 |
|
Brian Silvers Send message Joined: 11 Jun 99 Posts: 1681 Credit: 492,052 RAC: 0 |
@Zombie - I said it before and I'll say it again, I understand what you're saying. From looking at your graph on BOINCStats, it is clear that the majority of your crunching time here came after BOINC SETI-Enhanced was released, thus you've never seen the intentional braking (yes, as in applying brakes, not as in BREAKING) that is done to better utilize the horsepower that has become available since the app was first deployed... To understand where we are, and really where we're going, you need to know what happened in the past. If we were trying to run SETI Enhanced on P3 and early P4 machines that were prevalent back in 2003/2004, people would be complaining about how slow it performed. To be able to attract people to work on the project, it has to perform somewhat reasonably. To do that, certain sacrifices are made in the resolution of the science application. When processor performance has increased, like it had in 2006 when the prevalent platforms were HT-enabled P4 and AMD Athlon 64 machines, Enhanced was released which increased the resolution of the search. SETI has taken the path of incrementally increasing the resolution rather than building an ultra hi-res app up front so as to bring more people on board. Is that needed now (to bring more people on board)? Don't know. Depends on how Astropulse performs (thanks to MajorKong for pointing out my mistaken impression of multibeam, btw)... Fundamentally what I think you want, if you are really concerned about "future proofing" and not some short-term personal goal, is a model system like CPDN that has a highly complex and demanding science application that rewards with trickles. I don't know if a non-alpha / non-beta project for SETI could be made to be that intense. The alpha / beta apps have debugging which slows them down. In any case, upping the amount of work you can pull in a single day won't accomplish what you're thinking it will for you. It will help you recover faster from a lengthy complete outage, but with the caveat of having the work there to pull in the first place. It won't really help you pre-outage or to survive a lengthy outage. Only increasing the total number of days of cache can help that, but that has known and admitted negative impacts for the project and potential negative impacts (work availability issues) for the rest of us who don't run really large caches... Brian |
Brian Silvers Send message Joined: 11 Jun 99 Posts: 1681 Credit: 492,052 RAC: 0 |
Fundamentally, you just said that your expectations are reasonable because they are your expectations. I don't think that is very objective... |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13769 Credit: 208,696,464 RAC: 304 |
Grant, you remember the change from Classic 2.x to 3.x, right? That was the response of the project to harness the crunching power that had become available, by increasing the resolution of the search. The move to the BOINC "SETI-Enhanced" was a similar move. Yep. But you can only increase the resolution of the search in the data so much without making any results meaningless. I expect Multibeam & Astropulse when they come online will result in much longer crunch times & so a big reduction in the amount of data being processed per hour. But hardware will improve & we'll be back where we are now in not that much time from now. Grant Darwin NT |
Brian Silvers Send message Joined: 11 Jun 99 Posts: 1681 Credit: 492,052 RAC: 0 |
OK, my lack of sleep is really kicking in now... Due to noise enhancement or due to having sufficient samples so that any more samples just add redundant data that could be extrapolated (ala Integral Calculus), or something else entirely? |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
Brian Silvers: ...multibeam will increase processing time, reducing even further the number of results that can be processed in any "average" day. This will mean that 400/day will still be able to fill up a cache quicker. Actually, Multibeam will decrease processing time. Because of the difference in beam width, ar=0.441 Multibeam WUs like those we're running in SETI Beta now are processed like ar=0.732 Line feed WUs. In addition, one of the techniques used with the ALFA receivers is a basket weave scan which Kevin Douglas described as "nodding on the meridian". The nodding rate is such that those observations will produce VHAR WUs in abundance. Joe |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.