I am a hardcore seti cruncher, but I had joined Einstein to coverup outages. My RAC in Eintein used to be around 1000-2000, but now it has shot up to 90000. Hope Seti gets back online soon.

No need to panic, all the booze for nerve calming is on Juan today.

Happy Birthday

---

Pluto will always be a planet to me.

Seti Ambassador
Not to late to order an Anni Shirt

Happy B-Day Juan:)

---

rOZZ
Music
Pictures

Thanks friends. I wish some of you all could be here celebrating with me. I will rise a toast tonight in honor of you all.

---

Happy Birthday Juan.

---

With each crime and every kindness we birth our future.

Happy Birthday, Juan!

---

Member of the 20 Year Club

You are in this case, you're running a Boinc without OpenCL detection:

Another reason not to run Einstein anymore.
OpenCL app without own OpenCL detection.

---

With each crime and every kindness we birth our future.

You are in this case, you're running a Boinc without OpenCL detection:

Another reason not to run Einstein anymore.
OpenCL app without own OpenCL detection.

No, OpenCL support done the way it should be done - collaborating with BOINC, letting each part of the system play its proper part in the overall picture.

Taking OpenCL detection down a level to the science application layer (instead of the infrastructure layer) was a useful stopgap when the science app was ready before the infrastructure caught up, but that's history now. Time to grow up.

You are in this case, you're running a Boinc without OpenCL detection:

Another reason not to run Einstein anymore.
OpenCL app without own OpenCL detection.

No, OpenCL support done the way it should be done - collaborating with BOINC, letting each part of the system play its proper part in the overall picture.

Taking OpenCL detection down a level to the science application layer (instead of the infrastructure layer) was a useful stopgap when the science app was ready before the infrastructure caught up, but that's history now. Time to grow up.

Wrong way of thinking.
Einstein ignores thousands of users who wont use Boinc 7.
Their loss not mine.
Build in OpenCL detection wouldn`t hurt.

---

With each crime and every kindness we birth our future.

You are in this case, you're running a Boinc without OpenCL detection:

Another reason not to run Einstein anymore.
OpenCL app without own OpenCL detection.

No, OpenCL support done the way it should be done - collaborating with BOINC, letting each part of the system play its proper part in the overall picture.

Taking OpenCL detection down a level to the science application layer (instead of the infrastructure layer) was a useful stopgap when the science app was ready before the infrastructure caught up, but that's history now. Time to grow up.

Agreed, Don't send work to a client that doesn't report that it has a) OpenCL support, b) The correct OpenCL runtime for the app, anything else is a Bodge Job.

Claggy

You are in this case, you're running a Boinc without OpenCL detection:

Another reason not to run Einstein anymore.
OpenCL app without own OpenCL detection.

No, OpenCL support done the way it should be done - collaborating with BOINC, letting each part of the system play its proper part in the overall picture.

Taking OpenCL detection down a level to the science application layer (instead of the infrastructure layer) was a useful stopgap when the science app was ready before the infrastructure caught up, but that's history now. Time to grow up.

Agreed, Don't send work to a client that doesn't report that it has a) OpenCL support, b) The correct OpenCL runtime for the app, anything else is a Bodge Job.

Claggy

One of the reasons Einstein has only 10% of active members Seti has.
Simply stupid IMHO.

Richard always mentions bullet proof at Lunatics.
It seems not on einstein.

---

With each crime and every kindness we birth our future.

You are in this case, you're running a Boinc without OpenCL detection:

Another reason not to run Einstein anymore.
OpenCL app without own OpenCL detection.

No, OpenCL support done the way it should be done - collaborating with BOINC, letting each part of the system play its proper part in the overall picture.

Taking OpenCL detection down a level to the science application layer (instead of the infrastructure layer) was a useful stopgap when the science app was ready before the infrastructure caught up, but that's history now. Time to grow up.

Wrong way of thinking.
Einstein ignores thousands of users who wont use Boinc 7.
Their loss not mine.
Build in OpenCL detection wouldn`t hurt.

Totally agree Mike. Boinc 7 is an abomination. When I'm forced to upgrade to Boinc 7, because projects refuses to send work to older versions, then goodbye to those projects.

What exactly is the superiority of BOINC Version 6x?

Again wrong question.
The majority of Boinc users dont see a reason to update to Boinc 7.
Project admins shouln`t ignore this fact.
A lot of projects are still working with Boinc 6.
So what ?

---

With each crime and every kindness we birth our future.

What exactly is the superiority of BOINC Version 6x?

Whilst perhaps the wrong question, I would also like to know why so many people want to stay with Boinc 6.

What would I gain by switching back?

**Please note I never use the Boinc manager I only use Boinc Tasks, so any difference in the GUI I would not notice.

What other benefits would I see?.

---

What exactly is the superiority of BOINC Version 6x?

