Why not increase credits by 25% since electricity has gone up much more than thatin the last couple years? I've seen my retirement savings lose money the last year, my salary hasn't kept up with inflation for 3 years, and some useless project searching for little green men is concerned that people won't participate in their science fiction project unless credits are equal. So what? My disposable income has been reduced and now the credit I used to get for crunching is going down too? It isn't like I can pay my electric bill with cobblestones, but please give me some incentive to keep my six Pentium III's running! Longer WUs and 25% less credit is not the way.
Why not increase credits by 25% since electricity has gone up much more than thatin the last couple years? I've seen my retirement savings lose money the last year, my salary hasn't kept up with inflation for 3 years, and some useless project searching for little green men is concerned that people won't participate in their science fiction project unless credits are equal. So what? My disposable income has been reduced and now the credit I used to get for crunching is going down too? It isn't like I can pay my electric bill with cobblestones, but please give me some incentive to keep my six Pentium III's running! Longer WUs and 25% less credit is not the way.
Well you pretty much told yourself which project is a waste of electricity Slicker (seti) but I see you aren't running that now.
But you do have those PIII's doing lots of work........maybe too much as far as projects.
I remember when I started doing classic seti's back in 2000 on my PIII 500 which died many years ago.
I have done quite a few of the new S5R4's and the claimed credit does look bad but the granted credit isn't all that bad and is less than a 25% difference.
Instead of the previous 237 credits I get between 163 to 192's and most between 185 and 192 credits. (I see 6 of them are 192 credits and 7 are around 185)
The light bill is more where I live too so I started turning off all my monitors now instead of using the blank screensaver like I had been since 2000.
It isn't like I can pay my electric bill with cobblestones, but please give me some incentive to keep my six Pentium III's running!
If your Pentium III hosts are running mostly or entirely for BOINC, you might consider building a single new host. I think you'll find that for not very many hundred dollars at all you could considerably exceed the credit output of those six, vastly undercut their power consumption, and pay the initial cost back with not very many months of power bill savings.
If you don't like the thought of building your own, this is still true on the market.
I still run one 930 MHz Coppermine host. I currently am sketching intentions to replace it late this year with a Nehalem bottom end host. I think I can bring it in at about equal power, and get a little over twenty times the Einstein (or SETI) output, for well under $1000. Being a little less greedy in processor can push that initial outlay down a lot, and still come in over 6x.
Hint: within a given processor range, at default settings the lower frequency processors give a big power economy--you get linear power reduction from lower frequency, but a very big bonus from lower voltage. The specs are not much help on this, as TDP is only a weak indicator of actual power consumed by model.
It isn't like I can pay my electric bill with cobblestones, but please give me some incentive to keep my six Pentium III's running!
If your Pentium III hosts are running mostly or entirely for BOINC, you might consider building a single new host. I think you'll find that for not very many hundred dollars at all you could considerably exceed the credit output of those six, vastly undercut their power consumption, and pay the initial cost back with not very many months of power bill savings.
If you don't like the thought of building your own, this is still true on the market.
I still run one 930 MHz Coppermine host. I currently am sketching intentions to replace it late this year with a Nehalem bottom end host. I think I can bring it in at about equal power, and get a little over twenty times the Einstein (or SETI) output, for well under $1000. Being a little less greedy in processor can push that initial outlay down a lot, and still come in over 6x.
Hint: within a given processor range, at default settings the lower frequency processors give a big power economy--you get linear power reduction from lower frequency, but a very big bonus from lower voltage. The specs are not much help on this, as TDP is only a weak indicator of actual power consumed by model.
Must say I totally agree with archae86, for a $1000 you could build a reasonable quad, as long as you keep the graphics card at the low end. And it would use less electricity than 2 * P3's, possibly the same as one. And it would easily do a BOINC RAC of 3000+.
