Workunit 2581227344 (09no16aa.18442.2116.6.33.31)
Task 5821860064 (S=3, A=2, P=1, T=0, G=3) x41p_zi3t2b, Cuda 8.00 special
Task 5821860065 (S=3, A=2, P=1, T=0, G=3) v8.22 (opencl_nvidia_SoG) windows_intelx86

Cuda 8.00 special - Best gaussian: peak=6.385563, mean=0.5961245, ChiSq=1.203035, time=67.95, d_freq=1420305212.47,
score=2.122812, null_hyp=2.212901, chirp=-62.462, fft_len=16k
v8.22 SoG - Best gaussian: peak=5.745589, mean=0.583622, ChiSq=1.414207, time=59.56, d_freq=1420299457.13,
score=2.057846, null_hyp=2.326043, chirp=90.136, fft_len=16k

Got this one.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

I just posted the ones I've been using here, MBv8_Test-WUs.7z
Also, the OpenCL links in case someone else wants to run them in Linux;
http://boinc2.ssl.berkeley.edu/beta/download/setiathome_8.22_x86_64-pc-linux-gnu__opencl_nvidia_SoG
http://boinc2.ssl.berkeley.edu/beta/download/setiathome_8.22_x86_64-pc-linux-gnu__opencl_nvidia_sah
http://boinc2.ssl.berkeley.edu/beta/download/MultiBeam_Kernels_r3584.cl

Workunit 2573263722 (23se08ac.6875.22968.6.33.135)
Task 5805117074 (S=3, A=0, P=1, T=3, G=0) x41p_zi3t2b, Cuda 8.00 special
Task 5805117075 (S=3, A=0, P=1, T=3, G=0) v8.22 (opencl_nvidia_SoG) windows_intelx86

Cuda 8.00 special - Best gaussian: peak=3.252388, mean=0.5397108, ChiSq=1.344394, time=14.26, d_freq=1418816790.11,
score=-1.169299, null_hyp=2.144445, chirp=-39.071, fft_len=16k
v8.22 SoG - Best gaussian: peak=3.76217, mean=0.5480909, ChiSq=1.226871, time=39.43, d_freq=1418822660.68,
score=-1.169124, null_hyp=2.078196, chirp=43.425, fft_len=16k

This one should also still be available, for now.

EDIT: Looks like TBar beat me to it.

Got them too.
No linux GPU host availble currently so will try with Windows ATi SoG.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

The Benchmark App shows the Non-SoG App producing the correct result;

[size=9][pre]====================================================================
Current WU: 09no16aa.18442.2116.6.33.31.wu
----------------------------------------------------------------
Skipping default app MBv8_8.0r3305_ssse3_x86_64-pc-linux-gnu, displaying saved result(s)
Elapsed Time: ....................... 6283 seconds

Can I get res files from this run, please?

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

They're now at the same Place, Results_WU- 09no16aa.18442.2116.6.33.31.7z
You'll notice the Non-SoG App failed on 23se08ac.6875.22968.6.33.135, it passed on the others.

I'm almost done running 2000 tasks with x41p_zi3v on beta and have never seen the inconclusive rate go above 1.2% with invalid tasks remaining at 0. I would say that it passes with honors.

Does x41p_zi3v contain bugfix for <wrong Pulse selecting as reportable> issue?

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

Does x41p_zi3v contain bugfix for <wrong Pulse selecting as reportable> issue?

I could be corrected, but I believe it contains the shunt/workaround I recommended to serialise the race condition, though Petri implemented it and I haven't had a chance to examine it. Word is that it worked though, so validation characteristic should be more or less identical to Older Cuda variants.

---

"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.

I'm almost done running 2000 tasks with x41p_zi3v on beta and have never seen the inconclusive rate go above 1.2% with invalid tasks remaining at 0. I would say that it passes with honors.

It looks good, with one puzzling problem. Our machines ran the same task with the same App and yours found One less Pulse, https://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=9831637. I've checked one of those other tasks you had against a CPU and my CPU says it's your machine missing a Pulse. The only difference I know of is you're using the -pfb 64 setting. Try removing that setting and see if you still get the occasional One less Pulse. Too bad there aren't more people running the App at Beta.

Here's another one of those Bad Best Gaussians, https://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=9832304

Again...the Plot Thickens.

