Ned,
You running optimized? or Stock ?
S

I've run both.

Running an optimized application is a responsibility -- you are taking over the whole job of making sure you have a suitable app for the available work.

... and I decided a while ago, that I'd switch back from Optimized to Standard so I wouldn't have to carefully monitor the transition to Enhanced.

Because some of the go-fasts in the optimized clients found their way into Enhanced, there isn't the same boost for optimized-enhanced, so I'm probably staying on the standard client.

---

Yep, but if a work unit take about 45 hours, you get angry.
No matter what: 335089759 . And 5 of those, on a same computer . Grrr…………..
And someone finish it in 49,554 seconds.

cya2

Old units = smaller = LESS SCIENCE DONE

You are more interested in small units than the science.

Science is good, but when you want you know, when a computer left a building, is bad! At work……………….

RAID 1, Xp pro, 1 gigs memory 400, amd165 dual core.

So, you'd rather SETI did less science, so you can use SETI's connect times to keep track of your computers?

Okay, I'm done.

---

-----

Me too. No fancy stuff of any kind.
Only peculiarity I do is disable network most of the time.
And connect long enough to refill the cashe.

No stress. Generally if it's working for me, everyone else is working.
If I'm down, so is everyone else.

I'm gonna go take a peek at the Home PC and look at the WU times.
I don't understand why mine are all the same give or take 1/2 hour.
I'll post in a couple what I see.

S

CYA2 : I see your post. I haven't forgotten you

---

Me too. No fancy stuff of any kind.
Only peculiarity I do is disable network most of the time.
And connect long enough to refill the cashe.

I've been peeking at the project wide stats at BOINCSTATS and in general, it looks pretty flat -- but I don't think the graphs will mean that much until we've had a little more run-time.

My only cruncher is a Win2K workstation that is shut down nightly. I don't disable network access or anything like that. I participate in the BOINC Alpha project, so I'm often running a development version of BOINC, but I've been on 5.4.9 since the release.

I don't see any "red" in my results currently, but this machine isn't the fastest on the block.

---

CYA2

You said....
"Science is good, but when you want you know, when a computer left a building, is bad! At work………………."

And the 'Raid1'

Please explain these added details.

S

---

The only thing that get me back!

IS

SHORT WORKUNITS !

Please !

Yep, but if a work unit take about 45 hours, you get angry.
No matter what: 335089759 . And 5 of those, on a same computer . Grrr…………..
And someone finish it in 49,554 seconds.
cya2

No, I did not agree.

A WU with longer computing times results from the angle rate from which the data was recorded. The recivers are sensitver in some ranges.
The goal of enhanced is, to get more sentiviy and analyse the data more sentiver than before and look deeper into the sky for signals. This is very important for the goal of the science. With SETI Enhanced the sensitvy has improved by factor 10 compared with the old SETI.
And the science are the reason for my, why I do SETI. For me it is a logical and needed step, to improve the science. And I think, that the SETI devs has done a very good job to reach this goal.

With increasing the longer crunching times the bandwith on the berkley network conncetion will decreased. This was also a needed step, because the network connection (100 MBit) was often on the upper limit. A upgrade to a faster network connection in this range is really expensive...

Also, the increasement of the crunching times has the nice benefit, that the server load will decrease. The servers and the databases was often under a heavy load. If you will handle this load you have to reduce the data which will put on this servers, or upgrade your infrastructure. But stoagesystem and special servers are also expensive...

The other problem is, that SETI has not much funding. So the hardware and infrastructure invests are really hard to realise.

As summary, I must say, the higher sentivity and the reduced load for the servers are a good thing. I see the project on the right way.

---

http://setiweb.ssl.berkeley.edu/workunit.php?wuid=80224841
http://setiweb.ssl.berkeley.edu/show_host_detail.php?hostid=2140943
This sucks............. And 5/45 hours is still counting.

---

Ned,
You're right. Coincidence on the notebook.
Just took a look at the Dual Core home PC
and the times range from 3:25 to 7:03. There are about
20 or so ready to send.
So I suppose a variance of double the best isn't
beyond believable
S

---

Ned,
You're right. Coincidence on the notebook.
Just took a look at the Dual Core home PC
and the times range from 3:25 to 7:03. There are about
20 or so ready to send.
So I suppose a variance of double the best isn't
beyond believable
S

If Ingleside's table is accurate (I suspect it is) then 2:1 is not only believeable, but well within the expected 12:1 range.

