[QUOTE=nucleon;295524]I selected oldest exponent and it allocated based on lowest TF level. I can't rule out user error. I went back button in browser and it showed oldest.[/QUOTE]

The "select" was sorting by lowest TF level first, and then age. Changed to be solely dependent upon age.

[QUOTE=chalsall;295525]The "select" was sorting by lowest TF level first, and then age. Changed to be solely dependent upon age.[/QUOTE]

Cheers.

On the View Assignments page, the table at the top shows the overall average from start to date, which for the most part is fine, as LL/P-1 take a lot of time compared to TF. On TF though, if someone makes a system change, it takes a long time to get close to current capability, and mine shows ~1/2 of my current. Maybe it should show a 10 or 15 day average?

Heh, regarding my P-1 factors, I apparently found two back-to-back overnight, but of course they were both relatively small (76.x and 81.x).


Edit:

[url]http://mersenne.info/exponent_status_line_graph/1/0/[/url]
See the bump? :smile:

[QUOTE=chalsall;295455]
And as some might have noticed, we're now effectively out of all work except 71 -> 72.[/QUOTE]

That's a pity, because 2 of my GPUs are slow and old, with only Compute Capability 1.1. Therefore, they get ineffective above 71. Unfortunately, I forgot how ineffective.

At the moment a few exponents are available below 71, but I get this nasty
[COLOR=red]* Note: Minumum is 72.[/COLOR]
Their contribution is about 25 GHzDays/day, or it was.
I think it was useful to factor these exponents up to 71, and then someone else takes them to 72 or 73.

I could pledge to factor them to 72 and factor them to only 71.
Or I could switch to the DC range, which is far ahead of the "wave", or to CUDALucas.

Do you have better suggestions?

Thanks,
Rad

You might consider DCTF, they're at lower bit levels.

DCTF is not very useful, because it is so far ahead.
I could do the part up to 71 in the slow GPUs, and the remainder to 72 or 73 in a fast new GPU, but that is complicated.
Perhaps CUDALucas is the most useful option.

[QUOTE=Radikalinsky;295634]DCTF is not very useful, because it is so far ahead.
I could do the part up to 71 in the slow GPUs, and the remainder to 72 or 73 in a fast new GPU, but that is complicated.[/QUOTE]

I understand your concern. And I only did the "Minumum pledge is 72" after effectively everything below 60M was already taken to 71. Yes, a few below 71 trickle out every day, but it is a very small number and I want to see them processed in one assignment cycle (or two, if P-1 is needed).

[QUOTE=Radikalinsky;295634]Perhaps CUDALucas is the most useful option.[/QUOTE]

Unfortunately your cards can't run CUDALucas.

One thing I could do for people like you who have cards like you have is to bring in a few candidates at 60M which are already at 70, and make them available to only go to 71.

They're well ahead of the "wave", but the work would be more beneficial to the project than DCTF (in my opinion).

Thoughts?

[QUOTE]One thing I could do for people like you who have cards like you have is to bring in a few candidates at 60M which are already at 70, and make them available to only go to 71.[/QUOTE]

Thank you, this sounds more useful than DCTF.

[QUOTE=Radikalinsky;295641]Thank you, this sounds more useful than DCTF.[/QUOTE]

OK. A quick hack before I run out for a "Sundowner"...

If you go to the LLTF form and choose the new option "I have a slow card!!!" it will ignore the range and pledge fields, and issue you work between 60M and 61M to TF from 70 to 71.

[QUOTE=chalsall;295512]
No. We've got a large batch of TF results without P-1 expected to complete shortly, so I threw 600 back to PrimeNet to hand out as it feels is appropriate.[/QUOTE]
Hmm... now we're down to 63... :unsure:

While I'm at it, how did this happen?
[url]http://gpu72.com/reports/workers/ll/[/url]
[code] # Worker Own 68 69 70 71 72 73 >73 Facts Saved Done
1 Bdot 48 49 3,557
2 prime95 70 45 3,268
3 kcjaget 16 30 2,179
4 1997rj7 12 27 1,958
5 Pete 7 25 1,819
6 Chuck 7 18 1,304
7 LaurV 20 14 1,014
8 Jerry Hallett 20 13 859
9 rduerr 33 11 [U][B]1[/B][/U] [U][B]144[/B][/U] 820
10 markr 2 10 725
11 kdgehman 4 10 725
12 delta_t 6 9 655[/code]

Edit: Since I have the post, has anybody else noticed how creepily even the total LL/DC throughput has stayed, from a work/GHz-Days perspective? I know many of us are doing just one or the other, so it's remarkable how close they are. (At the time of this edit, the totals were 22,503/22,480 respectively.)

[QUOTE=Dubslow;295789]While I'm at it, how did this happen?[/QUOTE]Since it wasn't immediately obvious to me due to scrolling of the code box, I assume you mean "how did [i]rduerr[/i] find a factor with LL"?

[QUOTE=James Heinrich;295796]Since it wasn't immediately obvious to me due to scrolling of the code box, I assume you mean "how did [i]rduerr[/i] find a factor with LL"?[/QUOTE]

Indeed.

[QUOTE]Since it wasn't immediately obvious to me due to scrolling of the code box, I assume you mean "how did [I]rduerr[/I] find a factor with LL"? [/QUOTE]

rduerr did a TF up to 73 before the LL test. By finding a factor, he saved about 72 GHz days for the LL test.