In fact there are many quad cores you can buy already assembled for well under $1000 as long as you don't want the fastest ones (I have seen quad 2.4's for well under $1000)
Same with the faster dual cores (3.2's and higher)
I just wanted to let the crunchers here know that I am keeping an eye on this also. My latest check shows:
[pre]
mysql> select avg(cpu_time/granted_credit) from result where granted_credit>0 and name like '%S5R3%' ;
+------------------------------+
| avg(cpu_time/granted_credit) |
+------------------------------+
| 176.22502753646 |
+------------------------------+
1 row in set (1.47 sec)
mysql> select avg(cpu_time/granted_credit) from result where granted_credit>0 and name like '%S5R4%' ;
+------------------------------+
| avg(cpu_time/granted_credit) |
+------------------------------+
| 206.09569102025 |
+------------------------------+
1 row in set (1.62 sec)
[/pre]
So for the S5R4 run we have reduced our credits slightly in comparison with the S5R3 run, as needed to bring ourselves into alignment with SETI@Home. (See the BOINC project credit comparison page to see why this is needed.)
I'll continue to monitor the credit-related issues, and also will make some small scheduler/transitioner modifications so that claimed_credit and granted_credit are always the same as the current granted_credit.
Somebody competent (a moderator) please explain to me what it is that I don't comprehend about the new credit reduction. Is it based on credit per unit time or what measure?
Recent R3 tasks ran for approx 17.75 hours on my P4/3GHz/HT computer for let's say 235-237 credits. A 28 percent reduction would be 169 credits.
But the new R4 tasks are currently running 29.75 hours on my computer or 68 percent longer; therefore, I would expect to be granted approx 169 times 1.68 or 284 credits for a R4 task. I'm being granted around 221 credits or 22 percent less than I would expect even after the credit reduction.
Part of the answer is that you are not a typical Windows E@H user. The runtime you quoted for S5R3 was with a poweruser app, right? The vast majority of the Windows users were still using the stock app back in S5R3, and it was about 1.5 times slower.
So the average Windows stock app E@H user (by now migrated to the much faster S5R4 stock app) will hardly see any significant increase in runtime per workunit at all (also note the variation of runtime for different WUs of the same run which makes comparison difficult if based on only a few results).
If credits for S5R4 were raised according to your formula, the effect would be that the gap between SETI@Home credits/(CPU hour) and E@H credits/(CPU hour) would continue to widen, instead of shrink.
Bruce,
I wouldn't do any comparison's with Seti at the moment. They have not done a stats output for three days, due to server problems. And as they are also doing a credit realignment you will be comparing to out of date data.
I wouldn't do any comparison's with Seti at the moment. They have not done a stats output for three days, due to server problems. And as they are also doing a credit realignment you will be comparing to out of date data.
I've watched this ratio (Einstein@Home credit per CPU hour/SETI@Home credit per CPU hour) for a number of weeks now. It has not fluctuated much, though it is now slowly decreasing due to the Einstein@Home credit adjustments.
So I think that the comparison to SETI@Home is reasonable, even if their stats are a few days old. Do you know if their credit realignment will increase or decrease the average credit per CPU hour?
Why not increase credits by
)
Why not increase credits by 25% since electricity has gone up much more than thatin the last couple years? I've seen my retirement savings lose money the last year, my salary hasn't kept up with inflation for 3 years, and some useless project searching for little green men is concerned that people won't participate in their science fiction project unless credits are equal. So what? My disposable income has been reduced and now the credit I used to get for crunching is going down too? It isn't like I can pay my electric bill with cobblestones, but please give me some incentive to keep my six Pentium III's running! Longer WUs and 25% less credit is not the way.
RE: Why not increase
)
Well you pretty much told yourself which project is a waste of electricity Slicker (seti) but I see you aren't running that now.
But you do have those PIII's doing lots of work........maybe too much as far as projects.
I remember when I started doing classic seti's back in 2000 on my PIII 500 which died many years ago.
I have done quite a few of the new S5R4's and the claimed credit does look bad but the granted credit isn't all that bad and is less than a 25% difference.
Instead of the previous 237 credits I get between 163 to 192's and most between 185 and 192 credits. (I see 6 of them are 192 credits and 7 are around 185)
The light bill is more where I live too so I started turning off all my monitors now instead of using the blank screensaver like I had been since 2000.
Ok.......back to work.
RE: It isn't like I can pay
)
If your Pentium III hosts are running mostly or entirely for BOINC, you might consider building a single new host. I think you'll find that for not very many hundred dollars at all you could considerably exceed the credit output of those six, vastly undercut their power consumption, and pay the initial cost back with not very many months of power bill savings.
