Hi Hans:

That definitely proves beyond a doubt that there is something wrong with the way the Seti app 5.12 is calculating credit whether Crunch3r’s 51.12 or Seti’s standard 5.12.

When I received an average of of 48.20 cs for 13,976.81 seconds with TruXoft’s 4.2.13 BOINC client and Crunch3r’s 5.12 Seti app I was and still am extremely pissed off. If I had received the claimed 58.87 cs for 54,515.85 seconds as you just did I think I would have thrown my computer out the window and cursed Seti to no end.

Based on my above average, you should have received about 229 cs instead of the messily 58.87 cs per work unit. Based on the old TruXoft’s 4.2.13 BOINC client and Crunch3r’s 4.11 Seti app you should have received about 1,030 cs per work unit. Sorry, I hope I didn’t twist the knife to hard.

The reasons to downgrade to TruXoft’s 4.2.13 BOINC client and Crunch3r’s 5.11 Seti app are becoming more appealing by the second.

So, is your recommendation still to stick with the standard BOINC client 5.4.9 and the Seti 5.12 app?

Regards Franz

Agreed.

My recommendation was made because I thought it would be a good idea if all participants would claim the same amount of credit for the same WU.

After 2 days of reading this board, I just don't care anymore...

Run whatever combination pleases you the most :o)


Regards Hans

---

Here are 2 WUs that took me extremely long to complete:

No.1, ar=0.002535 and No.2, ar=0.002813

I'm using Crunch3rs 5.12 linux client.

Regards Hans

Hi Hans:

That definitely proves beyond a doubt that there is something wrong with the way the Seti app 5.12 is calculating credit whether Crunch3r’s 51.12 or Seti’s standard 5.12.

When I received an average of of 48.20 cs for 13,976.81 seconds with TruXoft’s 4.2.13 BOINC client and Crunch3r’s 5.12 Seti app I was and still am extremely pissed off. If I had received the claimed 58.87 cs for 54,515.85 seconds as you just did I think I would have thrown my computer out the window and cursed Seti to no end.

Based on my above average, you should have received about 229 cs instead of the messily 58.87 cs per work unit. Based on the old TruXoft’s 4.2.13 BOINC client and Crunch3r’s 4.11 Seti app you should have received about 1,030 cs per work unit. Sorry, I hope I didn’t twist the knife to hard.

The reasons to downgrade to TruXoft’s 4.2.13 BOINC client and Crunch3r’s 5.11 Seti app are becoming more appealing by the second.

So, is your recommendation still to stick with the standard BOINC client 5.4.9 and the Seti 5.12 app?

Regards Franz

Here are 2 WUs that took me extremely long to complete:

No.1, ar=0.002535 and No.2, ar=0.002813

I'm using Crunch3rs 5.12 linux client.

Regards Hans

---

I'd recommend to use Crunch3r's 5.12 version. It's been fixed to claim the same amount of credit as the official enhanced application.
Please use the current BOINC client, too.

Regards Hans

Hi Hans:

Thank you for your recommendation to stick with the standard BOINC client 5.4.9 and the Seti 5.12 app. My question is why???

From my initial look into the performance of Seti Enhanced it shows me that it takes my computer 4.49 times the time to complete a work unit and only increased my granted credit by 1.87 times compared to TruXoft’s 4.2.13 BOINC client and Crunch3r’s 4.11 Seti app.

When I downgraded to TruXoft’s 4.2.13 BOINC client and Crunch3r’s 5.12 Seti app there was an increase of 5.9% in performance in terms of computer time.

In response to your request, I have postponed downgrading to Crunch3r’s 5.11 Seti app. This has given me some addition time to process additional work units. While reviewing the results, I noticed the following inconsistencies:
16,849.14 sec. = 50.23 credits granted,
17,283.69 sec. = 50.24 credits granted,
12,498.41 sec. = 54.25 credits granted,
14,232.00 sec. = 61.88 credits granted,
13,778.61 sec. = 61.87 credits granted,
10,004.52 sec. = 31.28 credits granted.

From the above, I have to conclude that the people responsible for the new credit system still have not gotten it right.