This task, https://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=9832304
CUDA x41p_zi3v - Best gaussian: peak=4.661562, mean=0.5658302, ChiSq=1.327004, time=32.72, d_freq=1420757670, score=1.040808, null_hyp=2.254107, chirp=-4.8404, fft_len=16k
Win32 Build 3584 - Best gaussian: peak=4.340054, mean=0.5760618, ChiSq=1.40025, time=34.39, d_freq=1420757661.88, score=0.4676737, null_hyp=2.267228, chirp=-4.8404, fft_len=16k
SSE4.1xjf OS X Build 3344 - Best gaussian: peak=4.661561, mean=0.5658303, ChiSq=1.327021, time=32.72, d_freq=1420757670, score=1.040918, null_hyp=2.254123, chirp=-4.8404, fft_len=16k

But...But...This is a Windows ATI App! Yes, but it says SoG...
Run it Yourself, http://boinc2.ssl.berkeley.edu/beta/download/2f3/04oc08ab.19755.13978.13.47.78

*Shrugs*

Hey TBar

Got 2 machines on Beta running the zi3v, will get a 3rd over shortly, they consist of a GTX 980, 960 & 950.

I forgot about changing the version number in the app_info, but I have gotten 8 tasks down and 4 cruched already, so who knows what happened there.

You don't have to download the cuda libs again for those on the Special app already, you will have those on your computer already, just paste into the beta folder.

But then most on here would have thought of that already.

Cheers, I will run as many as I can.

Mark

Well, it may not be just the Special App vs. SoG where Best Gaussian differences are showing up. To make it easier and quicker for me to eyeball likely candidates, I decided to add Best Gaussian to the data I extract in generating my Inconclusives list (see BG= in the WUs, below). I found a few surprises, as you can (hopefully) see in the first couple examples.

Workunit 2537087160 (23no08ad.22055.7434.9.36.159)
Task 5729514952 (S=0, A=2, P=0, T=0, G=1, BG=3.829226) v8.05 windows_x86_64
Task 5729514953 (S=0, A=2, P=0, T=0, G=1, BG=3.376349) SSE3xj Win32 Build 3584

Workunit 2557372653 (27my08ag.22553.5798.13.40.11)
Task 5771884489 (S=0, A=5, P=3, T=2, G=1, BG=3.841448) v8.05 windows_x86_64
Task 5771884490 (S=0, A=5, P=3, T=2, G=1, BG=3.480158) v8.22 (opencl_nvidia_SoG) windows_intelx86

Workunit 2583000026 (24no16aa.11418.885.7.34.119)
Task 5825591817 (S=0, A=1, P=0, T=7, G=1, BG=5.74053) x41p_zi3t2b, Cuda 8.00 special
Task 5825591818 (S=0, A=1, P=0, T=7, G=1, BG=5.447097) v8.22 (opencl_nvidia_SoG) windows_intelx86
Task 5826607618 (S=17, A=11, P=2, T=0, G=0, BG=5.531063) v8.22 (opencl_nvidia_SoG) windows_intelx86
(The 3rd host here obviously has other problems, but the first two are what caught my eye.)

Workunit 2583437545 (24oc16aa.31883.5384.10.37.92)
Task 5826514855 (S=16, A=4, P=0, T=0, G=1, BG=3.85572) x41p_zi3t2b, Cuda 8.00 special
Task 5826514856 (S=16, A=4, P=0, T=0, G=1, BG=3.685971) SSE3xj Win32 Build 3557

Workunit 2583532710 (30oc16aa.8195.4161.5.32.175)
Task 5826713486 (S=0, A=1, P=0, T=0, G=0, BG=8.930679) v8.22 (opencl_nvidia_SoG) windows_intelx86
Task 5826713487 (S=0, A=1, P=0, T=0, G=0, BG=7.704001) x41p_zi3t2b, Cuda 8.00 special

Workunit 2583881219 (29oc16aa.21186.511408.9.36.188)
Task 5827437143 (S=1, A=2, P=0, T=0, G=0, BG=5.375575) v8.22 (opencl_nvidia_SoG) windows_intelx86
Task 5827437144 (S=1, A=2, P=0, T=0, G=0, BG=5.020687) x41p_zi3t2b, Cuda 8.00 special

One thing that I think I'm noticing is that when there is a reported Gaussian, that peak will match the Best Gaussian peak in SoG. However, in the other apps, the Best Gaussian will have a higher peak than the reported Gaussian. Perhaps there's some significance there. Or perhaps not. :^)

...
One thing that I think I'm noticing is that when there is a reported Gaussian, that peak will match the Best Gaussian peak in SoG. However, in the other apps, the Best Gaussian will have a higher peak than the reported Gaussian. Perhaps there's some significance there. Or perhaps not. :^)

Hmmm, I am seeing svn commits after January on stock CPU multibeam, though nothing immediately stands out as affecting best/reportable policy. Will do some trawling of the codebases.

