If we are processing 4 bit work units could you please explain why the Multibeam tasks don't take twice as long?

The 2 to 4bit change was to improve the S/N (Signal to Noise) ratio, we aren't actually doing any extra work on the Work Unit. ie the processing done remains unchanged.

From Eric's post in Beta-

I'll be distributing some workunits with 4-bit complex samples. They'll be twice the size of the current 2-bit workunits, but will give us 46% reduction in noise power.

It's possible that some of the app versions date from before the 4-bit quantization code existed in the client (but I don't think so). If things are going to break it should be obvious pretty quickly.

The applications at the time were able to process the 4bit WUs, so the splitters on Main were updated to produce them, no need for new applications, no change to the work being done.

+1
Yes, exactly.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

If we are processing 4 bit work units could you please explain why the Multibeam tasks don't take twice as long?

The 2 to 4bit change was to improve the S/N (Signal to Noise) ratio, we aren't actually doing any extra work on the Work Unit. ie the processing done remains unchanged.

From Eric's post in Beta-

I'll be distributing some workunits with 4-bit complex samples. They'll be twice the size of the current 2-bit workunits, but will give us 46% reduction in noise power.

It's possible that some of the app versions date from before the 4-bit quantization code existed in the client (but I don't think so). If things are going to break it should be obvious pretty quickly.

The applications at the time were able to process the 4bit WUs, so the splitters on Main were updated to produce them, no need for new applications, no change to the work being done.

From Raistmer post in Beta

It seems some questions regarding SETI's project bandwidth increase should be solved before full-scale deployment on main.
So, hard to predict when it will be implemented.
Beta with much less traffic is OK to use 4-bit tasks further though.

---

From Raistmer post in Beta

Not sure what that has to do with anything- that was posted near the end of 2016.

From Eric's original post,

I'll be distributing some workunits with 4-bit complex samples. They'll be twice the size of the current 2-bit workunits

From his second post in that thread,

There won't be any easy way to tell what you've got apart from the file size. 2-bit WUs are about 360K. 4-bit WUs are about 620K. The "2bit" or "8bit" designation in the file name is about the input data that the splitter operated upon. (And it's wrong anyway. The 2bit input data is actually 4 bits per complex sample.)

In the workunit file itself, there's a relatively new entry called wu_bits_per_sample which should be 2 past data, and in the new data should say 4.

If you can find the date that the WUs on main went from 360k to 700k in size, then you'll know the date that the 4bit WUs were released here.

---

Grant
Darwin NT

In the workunit file itself, there's a relatively new entry called wu_bits_per_sample which should be 2 past data, and in the new data should say 4.

If you can find the date that the WUs on main went from 360k to 700k in size, then you'll know the date that the 4bit WUs were released here.[/quote]
Those work unit results will be long gone from the server, I would think

---

As far as I am aware we are still processing 2 bit work units.

They've been 4bit WUs for at least 2.5 years. The larger WU file size was the result of that change.

Not according to the latest post in that the thread that I linked to. If we are processing 4 bit work units could you please explain why the Multibeam tasks don't take twice as long?

. . If you take another look at that thread you posted the link to you will see that it ended in 2016. A lot of water has passed under the bridge since then :)

Stephen

:)

Yes, we went from the 2 to 4 bit shortly after Eric posted that in the Beta thread back in 2016. No follow up post by him in Beta since it went live on Main.

---

Yes, we went from the 2 to 4 bit shortly after Eric posted that in the Beta thread back in 2016. No follow up post by him in Beta since it went live on Main.

Yes you are absolutely correct the post I am referring to was posted in 2016. Since there has been no update I still believe we are processing 2 bit work. Until I see 4 bit in the task name or hear from Jeff, Matt, Eric or Mark or see an official post on main I will have to agree to disagree that we are processing 4 bit work

---

Yes, we went from the 2 to 4 bit shortly after Eric posted that in the Beta thread back in 2016. No follow up post by him in Beta since it went live on Main.

Yes you are absolutely correct the post I am referring to was posted in 2016. Since there has been no update I still believe we are processing 2 bit work. Until I see 4 bit in the task name or hear from Jeff, Matt, Eric or Mark or see an official post on main I will have to agree to disagree that we are processing 4 bit work

. . Suit yourself, they say ignorance is bliss ....

Stephen

<shrug>

Since we haven't convinced you that we are already running 4 bit work, why don't you ask your question directly to Eric or Jeff in a PM.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Until I see 4 bit in the task name or hear from Jeff, Matt, Eric or Mark or see an official post on main I will have to agree to disagree that we are processing 4 bit work

The postings from Dr. Korpela:

I'll be distributing some workunits with 4-bit complex samples. They'll be twice the size of the current 2-bit workunits, but will give us 46% reduction in noise power.

It's possible that some of the app versions date from before the 4-bit quantization code existed in the client (but I don't think so). If things are going to break it should be obvious pretty quickly.

and also

There won't be any easy way to tell what you've got apart from the file size. 2-bit WUs are about 360K. 4-bit WUs are about 620K. The "2bit" or "8bit" designation in the file name is about the input data that the splitter operated upon. (And it's wrong anyway. The 2bit input data is actually 4 bits per complex sample.)