Franz

In one of the other threads the devs have said they will look at the credit calculation again when they have data from lots more computer/OS and ar spreads.

Andy

I'd recommend to use Crunch3r's 5.12 version. It's been fixed to claim the same amount of credit as the official enhanced application.
Please use the current BOINC client, too.

Regards Hans

Hi Hans:

Thank you for your recommendation to stick with the standard BOINC client 5.4.9 and the Seti 5.12 app. My question is why???

From my initial look into the performance of Seti Enhanced it shows me that it takes my computer 4.49 times the time to complete a work unit and only increased my granted credit by 1.87 times compared to TruXoft’s 4.2.13 BOINC client and Crunch3r’s 4.11 Seti app.

When I downgraded to TruXoft’s 4.2.13 BOINC client and Crunch3r’s 5.12 Seti app there was an increase of 5.9% in performance in terms of computer time.

In response to your request, I have postponed downgrading to Crunch3r’s 5.11 Seti app. This has given me some addition time to process additional work units. While reviewing the results, I noticed the following inconsistencies:
16,849.14 sec. = 50.23 credits granted,
17,283.69 sec. = 50.24 credits granted,
12,498.41 sec. = 54.25 credits granted,
14,232.00 sec. = 61.88 credits granted,
13,778.61 sec. = 61.87 credits granted,
10,004.52 sec. = 31.28 credits granted.

From the above, I have to conclude that the people responsible for the new credit system still have not gotten it right.

Franz

As a follow up to my previous post, by switching back to TruXoft’s 4.2.13 BOINC client and Crunch3r’s 5.12 Seti app from the new 5.4.9 BOINC client and the new 5.12 Seti app the following average results were obtained for 11 work units.

The CPU time to process a wu dropped from 13,976.81 sec. to 13,151 sec. real_cpu_time an increase of 5.9% in performance. The reported corrected_cpu_time was 13,889.09 sec. only a 0.6% difference.

There was no improvement in the claimed credit and the granted credit since everyone else using the new Seti version was within +/- 1 credit point. Unless you were using an older BOINC client such as 4.25, 4.43 and 4.45 then you were off by quite a bit.

Since there was no dramatic improvement in claimed credit I will now revert to Crunch3r’s 5.11 Seti app and see what happens.

Franz

BTW: To whoever deleted my previous post that was not very democratic. Whatever happened to freedom of speech? As a note, my posts are saved in MS Word and can be reposted at any time as this repost clearly demonstrates.

Hi Franz,

I'd recommend to use Crunch3r's 5.12 version. It's been fixed to claim the same amount of credit as the official enhanced application.
Please use the current BOINC client, too.

Regards Hans

---

As a follow up to my previous post, by switching back to TruXoft’s 4.2.13 BOINC client and Crunch3r’s 5.12 Seti app from the new 5.4.9 BOINC client and the new 5.12 Seti app the following average results were obtained for 11 work units.

The CPU time to process a wu dropped from 13,976.81 sec. to 13,151 sec. real_cpu_time an increase of 5.9% in performance. The reported corrected_cpu_time was 13,889.09 sec. only a 0.6% difference.

There was no improvement in the claimed credit and the granted credit since everyone else using the new Seti version was within +/- 1 credit point. Unless you were using an older BOINC client such as 4.25, 4.43 and 4.45 then you were off by quite a bit.

Since there was no dramatic improvement in claimed credit I will now revert to Crunch3r’s 5.11 Seti app and see what happens.

Franz

BTW: To whoever deleted my previous post that was not very democratic. Whatever happened to freedom of speech? As a note, my posts are saved in MS Word and can be reposted at any time as this repost clearly demonstrates.

Another Celeron/Win 98 result, for WU Nº77758096:
AR 0.7251, 7.16 hours, 33.20 credits granted.

Doom-and-gloomers please note that a v4.x BOINC client in the quorum, going by time-&-benchmarks instead of Flops, claimed almost 20% less credit than the current ones, showing that the new system can sometimes yield more cobblestones than the old.

---

[oops]

---

Boinc....Boinc....Boinc....Boinc....

This reinstall of 5.11 brings up the issue of whether the Seti project intends to enforce the FPOP count.