[Edit:] Superficially looks like certain incompatible changes to Gaussian best reporting were integrated circa 2011, By Eric from AK code, then 'bugfixed' a revision later. So will need more digging, but there could be as many as 3 different best gaussian reporting variants floating about. Will check the baseline and special CUDA variants against the 'bugfixed' variant logic. Ideally we'd want to match stock CPU logic, therefore if any number of applications require updates, will probably have to happen. Probably the actual intent of the change will need to be looked at, as it may possibly have holes in it. At the moment it appears as though the reportable gaussian should not be being used to update best if it isn't very gaussian-ey, though the intent is unclear to me on the first pass (a possible red flag)

---

"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.

Well, it may not be just the Special App vs. SoG where Best Gaussian differences are showing up. To make it easier and quicker for me to eyeball likely candidates, I decided to add Best Gaussian to the data I extract in generating my Inconclusives list (see BG= in the WUs, below). I found a few surprises, as you can (hopefully) see in the first couple examples.

Yes, I believe I mentioned earlier how the SoG tasks were failing against the CPUs with the Best Gaussian, it's on the last page. There is Currently a Brand New Windows CPU App at Beta and a number of Machines with Dozens of Cores a piece. They're quiet right now, but were running a little while ago, there is still a number of New CPU tasks around. During my 2200 recent CUDA tasks run, many of them were against the New CPU App. I think I saw maybe one or two inconclusives against the CPUs at Beta, and if you look right now, I don't have a single Inconclusive against a CPU. The New CUDA App gets along Very Well with the New CPU App at Beta, which is very nice. I always said just make it work against the CPU and the rest will shakeout in the end.
I'm still waiting on the links in that old post,https://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=9809103 & https://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=9796907

Thanks tazzduke, it's good to see a few more machines running the App at Beta.

Have Verified Cuda baseline matches Joe Segur's Bugfix/changes to stock best scoring

Revision: 1146
Author: korpela
Date: Wednesday, 17 August 2011 7:41:35 AM
Message:
- Fix to bug introduced in last change of gaussfit.cpp. Much of the new code
is from the AK8 branch.
- Version number to 6.97

----
Modified : /branches/sah_v7/seti_boinc/client/gaussfit.cpp
...

Coming from Baseline Petri's special should match that logic (to check). The original modification this fixes, by Joe Segur, contains comments by Raistmer, and appears to be from an AK commit by Raistmer (also committed via Eric) sometime before. I'm unfamiliar with the intent, as mentioned, though Joe's comments on the code seem reasonable.

// Gauss score used for "best of" and graphics.
// This score is now set to be based upon the probability that a signal
// would occur due to noise and the probability that it is shaped like
// a Gaussian (normalized to 0 at thresholds). Thanks to Tetsuji for
// making me think about this. The Gaussian has 62 degrees of freedom and
// the null hypothesis has 63 degrees of freedom when gauss_pot_length=64;
//JWS: Calculate invariant terms once, ala Alex Kan and Przemyslaw Zych

Probably each branch will need checking, in case a lack of this change propagated from AK sources into other builds.

---

"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.

Yes, I believe I mentioned earlier how the SoG tasks were failing against the CPUs with the Best Gaussian, it's on the last page. There is Currently a Brand New Windows CPU App at Beta and a number of Machines with Dozens of Cores a piece. They're quiet right now, but were running a little while ago, there is still a number of New CPU tasks around. ....

Maybe a new call for testing CPU apps is a good idea New binary to test on beta to get more CPUs running.

I fired mine back up.

Okay, without upsetting anyones apple cart lol, I have at least 9 cores I could run the BETA cpu on, being Core2Quad 9400;s though.

Just need a link to the CPU executable please, as I seem to have lost mine.

Regards

Here you go.
Linux Seti@Home apps

I fired up my CPUs as well, but realistically I can only do ~6 CPU vs >100 GPU tasks a hour.

Greetings Brent

Thankyou, I now have one machine up, the other 2 will be up shortly.

Well the old q9400s are getting on as well, but will see how they do in the meantime lol.

Regards