In the workunit file itself, there's a relatively new entry called wu_bits_per_sample which should be 2 past data, and in the new data should say 4.

Edit: I also checked and all of mine are 721KB and the Arecibo ones contain "<wu_bits_per_sample>4</wu_bits_per_sample>"

This has been posted twice in this thread...

Still doesn't seem like enough though... lol.

---

In the workunit file itself, there's a relatively new entry called wu_bits_per_sample which should be 2 past data, and in the new data should say 4.

This has been posted twice in this thread. Why don't you just open one of your work units with an editor and search on the wu_bits_per_sample term and you will see it is set at a value of 4.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Since there has been no update I still believe we are processing 2 bit work. Until I see 4 bit in the task name or hear from Jeff, Matt, Eric or Mark or see an official post on main I will have to agree to disagree that we are processing 4 bit work

Why on Earth would you believe what Eric posts now, when you are ignoring what he has posted previously???
Seriously?

Once again, here it is.

There won't be any easy way to tell what you've got apart from the file size. 2-bit WUs are about 360K. 4-bit WUs are about 620K. The "2bit" or "8bit" designation in the file name is about the input data that the splitter operated upon. (And it's wrong anyway. The 2bit input data is actually 4 bits per complex sample.)

As the WUs we used to download were 360k in size, and the WUs we are downloading now (and for some time) are 700k in size, we are processing 4 bit work.
It really is that simple.

---

Grant
Darwin NT

In the workunit file itself, there's a relatively new entry called wu_bits_per_sample which should be 2 past data, and in the new data should say 4.

Why don't you just open one of your work units with an editor and search on the wu_bits_per_sample term and you will see it is set at a value of 4.

Simply because currently I don't have any work on my system.

Keith wrote

Since we haven't convinced you that we are already running 4 bit work, why don't you ask your question directly to Eric or Jeff in a PM

Keith I'm sorry you don't think I'm convinced. I find it interesting that all talk of change to 4 bit work has been discussed on beta however since it has gone to main there has been no mention of it on these boards. Simple reason I am not going to PM Matt, Jeff & Eric is because I believe they only respond to a few people.

Grant (SSSF) wrote

Why on Earth would you believe what Eric posts now, when you are ignoring what he has posted previously???
Seriously?

It has come to my realisation that I think I used the wrong wording when originally posting. I believe what Eric says. For further clarification please read again what I put in my comment under what Keith wrote.

---

Thank you for all of the replies in regards to this topic. It is clear from what everyone has said that we are processing 4 bit work. I apologise if anybody doesn't think I believe them. I also apologise for taking up this thread with my discussion

---

Thank you for all of the replies in regards to this topic. It is clear from what everyone has said that we are processing 4 bit work. I apologise if anybody doesn't think I believe them. I also apologise for taking up this thread with my discussion

Yeah, but since it didn't drive us into a "Panic" then its ok! After all this IS the "Panic Mode On" thread :)

Tom

---

A proud member of the OFA (Old Farts Association).

Thank you for all of the replies in regards to this topic. It is clear from what everyone has said that we are processing 4 bit work. I apologise if anybody doesn't think I believe them. I also apologise for taking up this thread with my discussion

Yeah, but since it didn't drive us into a "Panic" then its ok! After all this IS the "Panic Mode On" thread :)

Tom

You are so right Tom

---

We'll just blame the interruption on that long white cloud. :-D

Cheers.

OK, Wiggo . . . . you had me going Huh? with that reference. So I had to look it up to see what it might be referring to.

Landed on this video. https://www.youtube.com/watch?v=3sHE_AzN3lU
And this. https://www.telegraph.co.uk/news/worldnews/australiaandthepacific/newzealand/4315307/Swirling-cloud-captured-above-New-Zealand-The-Land-of-the-Long-White-Cloud.html
Totally cool. I get it now.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Keith I'm sorry you don't think I'm convinced. I find it interesting that all talk of change to 4 bit work has been discussed on beta however since it has gone to main there has been no mention of it on these boards. Simple reason I am not going to PM Matt, Jeff & Eric is because I believe they only respond to a few people.

. . That is the simplest thing to understand. In Beta it was new and under test and the discussion there was about the manner in which it processed and whether there were any issues. When it moved to main that was all very much settled. But there were still some messages about the new format, I know that for a fact because I posted some of them, but the topic faded quickly because for the rank and file it was a non-event. There were no obvious changes apart from the larger files, which was pretty much the sole topic in main, allied with the fact they did NOT taken any longer to process which was the only major concern with the larger file sizes. I think there were one or two who were worried about the extra download time because they were on slow internet connections but it was all over quite quickly. I am sure if you want to go back over the closed versions of this thread dating back about 2 years or so you can still find those messages.

Stephen

:(

OK, Wiggo . . . . you had me going Huh? with that reference. So I had to look it up to see what it might be referring to.

Landed on this video. https://www.youtube.com/watch?v=3sHE_AzN3lU
And this. https://www.telegraph.co.uk/news/worldnews/australiaandthepacific/newzealand/4315307/Swirling-cloud-captured-above-New-Zealand-The-Land-of-the-Long-White-Cloud.html
Totally cool. I get it now.

Really nice finds, Keith.

Now returning to the standard Panic Mode ^^

---

Apr 3, 1999 - May 3, 2020