The question: When a WU is returned and has a vastly different FPOP count (and I mean larger) will the Seti project consider that WU invalid in terms of credit eligibility.

The new FPOP scheme should produce WU results that should return FPOP counts that are nearly identical for the same WU regardless of the hardware/software used to process the WU. A greatly different FPOP would seem to be equivalent to credit cheating unless there was some explainable hardware/software failure.

It is not sufficient to respond with the answer that it will be handled by the Quorum - that is not the purpose of the new FPOP scheme.

I have yet to see a statement of intent regarding this issue from the Seti project.

I'm pretty sure the enforcement of Fpops will go ahead and when that happens anybody using ver 4.11 will become immediately apparent and the servers will be able to adapt their granted credit to a justified amount, i.e. 0.

Andy

This reinstall of 5.11 brings up the issue of whether the Seti project intends to enforce the FPOP count.

The question: When a WU is returned and has a vastly different FPOP count (and I mean larger) will the Seti project consider that WU invalid in terms of credit eligibility.

The new FPOP scheme should produce WU results that should return FPOP counts that are nearly identical for the same WU regardless of the hardware/software used to process the WU. A greatly different FPOP would seem to be equivalent to credit cheating unless there was some explainable hardware/software failure.

It is not sufficient to respond with the answer that it will be handled by the Quorum - that is not the purpose of the new FPOP scheme.

I have yet to see a statement of intent regarding this issue from the Seti project.

That person ran that result with crunch3r 5.11.

Ah, I didn't notice that; thanks for pointing it out.

Crunch3r 5.12 has been on offer for some days now, and is reported to give claims closely comparable to the 5.12 distributed by the project. I think crunch3r 5.11 inherited its claiming scores from the project--which changed its mind on that subject between 5.11 and 5.12.

IMO publically releasing the Beta-based version for use here was completely irresponsible, especially in light of the official developer’s statement that the discrepant Flop-weighting factors in certain Windows builds were due to compiling accidents and not at all intentional. Over in Beta it’s no big deal, because the apps get automatically updated fairly often, but here is quite another kettle of fish, as it requires user intervention to replace an optimized worker. What incentive is there for users running the 5.11 app, with its credit claims inflated by a factor of 9/3.35 ~= 2.7 (IIRC from discussions in the Beta forum), to update to a current one?

---

[quote]WU Nº77539681:

I think crunch3r 5.11 inherited its claiming scores from the project--which changed its mind on that subject between 5.11 and 5.12.

With my P4 650 HT at 3.4 GHz and 2 GB RAM using TruXoft’s 4.2.13 Boinc client and Crunch3r’s 4.11 Seti app, I was averaging 3,110.58 sec./wu, a claimed credit of 32.16 and a granted credit of 25.76. Based on the last 96 - 4.18 good wus completed. After installing the new Boinc client 5.4.9 and the Seti 5.12 app, I averaged 13,976.81 sec./wu, a claimed credit of 50.23 and a granted credit of 48.20. Based on 10 – 5.12 wus completed.

This represents a 4.49x increase in computer time per work unit with only a 1.67x increase in claimed credit and only a 1.87x increase in granted credit. Secondly, with everyone getting the same credit for each work unit, we have come full circle back to Seti Classic where everyone received 1 credit per work unit.

Therefore, I have deleted the 5.4.9 BOINC client along with the standard 5.12 Seti app and have reinstalled TruXoft’s 4.2.13 Boinc client and both Crunch3r’s 4.11 and 5.12 Seti apps. With only 4.11 installed the server wouldn’t download any wus so I added Crunch3r’s 5.12 and received numerous 5.13 wus and a hand full of 4.18 wus.

Hopefully the claimed credits will go back up to reflect the actual time spent per wu as mentioned by Odysseus otherwise, I will downgrade further to Crunch3r’s 5.11 Seti app as mentioned by Archae86.

I’ve already uploaded 8 work units with the TruXoft/Crunch3r combination but the computer hasn’t actually reported them yet. Until then, I won’t be able to see the results on my results page at Seti. (Yes, uploading and reporting are 2 separate operations not necessarily carried out at the same time by your computer.)

