Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 133
Posts: 133   Pages: 14   [ Previous Page | 5 6 7 8 9 10 11 12 13 14 | Next Page ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 20943 times and has 132 replies Next Thread
Jim1348
Veteran Cruncher
USA
Joined: Jul 13, 2009
Post Count: 1066
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WCGrid Cost-Benefit Analysis

The cost part is easy. The benefit difficult, if not impossible. They don't know whether anything useful will be found, and if so, how widely it can be adopted. Maybe it will be a world-changer. Maybe it will never be heard of again.

I worked at a very high-tech company. That is the way it is. And the more basic the research, the greater the uncertainties.
[Apr 10, 2020 9:43:43 PM]   Link   Report threatening or abusive post: please login first  Go to top 
wcgridmember
Advanced Cruncher
Joined: Mar 30, 2005
Post Count: 110
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WCGrid Cost-Benefit Analysis

R.West, you make a very good point. If/when tasks run more cheaply not on your computer, but on one of those cloud services, then it makes more sense that you donate the money you would spend on electricity to pay the cloud computing service for crunching WCGrid. This might be even more dramatic for places such as Europe where energy is typically twice the price as in the US.

jackielan2000, you said "The electricity would have been wasted anyway.", but the way computers work implies the computer will use more energy if it is crunching while you do something than if you were doing something without crunching. Yes, if you are using it for heating then you are not incurring in greater costs or environmental impact than if you used a heater. Please also note that if you run it for relative short amount of time, you might be wasting lots of energy because the checkpoints might not be sufficiently frequent. Imagine you crunch only for two hours, but the next checkpoint would only come in 2 and a half hours; all the crunching during that time will be wasted.
----------------------------------------
[Edit 1 times, last edit by wcgridmember at Apr 10, 2020 10:01:08 PM]
[Apr 10, 2020 10:00:19 PM]   Link   Report threatening or abusive post: please login first  Go to top 
wcgridmember
Advanced Cruncher
Joined: Mar 30, 2005
Post Count: 110
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WCGrid Cost-Benefit Analysis

Jim1348, all help on the estimation of the costs is welcome. ;-)
[Apr 10, 2020 10:04:22 PM]   Link   Report threatening or abusive post: please login first  Go to top 
wcgridmember
Advanced Cruncher
Joined: Mar 30, 2005
Post Count: 110
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WCGrid Cost-Benefit Analysis

Sheridon, could you make that into a formula, please? At this point I would already be quite happy if I could get a very conservative estimation (meaning, that it would assume everyone is being the most efficient).
----------------------------------------
[Edit 1 times, last edit by wcgridmember at Apr 10, 2020 10:08:12 PM]
[Apr 10, 2020 10:07:28 PM]   Link   Report threatening or abusive post: please login first  Go to top 
wcgridmember
Advanced Cruncher
Joined: Mar 30, 2005
Post Count: 110
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WCGrid Cost-Benefit Analysis

Also, let's simplify and ignore the hardware costs. Let's just take the electricity costs in consideration for now.
[Apr 10, 2020 10:10:47 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Jim1348
Veteran Cruncher
USA
Joined: Jul 13, 2009
Post Count: 1066
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WCGrid Cost-Benefit Analysis

Jim1348, all help on the estimation of the costs is welcome. ;-)

OK, assume an equipment cost of around $1000 per machine. Running costs at around 15 cents per kw/hour will be about $0.36 per day (with about 100 watts devoted to the WCG application). Multiply that by the expected run time for a given project. That will give you the total capital and running costs per machine. You can refine that by separating desktops from laptops if you want to.

Multiply that by the number of machines; you can probably get an estimate of that from the statistics somewhere.

That will get you within a factor of two or three of the costs, which is a whole lot better than you will get with the benefits.
[Apr 10, 2020 10:35:21 PM]   Link   Report threatening or abusive post: please login first  Go to top 
jackielan2000
Advanced Cruncher
China
Joined: Dec 31, 2005
Post Count: 115
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WCGrid Cost-Benefit Analysis

jackielan2000, you said "The electricity would have been wasted anyway.", but the way computers work implies the computer will use more energy if it is crunching while you do something than if you were doing something without crunching. Yes, if you are using it for heating then you are not incurring in greater costs or environmental impact than if you used a heater. Please also note that if you run it for relative short amount of time, you might be wasting lots of energy because the checkpoints might not be sufficiently frequent. Imagine you crunch only for two hours, but the next checkpoint would only come in 2 and a half hours; all the crunching during that time will be wasted.