Whilst perhaps the wrong question, I would also like to know why so many people want to stay with Boinc 6.

What would I gain by switching back?

**Please note I never use the Boinc manager I only use Boinc Tasks, so any difference in the GUI I would not notice.

What other benefits would I see?.

How much time do you have ?
Boinc 7 prefers multi projects.
Running just one back offs are way to long for example.
The fact you can`t downgrade is a no go.
Thats enough for me.

---

With each crime and every kindness we birth our future.

What exactly is the superiority of BOINC Version 6x?

Whilst perhaps the wrong question, I would also like to know why so many people want to stay with Boinc 6.

What would I gain by switching back?

**Please note I never use the Boinc manager I only use Boinc Tasks, so any difference in the GUI I would not notice.

What other benefits would I see?.

How much time do you have ?
Boinc 7 prefers multi projects.
Running just one back offs are way to long for example.
The fact you can`t downgrade is a no go.
Thats enough for me.

The back offs are easy enough to work around using boinccmd & scripted commands to give BOINC a poke. Since I was doing that with BOINC v6.x.x I saw no change in behavior after switching to BOINC v7.x.x. I actually increased the time between pokes after switching to BOINC v7.

Not having the messages tab is dreadful. So I do stick with an older boincmgr.exe for the GUI on the machines where I even use Manager. Also I clear all of stuff in the files places in the Notices folder & then write protect them to prevent extra nonsense building up.

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

Running just one back offs are way to long for example.

LOL, you must be confusing it with 6.10 or before, which had long back offs on failed file transfers and project contacts. But since 6.12 the scheduler operations have a random back off, while file transfers slowly grow from 10 minutes to 12 hours.

This has not changed for 7.0, 7.2 or 7.4, only been tweaked.

Boinc 7 prefers multi projects.

BOINC 7 will run one project at a time, but can do multiple projects. They'll just run in order of priority (debt if you will). 6.12 could run different project applications at the same time.

But seeing how you run Seti only, I don't see where your problem is in this.

The fact you can`t downgrade is a no go.

You can downgrade, even from 7.0/7.2/7.4 back to 6.12
All you have to do is backup client_state.xml and client_state_prev.xml on your BOINC 6 installation and keep it in a safe place. Then when you do want to return to 6 from 7, overwrite the current client_state.xml and client_state_prev.xml files with the old ones, and that's it. Go on, uninstall 7, reinstall 6.

And if you don't have a present client_state.xml file, delete the ones of BOINC 7. The file will be remade with all the correct entries, the only thing that might happen is that you have to download new work if a project doesn't support resend lost work.

---

Nothing i dont know Jord.
I patched Boinc 6.10 and 6.12 to my needs.

I dont see a reason to upgrade since seti is my favorite project.
I just choose a project which works with what i have.
Shouldn`t be to hard to understand.

When i learned programming it was a necessity to make sure apps are working on as many hosts possible.
Now its smarter to find excuses.

---

With each crime and every kindness we birth our future.

Running just one back offs are way to long for example.

LOL, you must be confusing it with 6.10 or before, which had long back offs on failed file transfers and project contacts. But since 6.12 the scheduler operations have a random back off, while file transfers slowly grow from 10 minutes to 12 hours.

This has not changed for 7.0, 7.2 or 7.4, only been tweaked.

I'm running Seti only with v7.0.64 & the backoffs are ridiculous.
Doesn't take much time at all for CPU & GPU work requests to backoff for 20-22 hours.

---

Grant
Darwin NT

Running just one back offs are way to long for example.

LOL, you must be confusing it with 6.10 or before, which had long back offs on failed file transfers and project contacts. But since 6.12 the scheduler operations have a random back off, while file transfers slowly grow from 10 minutes to 12 hours.

This has not changed for 7.0, 7.2 or 7.4, only been tweaked.

I'm running Seti only with v7.0.64 & the backoffs are ridiculous.
Doesn't take much time at all for CPU & GPU work requests to backoff for 20-22 hours.

When there's no work available - deliberately, and sensibly, since there's no point in pestering the servers with requests that are bound to come back empty.

When work starts to flow again, hosts will gradually contact the servers and ramp up the server loads. A few people (like those who frequent these boards) will click the update button once and thereby clear the backoffs: the majority will let science take its course. Once an initial charge of work has been received, and is being processed and returned, you'll very rarely see a backoff longer than 600 seconds.

(my copy of BoincView - predecessor of BOINCtasks - is showing me the tasks being run by three hosts running BOINC 6, and four hosts running BOINC 7.4. Both sets are chosen to be fit for their respective purposes, and keep me familiar with both behaviours)

The 24 hrs max back off was a compromise, at my request, from the original back off of as much as 2 weeks for projects that were down or out of work.

---

Boinc V7.2.42
Win7 i5 3.33G 4GB, GTX470