Many thanks to Crunch3r, TruXoft ,Tetsuji Maverick Rai, Hans Dorn, Harold Naparst, and Ned Slider for their hard work. Without their efforts Seti wouldn’t be doing twice the science today or be as much fun crunching.

Franz

---

WU Nº77539681:

Only one other result submitted on that one so far, from a Crunch3r/TruXoft combination; it's claiming nearly 124 credits. I thought that kind of discrepancy wasn't supposed to happen any more in Enhanced! We'll see what the quorum says ...

That person ran that result with crunch3r 5.11. Crunch3r 5.12 has been on offer for some days now, and is reported to give claims closely comparable to the 5.12 distributed by the project. I think crunch3r 5.11 inherited its claiming scores from the project--which changed its mind on that subject between 5.11 and 5.12.

---

A pending result from the Celeron/Win 98, WU Nº77539681:
AR 0.6471, 22.3 hours, claiming 46.02 credits.

Only one other result submitted on that one so far, from a Crunch3r/TruXoft combination; it's claiming nearly 124 credits. I thought that kind of discrepancy wasn't supposed to happen any more in Enhanced! We'll see what the quorum says ...

---

609850 secs on a GX1@300Mhz :o)

Regards Hans

---

My first validated Enhanced result for Windows 98 (SE), on a 1.7-GHz (?) Celeron running Boinc Manager 5.2.13: WU Nº77408272
AR 0.5957, 9.51 hours, 49.69 credits granted.

No apparent timer problems (although I notice this machine almost never reports fractions of a second, on any project: a Win9x thing?). Credit per hour similar to that earned with the stock 4.18 app on this machine, maybe 10% lower. The other two claims in the quorum were identical, one of them from BOINC v5.2.8.

I also have pending from a 2-GHz Athlon, Win XP Pro, BM 5.2.13: WU Nº77037158
AR 0.4272, 10.1 hours, 62.25 credits claimed by this host and one other.

---

I'll let you know soon, my cashe has 12 of the 4.18 versions left, and about 15 of the enhanced version units. I currently complete one of the 4.18 versions between 2.0 hours and 2.2 hours...so we will see what happens......

Just started one of the new enhanced work units...results to follow soon I hope

Looks like 4.0 hours for the first one. I have noticed that the top screen where the graph searches for triplets and gausens seems to move quicker, but the lower screen is about the same. I also noticed that the entire screen saver application pauses every 30 seconds or so for about 1-2 seconds. Here is a breakdown of the hardware in use on this machine:

Motherboard Asus SK8V
Processor AMD Athlon 64 Fx 51
Memory Kingston 333 Regist DDR @ 1 Gig (2-512 DDR cards)
HardDrive Western Digital 80 Gig 7200rpm 8Meg Memory Cashe
Video Card ATI Radeon 9200 with 128 Megs DDR
Op Sys Windows Xp Pro, SP 2 fully updated.

Second enhanced workunit took 3.0 hours, so I guess it varies.

---

I'll let you know soon, my cashe has 12 of the 4.18 versions left, and about 15 of the enhanced version units. I currently complete one of the 4.18 versions between 2.0 hours and 2.2 hours...so we will see what happens......

Just started one of the new enhanced work units...results to follow soon I hope

Looks like 4.0 hours for the first one. I have noticed that the top screen where the graph searches for triplets and gausens seems to move quicker, but the lower screen is about the same. I also noticed that the entire screen saver application pauses every 30 seconds or so for about 1-2 seconds. Here is a breakdown of the hardware in use on this machine:

Motherboard Asus SK8V
Processor AMD Athlon 64 Fx 51
Memory Kingston 333 Regist DDR @ 1 Gig (2-512 DDR cards)
HardDrive Western Digital 80 Gig 7200rpm 8Meg Memory Cashe
Video Card ATI Radeon 9200 with 128 Megs DDR
Op Sys Windows Xp Pro, SP 2 fully updated.

---

I'll let you know soon, my cashe has 12 of the 4.18 versions left, and about 15 of the enhanced version units. I currently complete one of the 4.18 versions between 2.0 hours and 2.2 hours...so we will see what happens......

Just started one of the new enhanced work units...results to follow soon I hope

---