processed dc and tc posts

[ric]

Originally Posted by sdbardwick

We matched.

Yup! Thanks for this, as well as for the many others TC’ed.

[Uncwilly]

List updated.

[LaurV]

Originally Posted by kriesel

@LaurV, if the 20000K fft deviating residues are reproducible in CUDALucas, please isolate it to the granularity gpuowl accepts on logging intervals (10,000) or finer.

Will do this, please give me a day or two.

[Prime95]

Originally Posted by ATH

How many exponents where all tests, 2 or more, were done with CUDALucas?

If I did the query correctly:

34643591
34696567
35184673
35381377
35478853
36142801
36211067
36313813
36497473
36532159
36717713
36841111
37018711
37047167
38093491
38208713
38276081
38363993
38931791
38976211
38976221
39052267
39258293
39839603
40123351
40404289
40413371
40473841
40501819
40641659
41508253
41518229
41856721
42791519
43883923
44932729
45243557
45285043
48073099
48075583
48122471
48429497
48555343
48677777
49404263
49457687
53998811
54009271
54010013
55831921
56294479
56309111
57766307
57954781
58370549
58370563
72366587
73604719
73612841
73614041
73642033
73684073
73684703
73685609
73798027
73802059
73812071
73901071
77075387
77143147
88680457
132000191
137362691
666666667

[ATH]

Hmm I'm involved in 23 of the 74 exponents. I'm triple checking my lowest one now with Prime95 30.8 b17: 36532159

[LaurV]

Hmm... they are by far not so many as I expected. I thought there are more of them, especially in 332M, where I did myself some, but probably those which were LL and DC by myself were killed by Madpoo with Prime95, already.

I could "owl-LL" all those, except the bigger ones. As George said, I would be surprised a lot, if the random shift wouldn't catch this bug (and disappointed a lot too , because there it goes to the drain my advocacy for random shift ).

On the other hand, meantime, on a 2080 Ti, Windows 10:

FFT = 20000k  (wrong)

|  Nov 14  16:48:00  | M332329111 121482200  0x72ac700df6edc14e  | 20000K  0.05566 267.1888    2.67s  |   6:04:40:47  36.55%  |

|  Nov 14  17:10:34  | M332329111 121482201  0x1d15bc664e50aa21  | 20000K  0.25000   1.#INF    0.03s  |   6:04:40:47  36.55%  |
|  Nov 14  17:10:34  | M332329111 121482202  0x495c10fac3cb687b  | 20000K  0.12500  37.6750    0.03s  |   6:04:40:48  36.55%  |
|  Nov 14  17:10:34  | M332329111 121482203  0x87be878e3f8a71ba  | 20000K  0.06250  36.5630    0.03s  |   6:04:40:48  36.55%  |
|  Nov 14  17:10:34  | M332329111 121482204  0xfa90c31f9f4db434  | 20000K  0.05371  36.3340    0.03s  |   6:04:40:49  36.55%  |
|  Nov 14  17:10:34  | M332329111 121482205  0xbe66bb2afd9a4d8a  | 20000K  0.05371  36.7410    0.03s  |   6:04:40:49  36.55%  |
|  Nov 14  17:10:34  | M332329111 121482206  0xd9db0fb42ccfebae  | 20000K  0.05103  31.4970    0.03s  |   6:04:40:50  36.55%  |

FFT = 19600k (correct, I mean, like gpuOwl, and like other FFTs I tried at this size)

|  Nov 14  16:55:52  | M332329111 121482200  0x72ac700df6edc14e  | 19600K  0.09570 271.3789    2.71s  |  37:04:05:54  36.55%  |

|  Nov 14  17:13:36  | M332329111 121482201  0x1d15bc664e50aa21  | 19600K  0.25000   1.#INF    0.03s  |  37:04:06:07  36.55%  |
|  Nov 14  17:13:36  | M332329111 121482202  0x495c10fac3cb687b  | 19600K  0.12500  39.0810    0.03s  |  37:04:07:09  36.55%  |
|  Nov 14  17:13:36  | M332329111 121482203  0x87be878e3f8a71ba  | 19600K  0.06250  36.9730    0.03s  |  37:04:08:06  36.55%  |
|  Nov 14  17:13:36  | M332329111 121482204  0xfa90c31f9f4db434  | 19600K  0.07324  36.5400    0.03s  |  37:04:09:01  36.55%  |
|  Nov 14  17:13:36  | M332329111 121482205  0xbe66bb2afd9a4d8a  | 19600K  0.07324  36.8750    0.03s  |  37:04:09:57  36.55%  |
|  Nov 14  17:13:36  | M332329111 121482206  0xb3b61f68599fd75e  | 19600K  0.06958  36.6120    0.03s  |  37:04:10:53  36.55%  |

