I understand the theory behind the Enhanced app - I'm saying if they take 30-40 times longer than the usual, that goal will be achieved, not only because WUs take longer, but also because fewer people will be willing to run them.

I am using optimised apps on the current SETI, with the 3.4GHz P4 doing a little over 2 WU per hour and even my 1.7GHz Celeron only taking 1.5 hours. Are you (KB7RZF) saying there are optimised apps for Enhanced that will go just as fast? If there are (or will be), perhaps I'll still be here when Enhanced hits.

@MJ: I'll keep it running for a bit longer and see what happens. Sounds a bit like Rosetta (except we're not moving in 10% steps)

---

*** Join the #1 Aussie Alliance on SETI ***

I am using optimised apps on the current SETI, with the 3.4GHz P4 doing a little over 2 WU per hour and even my 1.7GHz Celeron only taking 1.5 hours. Are you (KB7RZF) saying there are optimised apps for Enhanced that will go just as fast? If there are (or will be), perhaps I'll still be here when Enhanced hits.

@Yoda,

Unfortunately I don't know how the optimized apps are comparing for times, since I am not one of those optimizers. I highly doubt they will cut the times to what they are here. But I would expect them to be cut at least in the same percentage wise. Maybe one of the optimizer guys can jump in and relate what they have done so far as far as the times are concerned.

Jeremy

---

I understand the theory behind the Enhanced app - I'm saying if they take 30-40 times longer than the usual, that goal will be achieved, not only because WUs take longer, but also because fewer people will be willing to run them.

My experience with completed Enhanced work is more in the range of 10x longer. Is the 30-40 number you're thinking of in hours, rather than multiples of current work?

I am using optimised apps on the current SETI, with the 3.4GHz P4 doing a little over 2 WU per hour and even my 1.7GHz Celeron only taking 1.5 hours. Are you (KB7RZF) saying there are optimised apps for Enhanced that will go just as fast? If there are (or will be), perhaps I'll still be here when Enhanced hits.

I know there are optimized apps in the pipeline, but the people writing them are not going to release them until Enhanced is released. It kindof defeats the purpose of running a Beta if your testers aren't using the application you're trying to test!

@MJ: I'll keep it running for a bit longer and see what happens. Sounds a bit like Rosetta (except we're not moving in 10% steps)

I think it won't be as big a hit as you think. 8-)

MJ

---

My experience with completed Enhanced work is more in the range of 10x longer. Is the 30-40 number you're thinking of in hours, rather than multiples of current work?

It's a pro-rata: 25% done in 9.75 (CPU) hours, 100% would be 39 (CPU) hours if it were linear, but it sound like it will be less.

Edit: and that same CPU typically does a standard SETI in under an hour.

Still, if you're saying 10x longer than the standard SETI with the standard SETI app, it would get close to the 30 hours. I guess I'll find out tomorrow. Will let it run.

My AMD64 3700 (mobile AMD, 754 socket) estimates them at 16:53:57. I'm currently at 16:40:21, 57.66%, 14:12:52 remaining. It shouldn't take 30 hours on this machine.

PIII - 1000 - 284,515.91 seconds 79.3 hours 3.29 days - 190.71 claimed credits (pending)

The machine above has been taking about 10825 with an optimized science app, which means that the Beta/Enhanced work unit took somewhat over 25 times as long. However, if it weren't for the optimized application, the difference would be closer to 10-12 times as long ... the same machine was taking more than twice as long with the official app as it takes with the optimized.

Hope it helps in the way of comparison.

My experience with completed Enhanced work is more in the range of 10x longer. Is the 30-40 number you're thinking of in hours, rather than multiples of current work?

It's a pro-rata: 25% done in 9.75 (CPU) hours, 100% would be 39 (CPU) hours if it were linear, but it sound like it will be less.

Edit: and that same CPU typically does a standard SETI in under an hour.

Still, if you're saying 10x longer than the standard SETI with the standard SETI app, it would get close to the 30 hours. I guess I'll find out tomorrow. Will let it run.

You might want to look at the results turned in by Tetsuji Maverick Rai's hosts using his personal optimized applications. Crunch3r has also turned in some optimized results for his Athlon(tm) 64 Processor 3200+ host.

                                                    Joe

You might want to look at the results turned in by Tetsuji Maverick Rai's hosts using his personal optimized applications. Crunch3r has also turned in some optimized results for his Athlon(tm) 64 Processor 3200+ host.

                                                    Joe

Yes, that's why I wrote optimization worked great with _Enhanced. Notice my 2.8G HT P4 is almost twice as fast as 2.4G single thread P4 (2 wu's at once for about the same time, and none of my machines are overclocked). This is probably due to larger cache of 2.8G P4; 1mb vs. 512kb. The reported cache sizes on the host page are incorrect. My 2.8G P4, 2.4G P4 and 1.8G Celeron have 1MB, 512kb, 128kb caches respectively. My crunchers are not modified except that automatic memory size detection and graphics are disabled.

If any tweaks can be done (I have done as much as possible in official source), _Enhanced will be faster. and at the same time, we will have less (little) connection problems with _enhanced.

But beta's credits are not emitted to stats sites and my RAC is very low as in my sig :) Now I'm attaching only to beta...

---

Luckiest in the world. WMD = Weapon of Mass Distraction.
Click this table.

You might want to look at the results turned in by Tetsuji Maverick Rai's hosts using his personal optimized applications. Crunch3r has also turned in some optimized results for his Athlon(tm) 64 Processor 3200+ host.

