Looking at the comments, and the credits being granted, I can't help feeling that someone has forgotten to include a chunk of the calculations in the credit awarded equation. I suggest this because, like others, I'm seeing tasks taking "twice" as long, but only getting "half" the credits. Surely if a task takes twice as long to execute on a given processor it should get double the credit, not half.

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

Looking at the comments, and the credits being granted, I can't help feeling that someone has forgotten to include a chunk of the calculations in the credit awarded equation. I suggest this because, like others, I'm seeing tasks taking "twice" as long, but only getting "half" the credits. Surely if a task takes twice as long to execute on a given processor it should get double the credit, not half.

That´s exactly what we all wants, but with Creditnew seem´s to be an impossible target.

---

Looking at the comments, and the credits being granted, I can't help feeling that someone has forgotten to include a chunk of the calculations in the credit awarded equation. I suggest this because, like others, I'm seeing tasks taking "twice" as long, but only getting "half" the credits. Surely if a task takes twice as long to execute on a given processor it should get double the credit, not half.

That´s exactly what we all wants, but with Creditnew seem´s to be an impossible target.

I was just asleep and woke up with a start, having had an idea that might just work... Calculate David Anderson's pay cheque with CreditNew....

Back to sleep

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

Looking at the comments, and the credits being granted, I can't help feeling that someone has forgotten to include a chunk of the calculations in the credit awarded equation. I suggest this because, like others, I'm seeing tasks taking "twice" as long, but only getting "half" the credits. Surely if a task takes twice as long to execute on a given processor it should get double the credit, not half.

That´s exactly what we all wants, but with Creditnew seem´s to be an impossible target.

I was just asleep and woke up with a start, having had an idea that might just work... Calculate David Anderson's pay cheque with CreditNew....

Back to sleep

LOL....
Lessee...hours worked X mystery algorithm = $1.50 for you this week, Dr. Anderson.

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

... Calculate David Anderson's pay cheque with CreditNew....

ROFL, I've been thinking that for some time.

It would have to be the whole Campus, though, otherwise there wouldn't be a large enough population to make the statistics significant. Heck, make it the whole State of California public payroll.

That feedback thread would be an interesting read....

I'm not interested in deep technical analisys over the different hardware and software platforms, what I know is that some copyright issues leaded to a renowed production of optimized apps, which overturned Intel-Nvidia superiority over AMD-ATI based platforms.
Now part of that new brand optimized code has been enbodied also in the stock apps within the projects. It's actually good for science but at my eyes not all looks so crystal-clear.
Someone complaining for the diminishing credits in the vast number crunchin' arena was answered this way: "If you're crunching SETI you should not care about absolute credits only relative credit within the project because our credit system makes no sense to most of us" or "Both will recive less credits.
Continue to compete but with shortened сarrots :)".
The fact is that carrot has shortened for Intel/Nvidia/Windows platforms and has lenghtened for AMD/ATI/Linux ones.
Creditnew refined the picture, making credit granting really no sense to most of us.
I don't understand why a so complicated policy for credits granting was adopted. I would prefer a plain Flopcounter based one. The same work has to give the same credits, fullstop.
Remember users are spending of their own for the Seti@home project's advance, both for the hardware purchase and for the energy bill, You have to keep the whole thing honest and clean.

Looking at the comments, and the credits being granted, I can't help feeling that someone has forgotten to include a chunk of the calculations in the credit awarded equation. I suggest this because, like others, I'm seeing tasks taking "twice" as long, but only getting "half" the credits. Surely if a task takes twice as long to execute on a given processor it should get double the credit, not half.

That´s exactly what we all wants, but with Creditnew seem´s to be an impossible target.

I was just asleep and woke up with a start, having had an idea that might just work... Calculate David Anderson's pay cheque with CreditNew....

Back to sleep

LOL!!! Brutal...

---

Something Raistmer said at beta kicked a couple of my maths neurons into action.
Now wait for the others to see if my reasoning is sound and whether Eric follows it.

---

A person who won't read has no advantage over one who can't read. (Mark Twain)

Something Raistmer said at beta kicked a couple of my maths neurons into action.
Now wait for the others to see if my reasoning is sound and whether Eric follows it.

Ooooooo nooooooooo...not MATH neurons....

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

Credit new has no clue what so ever what to do yet. This is a great example here : http://setiathome.berkeley.edu/workunit.php?wuid=1261682117

Something wrong with that, Explanation needed! The less work, the more credit!
Maybe there is a wrong sign (positive, negative) in the formula!

Don't think so, got one that ran for 5 seconds and gave 0.26 credits.

Granted, that one ended with an overflow, so it could be that the end result is being factored in as well?