All tests were run till the next checkpoint matched, to make sure it is not an error. I mean, next checkpoints on both branches (which were different as in the former post). Then, where the residues started to differ, the range split in 10 and ran again, full range, so the checkpoint at the end matches. When the split reached "1", I ran every branch twice to make sure it is not a hardware error.

Once we switched to smaller ranges, all tests were done with error checking for every iteration. No error catch.

I will share residue file(s) at 121482200 with George (cudaLucas can show every residue on screen, but the smallest granulation for checkpoints is 10, even if you set it to 1). I mean, it is no big secret, just they are 40MB+.

[kriesel]

Originally Posted by Prime95

If I did the query correctly:

So, no 100Mdigit exponents or thereabouts found with all LL tests by CUDALucas, for which 20000K would be a reasonable FFT size selection, and so make them suspect.

https://www.mersenne.org/report_ll/?...r_id=laurv&B1= shows lots of self-checking, probably gpuowl/CUDALucas pairs since they're nonzero and zero shift pairs, and an occasional different-user check in the verified exponents.

A gpuowl v6.11-380 result for M34643591 matched the previous two runs (the first is the outlier) https://www.mersenne.org/report_expo...4643591&full=1

[kriesel]

I've created a thread in my blog to track rechecks for the 74 flagged by George's query as CUDALucas-only verification, at https://mersenneforum.org/showthread.php?t=28237. Feel free to post claims & completions there, and I'll update post one of that thread as work progresses.
(I think a recheck of LaurV's monster will be a long time coming, but the rest look tractable.)

[frmky]

Just as a comment, those than I (Greg) did were run on nVidia Tesla hardware with ECC enabled both on the system and GPU memory. Doesn't make them error-proof but I would be very surprised if there's an error with matching residues with different shifts.

[LaurV]

Originally Posted by kriesel

I've created a thread in my blog to track rechecks for the 74 flagged by George's query as CUDALucas-only verification, at https://mersenneforum.org/showthread.php?t=28237. Feel free to post claims & completions there, and I'll update post one of that thread as work progresses.
(I think a recheck of LaurV's monster will be a long time coming, but the rest look tractable.)

Ok, I will "owl" everything under 70M in two Rvii cards, for a start. As Mihai did his own FFT, the combination of cudaLucas+gpuOwl should be tought-proof in any case. Colab is no-go, as it became too expensive. Later, if no taker for the larger ones, I may adventure into 7xM too, depending on how long it takes for the smaller ones, but for now, please reserve for me only what's under 70M, all of them.
Please ignore, I went to the thread, I see people already jumping into it, so I will reserve there, to avoid stepping on toes.

On the other hand, here is the link for the files with the residues as discussed in the former post. You should NOT waste time on it, unless your name is George, hihi, but if you have itching on the fingers, you can download and see if you can reconstruct the bug in your card (need nVidia card and cudaLucas). I don't think is card-dependent, but I would be curious if older cufft/cudart libraries or versions of cudaLucas show this behavior. Anyhow, this is minor importance, I don't think it will help, except satisfying my personal curiosity.

About the zip: there are 4 files. The last two (with the larger iteration number) are not useful, except to confirm that you get the same residue after 10 iterations. The first two are the same file, except the FFT size inside is set to 19600k and 20M respectively (you can binary compare them to see what is changed), because the test is so fast that you don't have time to press the "f/F" to change the FFT. But you can change the FFT if you want. You have to rename the starting file (lower-iteration count) to c332329111 into your cudaLucas folder and run a test for this expo. The test takes less than a minute. You have to do this only for the 20M FFT, the other is provided as witness only. To see anything on screen you may change the output every 1 iteration, and checkpoint every 10.

[Jan S]

Registered:

Cat 2
DoubleCheck=63406349,74,1
DoubleCheck=65028101,75,1

Cat 3
DoubleCheck=66654829,75,1
DoubleCheck=67070827,75,1