                                                    Joe

Yes, that's why I wrote optimization worked great with _Enhanced. Notice my 2.8G HT P4 is almost twice as fast as 2.4G single thread P4 (2 wu's at once for about the same time, and none of my machines are overclocked). This is probably due to larger cache of 2.8G P4; 1mb vs. 512kb. The reported cache sizes on the host page are incorrect. My 2.8G P4, 2.4G P4 and 1.8G Celeron have 1MB, 512kb, 128kb caches respectively. My crunchers are not modified except that automatic memory size detection and graphics are disabled.

If any tweaks can be done (I have done as much as possible in official source), _Enhanced will be faster. and at the same time, we will have less (little) connection problems with _enhanced.

But beta's credits are not emitted to stats sites and my RAC is very low as in my sig :) Now I'm attaching only to beta...

The stats have been released, certainly available on Zain's and Willy's sites.

---

But beta's credits are not emitted to stats sites and my RAC is very low as in my sig :) Now I'm attaching only to beta...

That depends on which stats site you use. The XML feed was set up, and both BOINCStats and BOINC Synergy have SETI-Beta stats now 8-)

MJ

---

The stats have been released, certainly available on Zain's and Willy's sites.

Thanks. Didn't know that...

If any tweaks can be done (I have done as much as possible in official source), _Enhanced will be faster. and at the same time, we will have less (little) connection problems with _enhanced.

Thanks Tetsuji. I look forward to seeing the optimized apps. Without them, I'm not likely to keep crunching SETI when they switch to Enhanced. 1WU a day is just not enough to satisfy my appetite (and I'd get more credit running CPDN)

Thanks Tetsuji. I look forward to seeing the optimized apps. Without them, I'm not likely to keep crunching SETI when they switch to Enhanced. 1WU a day is just not enough to satisfy my appetite (and I'd get more credit running CPDN)

You do realize that, with the longer running results will come more credit per result? I think the goal is to have about the same credit per day as you'd be getting with the current. My last two credited results got 190.69 and 191.56 credits, and there are two more in that vicinity waiting for the third result to be returned.

MJ

---

I have to agree with WM Yoda, using standard Seti enhanced app is not going to bring in credits/day equal to either Einstein or CPDN. I could get about 240/day on Einstein if all my crunching went there for this computer, compared to 190 for 28+ hrs on enhanced (24/28 * 190 = 163) only about two thirds.

---

compared to 190

I believe there is still a "conversion factor" that has yet to be added... which SHOULD make it quite a bit more than 190 for that angle range.

You do realize that, with the longer running results will come more credit per result?

That's the theory, yes. In practice it's not that simple. I guess I'm spoilt, crunching about 50 SETI WU a day on a 3.4GHZ P4 with the Crunch3r optimised app and getting more credit than claimed on most occasions. Still, that same computer can get 600 or more credits a day on CPDN (~7 trickles a day, running two slab models)

My Enhanced WU is now at 35.03% after 48,482 seconds. Hope it gets to the quicker part soon, or it could still take over 38 hours total (in which time it could do over 40 regular SETI WU)

There seems to be so much speculation here as to what could happen when enhanced is released. IMO we must wait and see what actually happens before deciding/ranting about staying with/leaving the SETI project.

One big problem is trying to count WUs completed. If one tries to count WUs then whatever happens you're going to be dissapointed - but that's just why BOINC dropped the WU count in favor of credits.

I'm not saying that credits won't be affected by running enhanced, I really don't know how they will change. In a perfect world, they won't change but change they will. A given machine earning say 500 credits a day now will get 500 +/- X. IMO 'sucess' will be measured by the size of X.

We'll only know X once a significant number of hosts run enhanced and we see some real RAC figures - those from the beta project are too few to be meaningful.

Also, IMO when X is known, there would always be a (Project Managers) option to add a correction factor. I think Berkeley are aware of participants motivations, and are not planning to upset any large percentage of the participants whose motivation is credit.

---

Now that I'f finished my firs 5.02 enhaced usint 112,822.54 seconds => 31:20:22.54
191.63 credits

or ~6 credits/hour

on the same machine

Seti gets me 24,381.37 seconds => 6:46:21.37
20.19 credit

or ~3credit/hour

---

uninstall dyslexica.o : Permission denied


AMD Athlon 64 3000+ w/Windows
AMD Athlon 1800+ w/Linux

Now that I'f finished my firs 5.02 enhaced usint 112,822.54 seconds => 31:20:22.54
191.63 credits

Is that claimed credit or granted credit?

or ~6 credits/hour

on the same machine

Seti gets me 24,381.37 seconds => 6:46:21.37
20.19 credit

or ~3credit/hour

Double credits for enhanced sounds good on the face of it, but it should not happen. It would spell death for other less generous projects as far as credit counters are concerned.

---

Now that I'f finished my firs 5.02 enhaced usint 112,822.54 seconds => 31:20:22.54
191.63 credits

Is that claimed credit or granted credit?

Only Claimed people are only just starting to claim credit on beta, with a small userbase it take longer to Grant credit, but that simply a userbase issue

[I chose "current seti" unit which claimed and granted was the same]

BUT remeber the bit this with Enhance is the Credit is much more depended on unit contents than it is on the machine chruniching it. the machine will vary how long the crunch time is.

---

uninstall dyslexica.o : Permission denied


AMD Athlon 64 3000+ w/Windows
AMD Athlon 1800+ w/Linux