Inconclusive?

[[Ion]]

I was checking through my results status page on WCG (to make sure there weren't any errors I was unaware of), and I saw two WUs that are marked as "Inconclusive". What does this mean?

 

[ERazer]

I was checking through my results status page on WCG (to make sure there weren't any errors I was unaware of), and I saw two WUs that are marked as "Inconclusive". What does this mean?
http://img.techpowerup.org/091112/wcg.png

I got 8 inconclusive

 

[[Ion]]

bump

 

[Velvet Wafer]

I was checking through my results status page on WCG (to make sure there weren't any errors I was unaware of), and I saw two WUs that are marked as "Inconclusive". What does this mean?
http://img.techpowerup.org/091112/wcg.png

I got 8 inconclusive

watch out guys, if one of these inconclusives, gets invalid,
your OVERCLOCK ISTN STABLE
my 939 cruncher produced masses of inconclusives endig in valids and invalids, until i dropped it a notch
btw invalids give you only 50% points ;-)

 

[[Ion]]

None of my rigs are OCed
Q9400 is at stock 2.66, E6550 is at stock 2.33

 

[Velvet Wafer]

None of my rigs are OCed
Q9400 is at stock 2.66, E6550 is at stock 2.33

then it will possibly turn out to be valid. might take a day,just

 

[hat]

watch out guys, if one of these inconclusives, gets invalid,
your OVERCLOCK ISTN STABLE
my 939 cruncher produced masses of inconclusives endig in valids and invalids, until i dropped it a notch
btw invalids give you only 50% points ;-)

I don't understand this. Aren't we supposed to be crunching these work units because nobody else has and we're bringing light to the unknown? How do they know if the work unit is bad or not?

 

[Jstn7477]

Thanks ION! I just checked my stats and nearly all my WUs coming from my desktop are "ERROR" and a few "Inconclusive". All the ones coming from my laptop are Valid. I need to figure out wtf is wrong with my desktop, between intermittent graphics issues that are fixed when I restart to a ton of bad WUs that it is pumping out.

 

[Kreij]

Depending on what project you are working on , depends on how the returned WUs are validated. You ARE crunching something new, but they must make sure that the data returned is valid.

Take for example the childhood cancer project. They will give you a WU and also give the same WU to another computer. If the results returned are the same, they deem them valid. If not, they will toss the WU out to more computers and then look at the results. They get a high level of reliability by using a consensus model. If a majority of computers return a particular results, they deem it valid and then compare it to the orginal responses. If yours matches it is valid.

That's why you see inconclusives. Something did not come out in the wash, so they are waiting on more returned results for the WU to determine what the valid outcome should be.

Several things can cause invalid results, so they have to be sure the returned data has a high likelyhood of being correct, before accepting the results.

 

[p_o_s_pc]

Thanks ION! I just checked my stats and nearly all my WUs coming from my desktop are "ERROR" and a few "Inconclusive". All the ones coming from my laptop are Valid. I need to figure out wtf is wrong with my desktop, between intermittent graphics issues that are fixed when I restart to a ton of bad WUs that it is pumping out.

that means its not stable..bump voltage a notch

 

[Jstn7477]

I just put my CPU back at stock 3.2GHz/1.325v with 2000HT/NB. Was previously at 2.8/1.1375v/2.2 which I guess is unstable apparently. If I start returning all good WUs, means my underclock sucked. When it is stable again I will try to slowly undervolt it and make sure it is working 100% of the time.

 

[p_o_s_pc]

I just put my CPU back at stock 3.2GHz/1.325v with 2000HT/NB. Was previously at 2.8/1.1375v/2.2 which I guess is unstable apparently. If I start returning all good WUs, means my underclock sucked. When it is stable again I will try to slowly undervolt it and make sure it is working 100% of the time.

why not just stop it and run linpack for a few hours

 

[Velvet Wafer]

I just put my CPU back at stock 3.2GHz/1.325v with 2000HT/NB. Was previously at 2.8/1.1375v/2.2 which I guess is unstable apparently. If I start returning all good WUs, means my underclock sucked. When it is stable again I will try to slowly undervolt it and make sure it is working 100% of the time.

if it survives one Hour Linx,or any linpack you prefer, on about 1024mb, it shall crunch at least 24 hours if not 48, without crashing

 

[p_o_s_pc]

if it survives one Hour Linx,or any linpack you prefer, on about 1024mb, it shall crunch at least 24 hours if not 48, without crashing

I run it for 4hrs most of the time.If it survives that then i let it crunch it has yet to fail me.

 

[Velvet Wafer]

I run it for 4hrs most of the time.If it survives that then i let it crunch it has yet to fail me.

on k9, its easier to get stable.
believe me, the 955 is a bitch.

 

[p_o_s_pc]

on k9, its easier to get stable.
believe me, the 955 is a bitch.

I think you mean K8 K9 was dropped before it was released.

I'm sure its harder on the newer chips to get stable but i would still think that the same rule would fallow.More time stress testing the more chance its going to be stable under crunching and everyday use.

 

[Jstn7477]

Alright, I ran Linpack in 64bit mode (as part of OCCT). I ran it stable for just over 2 hours (I cancelled it after that) at 3.2GHz/1.275v/2000NB. I then ran it for an hour at 2.8GHz/1.175/2000NB stable. I will probably run it for 4hrs tomorrow to see if it is definitely stable.

 

[Velvet Wafer]

I think you mean K8 K9 was dropped before it was released.

I'm sure its harder on the newer chips to get stable but i would still think that the same rule would fallow.More time stress testing the more chance its going to be stable under crunching and everyday use.

youre right, but that hurts the crunching. you need to balance, testing and crunching.
i thought K9 is the Name for a double K8... no one corrected me till yet

Alright, I ran Linpack in 64bit mode (as part of OCCT). I ran it stable for just over 2 hours (I cancelled it after that) at 3.2GHz/1.275v/2000NB. I then ran it for an hour at 2.8GHz/1.175/2000NB stable. I will probably run it for 4hrs tomorrow to see if it is definitely stable.

sounds like you got a base to work on. enjoy

 

[p_o_s_pc]

@velvet it doesn't do any good to crunch if you return work that has errors in it. It doesn't benefit them at all.That is why i do my testing then crunch.I also keep an eye on the stats to make sure my OC is stable(doesn't return errors)

 

[Velvet Wafer]

@velvet it doesn't do any good to crunch if you return work that has errors in it. It doesn't benefit them at all.That is why i do my testing then crunch.I also keep an eye on the stats to make sure my OC is stable(doesn't return errors)

it can crash, even if it doesnt return errors. see my 955