Project Update

[chooka03]

Well I've decided to stop crunching Seti becuase the Geckos/U1 just can't handle 20GHs WU's very well.
This morning I've woken to find that none of the miners are stuck but 2 of my WU's are ticking over 15hrs.....so obviously they're not going to complete :( Just have to snooze BOINC to reset all the WU's.

Bit of a bummer as I'd rather crunch #5 over #12. Oh well.

[Dingo]

Well I've decided to stop crunching Seti becuase the Geckos/U1 just can't handle 20GHs WU's very well.
This morning I've woken to find that none of the miners are stuck but 2 of my WU's are ticking over 15hrs.....so obviously they're not going to complete :( Just have to snooze BOINC to reset all the WU's.

Bit of a bummer as I'd rather crunch #5 over #12. Oh well.

My Geckos rarely stop unless the PC stops running for some reason, or I see that the blue light is not flashing, or a work unit is over 2 hours.  Then if there is one or two that are not running I just unplug the Gecko from the powered hub and re insert and abort the work unit.  It usually starts up right away with the blue light flashing.  I do not stop BOINC because that stops all four of the Geckos and they restart at zero.

[chooka03]

G'day Dingo,
Nice to put a face to the name by the way 

I hear what you're saying. I've just arrived home from work and again 2WU's have been ticking away for 8hrs 22Min......... none of the lights are stuck on my ASIC miners though. I have a feeling its the U1's that struggle with the WU's, not the Gecko's.
I could test that theory can't can't be bothered at the moment. I'll just discontinue campaign #5 till we get some 1.6GH/s WU's (apparently they are coming soon according to rebirther)

[chooka03]

Actually, if I've done the maths right, crunching the 20GH/s tasks is going to be a lot worse for me in terms of credits.

225 sec on average for 30,000 credits with a 1.6GH/s task.
7000 sec on average for 1,200,000 credits with the 20GH/s tasks.

225 x 4 = 900. So that's 900 sec on 1.6 WU's to get 1.3Mil credits.

Did I get that right?

[Dingo]

I think there is a big difference and doing the  20+GH/s  gives better credit I think.  These are two of your tasks

run time / 3600 = time in hours        credit for wu / time in hours = credit per hour

  TASK     RUN TIME   credit for wu    credits per hour  

158055504    7,663.10    1,200,000.00    563,740.5227649385          20+GH/s  task        
158070831    2,122.41     30,000.00         50,885.54991731098           1.6+GH/s task.  

EDIT: Added a quich 1.6 GH task

158072470     171.46       30000.00         629,884.5211711186

So if all the 1.6 GH tasks ran in this short time you would be better off.  Looking at the tasks that have returned they are all over the place some in the 100's and others in the 1000's but if you round them all up I guess they will be about the same ????

[chooka03]

Hmmm..... that 2000 sec though isn't an average time for a 1.6GH/s WU though. I think they are closer to 250 sec. (websites down at the moment)
Also again....2 stuck WU's overnight. It seems to happen ALOT which has got to affect the amount of work done.

Having said that, the graph for average work done in BOINC is steadily increasing.

Thanks for crunching those numbers Dingo 

[JugNut]

Yep Dingo's right there chooka.  To put it another way...

If you divide 30, 000cr into 1,200,000cr .. you get 40.  So the 20GH tasks pay 40 times more than the 1.6GH tasks.  Meaning you'd have to do 40x 1.6GH tasks to equal just 1x 20GH task.

But when you divide 225sec's into 7000sec's you get close enough to 31.  So in 7000sec's you can only do 31x 1.6GH tasks not 40.

End result every time you do 1x 20GH task you basically get 9x 1.6GH tasks worth of credit for free.

Err I hope that made sense. :(


EDIT: LOL  Of coarse if they keep stopping all the time and wasting precious time then all of what I said above is meaningless,  so then my friend as you've already figured out you just gotta do what works ;)

[JugNut]

Eureka!!! I found the problem. I config'ed this array for 4 r-boxes and ran the 4 USB cables into an old Belkin 4-port hub (not powered). Either the hub is bad or they don't like hubs because when I gave each cable its own USB port, both started working.  It looks like one miner has a burned out status light. I consider that as a benefit as the last thing we need around here is another friggin', blinkin' light!!!  The down side is that when both are running the run times increased to 30 min. How can I tell what the hash rate of a device is?
I should get 2 more r-boxes tomorrow. The R4 has not shipped yet and I don't know when I will get it.

Way to go buddy you did well today and gave me a little spanking too..   Mmm and there's still more gear to come? Rocket ship Dataman about to leave orbit LOL     
On the serious side be careful that your fuses can handle the extra load.

You can find the hash rate by looking in stderr.txt file the slot directory.  Just click on the task you want to check in boinc manager then click on the properties button. The slot directory is listed on the forth line from the bottom of the box that opens.

[Dataman]

Eureka!!! I found the problem. I config'ed this array for 4 r-boxes and ran the 4 USB cables into an old Belkin 4-port hub (not powered). Either the hub is bad or they don't like hubs because when I gave each cable its own USB port, both started working.  It looks like one miner has a burned out status light. I consider that as a benefit as the last thing we need around here is another friggin', blinkin' light!!!  The down side is that when both are running the run times increased to 30 min. How can I tell what the hash rate of a device is?
I should get 2 more r-boxes tomorrow. The R4 has not shipped yet and I don't know when I will get it.

Way to go buddy you did well today and gave me a little spanking too..   Mmm and there's still more gear to come? Rocket ship Dataman about to leave orbit LOL     
On the serious side be careful that your fuses can handle the extra load.

You can find the hash rate by looking in stderr.txt file the slot directory.  Just click on the task you want to check in boinc manager then click on the properties button. The slot directory is listed on the forth line from the bottom of the box that opens.