---

CYA2

What was that link for.
It isn't relavent.

Lets move on. If you want to.
If not, lets just pull the plug.

S

---

CYA2

What was that link for.
It isn't relavent.

Lets move on. If you want to.
If not, lets just pull the plug.

S

Yep, done...........

Resource share 0.1

---

Why do people say that like they think they're hurting us?

If you want to leave, fine. Your removing of the project doesn't hurt us one bit. More work units for the rest of us.

---

-----

CYA

Please:

1. Open up your BOINC Manager
2. In the activity tab - Suspend Network Activity.
Leave it suspended. You don't seem to need it but
every 45 hours anyway. ( It's what I do )
3. Open the task tab.
If there are any 100% ones ( Probably not )
record the CPU time
4. For the ones not yet started.
Record the longest and shortest
to completion time.

Let us know what you get.

Steve

PS: Everyone else. Please hold off on the comments not directly related
to CYA's run time issue. Lets keep CYA focused.

---

Steve,

If you follow those links he gave you can get to the host results for his Opteron, which look ok to me.

There is no issue, it's just more P & M'ing about the SE credit landcape.

Alinator

Steve,

If you follow those links he gave you can get to the host results for his Opteron, which look ok to me.

There is no issue, it's just more P & M'ing about the SE credit landcape.

Alinator

You must be out this universe, a true alien!

Thanks!

---

Steve,

If you follow those links he gave you can get to the host results for his Opteron, which look ok to me.

There is no issue, it's just more P & M'ing about the SE credit landcape.

Alinator

So you're telling me that you see "no issue" with WU 80224841

????

Someone spent 30 hours for a result that is trash and you don't see an issue?

---

You must be out this universe, a true alien!

Thanks!

I looked at your Opteron, and you've had 1 abnormal run out of the almost 50 showing, far better than some folks here who have had real trouble.

As Matt pointed out, as for not being able to use SAH as you own personal remote box heartbeat monitor hour by hour, tough. Join Predictor they have nice short WU's.

BTW, introducing yourself to the fora by spam crossthread posting your Farewell Address kinda indcates you're at least as clueless as you think I am.

Alinator

Steve,

If you follow those links he gave you can get to the host results for his Opteron, which look ok to me.

There is no issue, it's just more P & M'ing about the SE credit landcape.

Alinator

So you're telling me that you see "no issue" with WU 80224841

????

Someone spent 30 hours for a result that is trash and you don't see an issue?

Thanks! Brian, I am not lost, maybe…….

One of those 45-hour work units is now counting:

2140943 on my 165 computer and this
http://setiweb.ssl.berkeley.edu/workunit.php?wuid=80341582
seems really strange !

cya2

---

Thanks! Brian, I am not lost, maybe…….

You're welcome... However, please do me a favor and try to settle down a little, ok? The multiple posts can and likely will be seen as spamming or, as was already said, Pissing and Moaning...

Some people are trying to help you. Please try to let them. If it is determined that you're getting these long-running computation error units, that will be seen by developers and looked into.

Thanks!

Brian

---

Steve,

If you follow those links he gave you can get to the host results for his Opteron, which look ok to me.

There is no issue, it's just more P & M'ing about the SE credit landcape.

Alinator

So you're telling me that you see "no issue" with WU 80224841

????

Someone spent 30 hours for a result that is trash and you don't see an issue?

I don't think anyone is saying that there are no issues, just that it's hard to work on the facts with all the rhetoric flying around.

So, let's work through the results one-by-one.

This one errored out in zero time. The "no child process" error is being discussed on the alpha testers list, and has only been observed in the wild. May take a while to find that one. This machine seems to error every time, and has been limited by BOINC to one work unit per day until it starts returning results.

This one says that this work unit was noisy. I don't recognize the BOINC version (5.2.8), but it should have claimed much higher credit. I wonder if the same would have happened on 5.4.9 (by which I mean I don't know).

This machine turned in a valid result (many signals in this WU).

This one is the same. BOINC 5.2.13.

This machine returned a valid result (many signals in this WU).

This one crashed. Speaking as a programmer, this is the only one here that really bothers me because the science application ought not to crash. Did it crash because it's overclocked, or is the hardware failing, or is there a conflict with other software -- or is it just a bug?

One work unit doesn't tell us a whole lot about BOINC in general. We need to look at others -- especially those running standard apps on the latest standard version of BOINC.

---