I shall crunch my last seti/boinc work unit sometime today! Adiós my friends. If credit returns so shall I. looks like seti@home will not get a taste of my new i7-4770K machine on a GIGABYTE GA-Z87X-OC Force motherboard.

This is the only thing "the dictatorship" will listen to.

---

I was just asleep and woke up with a start, having had an idea that might just work... Calculate David Anderson's pay cheque with CreditNew....

Back to sleep

ha, i like it jason_gee.

---

I was just asleep and woke up with a start, having had an idea that might just work... Calculate David Anderson's pay cheque with CreditNew....

Back to sleep

ha, i like it jason_gee.

+ 1

---

Credit new has no clue what so ever what to do yet. This is a great example here : http://setiathome.berkeley.edu/workunit.php?wuid=1261682117

Something wrong with that, Explanation needed! The less work, the more credit!
Maybe there is a wrong sign (positive, negative) in the formula!

I ask and the answer was something like, one of the hosts don´t reach the 10 WU minimum to Creditnew works, so it credit the WU as a mean WU...

OK i agree is something totaly wierd... that way to think could only come from a "mad scientist mind". I just try to answer your question, not like that way too. I prefear the "old" way, more time/more work done more credit, like most of the others Boinc project does, easy & simple.

---

I shall crunch my last seti/boinc work unit sometime today! Adiós my friends. If credit returns so shall I. looks like seti@home will not get a taste of my new i7-4770K machine on a GIGABYTE GA-Z87X-OC Force motherboard.

This is the only thing "the dictatorship" will listen to.

I'm currently top 20 credit, at least you guys will have to work twice as hard to catch me:P has anyone looked at their total credit graph, it really tells the story!!! mine says my RAC should be half, yes half.

---

Credit new has no clue what so ever what to do yet. This is a great example here : http://setiathome.berkeley.edu/workunit.php?wuid=1261682117

Something wrong with that, Explanation needed! The less work, the more credit!
Maybe there is a wrong sign (positive, negative) in the formula!

The thoretical approach is that you get more credit for more efficiency. The app was 100% blanked so was very fast - in MB terms it was a -9. Credit however is awarded on the relative speed - very fast - lots of credit. credit doesn't know that the unit didn't slowly walk 100 steps but really took a shortcut from 1 directyl to 100.

Apart from that, I'm pretty sure there are numerous bugs in an already flawed design.

I think I'm really going to have to make time for that code walk.

---

A person who won't read has no advantage over one who can't read. (Mark Twain)

Back in the olden times they thought the sun revolved around the Earth & had the maths to back it up.
I would categorize Creditnew the same as these early astronomers calculations. It may work on paper, but in reality something is a bit wrong.

---

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

Back in the olden times they thought the sun revolved around the Earth & had the maths to back it up.
I would categorize Creditnew the same as these early astronomers calculations. It may work on paper, but in reality something is a bit wrong.

Ok, to take that picture - the therory is meant for the sun to revolve around the moon (*) and the calculations make Earth revolve around Jupiter.

(*) somebody took a theory in which the Sun revolves around the moon and tried to make it fit to data which match the sun revolving around earth.

---

A person who won't read has no advantage over one who can't read. (Mark Twain)

One of the things that seems to be missing from this discussion is the "feature" of CreditNew that attempts to adjust for the relative crunching power of the computers that process a particular WorkUnit. I don't pretend to understand how it works, but I remember from the discussions when CreditNew was first inflicted on us that one of the things it does is take a weighted average of the processing time of the 2 (or more) computers on that WU.

I have not kept detailed statistical data on my past results, but my general impression is that when my relatively old, slow Macs or P4s were paired with a similar machine, the result gave higher credit than when they were paired with GPUs or Anonymous Platform crunchers. Eric made a comment about the lower credit returns being due to only "faster" crunchers returniong work the first week after roll-out, and waiting to see what happened as more slower machines crunched and returned v7 work.

Do I like CreditNew? No. I don't like spending 42,000 seconds processing a Tasks that ends in a -9 overflow and getting 0.32 credits for it because my wingman only took 300 seconds to get the same result on his GPU. But it is what it is, and I'm not skilled enough to devise a more fair and equitable system. So until either Dr. Anderson deigns to change it, or Eric and co. come up with something better, it's what we have.

What is really starting to irritate me is all the whining and threats to leave the project because certain people feel that sudenly they are not getting the credit the think their work deserves, and that the project inflicted this on them on purpose. It ain't personal, and Eric has said elsewhere that it wasn't done on purpose. But if you really feel so personally offended and demeaned that you have to leave, I won't block the door......

It may take another couple of weeks for Credit and RACs to stabilize. There are a lot of v6 Tasks out there waiting to time-out and be resent. I'm going to wait and see how things look after the 4th of July. And enjoy a few cool beverages.

---

Donald
Infernal Optimist / Submariner, retired