Thanks JN! I looked at one and it shows: Directory: slots/6     ??? Don't know what that means.
Yep, I have two more r-boxes that are "vacationing" in Salt Lake City due to heavy snow and 0 visibility over the Sierra's. Been there for two days now. Also have a Rockminer R4 that has not yet shipped.
I am not happy with the performance of the 2 r-boxes. The two ant U3's are outperforming them. The only difference between the two PC's is the rocks are on 2.67 GHz processor and the ants are on a 3.33 GHz machine. Otherwise they are identical.  ???
I have the farm spread over 3 circuits so no problem there. The rocks are not producing much heat and they are more quiet than I thought they would be.

[kashi]

Good to hear of your upgrade adventures.

Directory: slots/6 is found in C:\ProgramData\BOINC\slots 
Record overnight temps here, send snow please.

[JugNut]

@Dataman: Sorry for the late reply I just had family arrive.  Yep as kashi said it's in the boinc/slots folder, just open the file called stderr.txt in notepad. 

Also my R-Box does 110GH which credits about 115 - 130 million a day so it give almost twice the credit as the U3's do.  I find it best to only use one type of miner per box as some of them don't seem to play well together. Are the R-boxes on a box by themselves?(no other U3's ect)

[Dataman]

Thanks kashi and JN. I will look at it sometime today. Busy schedule today.
The two r-boxes are by themselves on one machine and not crunching CPU or GPU. ??? Same with the U3's. The r-box run times are all over the place from 18 min to 55 min. The U3's are consistent at ~28 min.
My turn to make breakfast today. Oatmeal anyone? Gotta run ...

EDIT: Found it. Do not know what most of it means but am thinking "(avg.) 0.000 h/s" is not good! Also lots of lines "Failed to get *filename* device in use" and "no device detected" are not good either. Having said that, both are completing wu's eventually. Will take a deeper look when we get back today. I need to complete the build and restack before starting to debug them. At least they are doing something. Cheers!

EDIT2: I still cannot determine a hash rate. All of the numbers in both files are 0.000 h/s. What am I supposed to be looking for? One is stable and doing wu's in 18-19 min. The other one is all over the place from 19 to 59 min. I don't know which is which. Next week I will take one down and just run them one at a time to find the unstable one. Until then something is better than nothing. My other two are still stuck in Salt Lake City. The R4 still has not shipped.

[JugNut]

Which of your PC's has the R-Boxes attached them? Can tell me what the computer ID# is from this list of your computers?  http://www.bitcoinutopia.net/bitcoinutopia/hosts_user.php?sort=expavg_credit&rev=0&show_all=0&userid=7236 

I'm guessing it's the PC with ID: 14115 (with the i7 920)  If I click on tasks then pick the first task I see it shows this..  http://www.bitcoinutopia.net/bitcoinutopia/result.php?resultid=158371377 
This is what the stderr.txt file in the slot directory should look like as well.(or very similar)
As you see your task started at 83.77Gh/s at the top and finished at 78.04Gh/s at the bottom. In the "Summary of runtime statistics:" Section it shows the Average hashrate: 78015.4 Mhash/s (78.015GH)

This is one of mine http://www.bitcoinutopia.net/bitcoinutopia/result.php?resultid=158337006  It starts at 94.44Gh/s and ends at 101.5Gh/s.  The Average is 101378.6 Mhash/s.(101.378GH) When I first got it, it was 108-110GH(it still needs a dust out) 

[chooka03]

Good to see 1.6GH/s WU's now available for campaign #5.
I really was having to abort too many WU's with the 20GH/s WU's :(

[Dataman]

Which of your PC's has the R-Boxes attached them? Can tell me what the computer ID# is from this list of your computers?  http://www.bitcoinutopia.net/bitcoinutopia/hosts_user.php?sort=expavg_credit&rev=0&show_all=0&userid=7236  

I'm guessing it's the PC with ID: 14115 (with the i7 920)  If I click on tasks then pick the first task I see it shows this..  http://www.bitcoinutopia.net/bitcoinutopia/result.php?resultid=158371377  
This is what the stderr.txt file in the slot directory should look like as well.(or very similar)
As you see your task started at 83.77Gh/s at the top and finished at 78.04Gh/s at the bottom. In the "Summary of runtime statistics:" Section it shows the Average hashrate: 78015.4 Mhash/s (78.015GH)

This is one of mine http://www.bitcoinutopia.net/bitcoinutopia/result.php?resultid=158337006  It starts at 94.44Gh/s and ends at 101.5Gh/s.  The Average is 101378.6 Mhash/s.(101.378GH) When I first got it, it was 108-110GH(it still needs a dust out)  

DM08 has the 2 r-boxes (for now). Will move them to DM02 when I get the other 2.
Thanks, I will check it out later. One box is working OK but the other one had been sleeping for 2 hours when I got up. Hot plugged it and it immediately finished. The other two finally made it over the mountains and I should get them Monday. By then I should know which one of the existing ones is good and which is bad. Depends what Karen has planned for us today. Thanks for looking into it mate.
EDIT:OK, I understand now. Thanks for the help. The one thing that still confuses me is am looking at wu's that ran in 17 min that have an average rate of 80 Gh/s and others that ran in 1 hour and 20 min with a hash rate of 80 Gh/s. So the time it is not doing anything is not averaged in? But this helps as it looks like the 2 are doing about 80 Gh/s not 110 Gh/'s. But it is used equipment. I'll take what I can get.
EDIT2: Rather dismal for one r-box. Starts as low as 4 Gh/s; runs most of the time at 15 Gh/s but finishes with an overall Average of 79 Gh/s ??? The math make no sense to me.