If you don't like the thought of building your own, this is still true on the market.
I still run one 930 MHz Coppermine host. I currently am sketching intentions to replace it late this year with a Nehalem bottom end host. I think I can bring it in at about equal power, and get a little over twenty times the Einstein (or SETI) output, for well under $1000. Being a little less greedy in processor can push that initial outlay down a lot, and still come in over 6x.
Hint: within a given processor range, at default settings the lower frequency processors give a big power economy--you get linear power reduction from lower frequency, but a very big bonus from lower voltage. The specs are not much help on this, as TDP is only a weak indicator of actual power consumed by model.
RE: RE: It isn't like I
)
Must say I totally agree with archae86, for a $1000 you could build a reasonable quad, as long as you keep the graphics card at the low end. And it would use less electricity than 2 * P3's, possibly the same as one. And it would easily do a BOINC RAC of 3000+.
In fact there are many quad
)
In fact there are many quad cores you can buy already assembled for well under $1000 as long as you don't want the fastest ones (I have seen quad 2.4's for well under $1000)
Same with the faster dual cores (3.2's and higher)
I just wanted to let the
)
I just wanted to let the crunchers here know that I am keeping an eye on this also. My latest check shows:
[pre]
mysql> select avg(cpu_time/granted_credit) from result where granted_credit>0 and name like '%S5R3%' ;
+------------------------------+
| avg(cpu_time/granted_credit) |
+------------------------------+
| 176.22502753646 |
+------------------------------+
1 row in set (1.47 sec)
mysql> select avg(cpu_time/granted_credit) from result where granted_credit>0 and name like '%S5R4%' ;
+------------------------------+
| avg(cpu_time/granted_credit) |
+------------------------------+
| 206.09569102025 |
+------------------------------+
1 row in set (1.62 sec)
[/pre]
So for the S5R4 run we have reduced our credits slightly in comparison with the S5R3 run, as needed to bring ourselves into alignment with SETI@Home. (See the BOINC project credit comparison page to see why this is needed.)
I'll continue to monitor the credit-related issues, and also will make some small scheduler/transitioner modifications so that claimed_credit and granted_credit are always the same as the current granted_credit.
Cheers,
Bruce
Director, Einstein@Home
Somebody competent (a
)
Somebody competent (a moderator) please explain to me what it is that I don't comprehend about the new credit reduction. Is it based on credit per unit time or what measure?
Recent R3 tasks ran for approx 17.75 hours on my P4/3GHz/HT computer for let's say 235-237 credits. A 28 percent reduction would be 169 credits.
But the new R4 tasks are currently running 29.75 hours on my computer or 68 percent longer; therefore, I would expect to be granted approx 169 times 1.68 or 284 credits for a R4 task. I'm being granted around 221 credits or 22 percent less than I would expect even after the credit reduction.
Hi! Part of the answer is
)
Hi!
Part of the answer is that you are not a typical Windows E@H user. The runtime you quoted for S5R3 was with a poweruser app, right? The vast majority of the Windows users were still using the stock app back in S5R3, and it was about 1.5 times slower.
So the average Windows stock app E@H user (by now migrated to the much faster S5R4 stock app) will hardly see any significant increase in runtime per workunit at all (also note the variation of runtime for different WUs of the same run which makes comparison difficult if based on only a few results).
If credits for S5R4 were raised according to your formula, the effect would be that the gap between SETI@Home credits/(CPU hour) and E@H credits/(CPU hour) would continue to widen, instead of shrink.
Bikeman
Bruce, I wouldn't do any
)
Bruce,
I wouldn't do any comparison's with Seti at the moment. They have not done a stats output for three days, due to server problems. And as they are also doing a credit realignment you will be comparing to out of date data.
RE: I wouldn't do any
)
I've watched this ratio (Einstein@Home credit per CPU hour/SETI@Home credit per CPU hour) for a number of weeks now. It has not fluctuated much, though it is now slowly decreasing due to the Einstein@Home credit adjustments.
So I think that the comparison to SETI@Home is reasonable, even if their stats are a few days old. Do you know if their credit realignment will increase or decrease the average credit per CPU hour?
Director, Einstein@Home