Well, I usually left the computer on after throwing some commands to my online game. Then I go eating, having shower and watching TV etc while the PC heating the room (a very slow process though laughing ). After these routins, I play the game more and go to sleep with the PC on for the night.

Old PC + WCG = Heater laughing

Of course, I don't do this in the summer. AC costs too much. Only Android devices.
----------------------------------------
AMD Athlon64X2 5400+ 2.8G | 2c
MT6735 1.4G | 4c
Helio G85 1.8G |8c
Allwinner H2 1G | 4c
SnapDragon 810 2.1G | 8c
SnapDragon 801 2.5G | 4c
[Apr 11, 2020 9:14:31 PM]   Link   Report threatening or abusive post: please login first  Go to top 
jackielan2000
Advanced Cruncher
China
Joined: Dec 31, 2005
Post Count: 115
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WCGrid Cost-Benefit Analysis

I can not believe those PIII's are getting much done, but more power to you if you wish to use them.
Cheers


The PC, with dual PIIIs, used to be a server in the old days. It can generate a lot of heat if both CPUs were utilized w/ 100% CPU time. So, I only use it during winter. Why do I need to use a heater when this thing can do the same and at the same time do something good for science? tongue Anyway, PIII is not that slow though, almost like a 2-core Android phone. Roughly 2 WUs every 12 hours.
----------------------------------------
AMD Athlon64X2 5400+ 2.8G | 2c
MT6735 1.4G | 4c
Helio G85 1.8G |8c
Allwinner H2 1G | 4c
SnapDragon 810 2.1G | 8c
SnapDragon 801 2.5G | 4c
[Apr 11, 2020 9:37:20 PM]   Link   Report threatening or abusive post: please login first  Go to top 
wcgridmember
Advanced Cruncher
Joined: Mar 30, 2005
Post Count: 110
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WCGrid Cost-Benefit Analysis

From:
$0.36 daily cost for running the machine (Jim1348)
Average Run Time Per Calendar Day: 352 years per day (BOINC stats)

$0.36 price -> 1 day runtime
X -> 352 years * 365 days (Run Time in days per Calendar Day)

X = $46 252.8 Cost (electricity costs from all machines per Calendar Day)

Total Yearly Costs = $46 252.8 * 365 days = $16 882 272

Now a few things that could be inflating this estimation: devices that work on less than 100W, price lower in some areas;
things that could be deflating this estimation: devices that work on more than 100W, price higher in some areas.

These are direct costs, of course. I'm not taking into account the neutral impact on costs from WCGrid that crunching takes when it's used for heating that would be spent anyway. If this was taken into account, then the relevant costs would be lower. But then, it could be the case that people need more A/C when crunching in the summer and this would make relevant costs higher.

Any thoughts? Is there anything from this estimation that should be fixed?

UPDATE: These calculations assume one computer with only one core. For a more realistic estimation, assume: $0.36 price -> 6 day runtime (same as assuming an average of 6 cores per computer) and yesterday's runtime: 551 years.

Total yearly costs: 551 * 365^2 * 0.36 / 6 = $4 400 000
----------------------------------------
[Edit 4 times, last edit by wcgridmember at Apr 20, 2020 8:25:50 AM]
[Apr 11, 2020 9:48:37 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Jim1348
Veteran Cruncher
USA
Joined: Jul 13, 2009
Post Count: 1066
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WCGrid Cost-Benefit Analysis

I was estimating the power (100 watts) for the entire CPU and supporting motherboard, which would typically be from 8 to 16 cores these days, so just assume 10 cores as an average. That cuts it down considerably.
----------------------------------------
[Edit 1 times, last edit by Jim1348 at Apr 11, 2020 10:07:25 PM]
[Apr 11, 2020 10:05:38 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 133   Pages: 14   [ Previous Page | 5 6 7 8 9 10 11 12 13 14 | Next Page ]
[ Jump to Last Post ]
Post new Thread