[QUOTE=chalsall;295645]OK. A quick hack before I run out for a "Sundowner"...

If you go to the LLTF form and choose the new option "I have a slow card!!!" it will ignore the range and pledge fields, and issue you work between 60M and 61M to TF from 70 to 71.[/QUOTE]

I would again like to point out on behalf of BDot, Jerry, myself, and all of the other AMD owners that AMD's are around 10-15% faster going to 70 and below than to 71 and 72 due to how the optimal kernel is constructed.

I won't click the slow card button, and I will grab some 72's if the project thinks that is what is better, but I just hope you have this in your mind as a factor to think about. The performance drop is very significant for us above 70.

[QUOTE=KyleAskine;295907]I would again like to point out on behalf of BDot, Jerry, myself, and all of the other AMD owners that AMD's are around 10-15% faster going to 70 and below than to 71 and 72 due to how the optimal kernel is constructed.

I won't click the slow card button, and I will grab some 72's if the project thinks that is what is better, but I just hope you have this in your mind as a factor to think about. The performance drop is very significant for us above 70.[/QUOTE]

Hmm, that's a factor I did not take into consideration when I worked out the charts. It wouldn't affect the overall outcome much though, at least until a program to run LL's on those GPUs is made.

Periodically the reservation system backfires.
When it happened first time, I thought wth but I couldn't be bothered.
Now, this time I happened to have a snapshot of my reservation page, so I defintely know that it was my asignment:
[CODE]52456903 P-1 72 - - 2012-04-08 19:30:36 2.46[/CODE]
...yet, the final result was "not needed", -- it was done by someone else (J.H.).


P.S. The [I]other[/I] J.H.

[QUOTE=Batalov;295940]...yet, the final result was "not needed", -- it was done by someone else (J.H.).[/QUOTE]It wasn't me! :surrender

[QUOTE=Batalov;295940]Periodically the reservation system backfires.
When it happened first time, I thought wth but I couldn't be bothered.
Now, this time I happened to have a snapshot of my reservation page, so I defintely know that it was my asignment:
[CODE]52456903 P-1 72 - - 2012-04-08 19:30:36 2.46[/CODE]
...yet, the final result was "not needed", -- it was done by someone else (J.H.).[/QUOTE]
It says it was assigned to me today, but I haven't grabbed any P-1 assignments today?
[CODE]52456903 P-1 72 --2012-04-09 02:20:39 2012-04-09 09:33:23[/CODE]

Edit: Forgot... I did grab 20 P-1s last night, so today Zulu time, for a new system. I don't know what happened?

@chalsall: If I interpret the charts correctly, GPUto72 holds 6588 assignments between 53M and 54M. Of those, 0059 are assigned. The other 99+% are simply being [B]hoarded[/B]. While the hoarding isn't as bad in all ranges, the hoarding remains egregious. As of this moment, PrimeNet offers [B]nothing[/B] for TF below 65M

On my own behalf, I am requesting that you release some of your hoarded assignments in all available ranges, so that I may acquire some assignments directly from the PrimeNet server. [I respectfully suggest you release 50% of your hoarded assignments. If you run out first, you can ask PrimeNet for 50% of what remains. if PrimeNet runs out first, you should release 50% of what you still hold.]

Thank you.

[QUOTE=rcv;295949]@chalsall: If I interpret the charts correctly, GPUto72 holds 6588 assignments between 53M and 54M. Of those, 0059 are assigned.[/QUOTE]

No; you are not interpreting the charts correctly. Please look, and come again.

[QUOTE=rcv;295949]On my own behalf, I am requesting that you release some of your hoarded assignments in all available ranges, so that I may acquire some assignments directly from the PrimeNet server. [I respectfully suggest you release 50% of your hoarded assignments. If you run out first, you can ask PrimeNet for 50% of what remains. if PrimeNet runs out first, you should release 50% of what you still hold.[/QUOTE]

I have already released 11,000 candidates back to PrimeNet to issue for Trial Factoring. Why don't you take some of them?

As an aside rcv, just how much TF work can you do compared to what we can do?

[QUOTE=Batalov;295940]...yet, the final result was "not needed", -- it was done by someone else (J.H.).[/QUOTE]

Sorry about that.

I made a mistake in an SQL update intended to release several thousand candidates back to PrimeNet to issue to those who refuse to use GPU72.

The good news is only three (inclusive of this) candidate assignments were effected.

That's all right, no big deal. I was worrying that there was a bigger leak, but three is negligible indeed.

[QUOTE=chalsall;293790][URL="http://www.gpu72.com/reports/estimated_completion/primenet/"]View added.[/URL] Answer: less than 200 days to Trial Factor everything to the new GPU levels up to 60M for every candidate not already LLed.

Wow!!! GPUs are just amazing!!! :smile:[/QUOTE]

18 days later and we are down to 150 days.

[QUOTE=petrw1;295963]18 days later and we are down to 150 days.[/QUOTE]

if the rate is linear it looks like 54 more days , if the rate is exponential it looks like maybe 41 more days possibly less. at least according to the math I did.

[QUOTE=science_man_88;295964]if the rate is linear it looks like 54 more days , if the rate is exponential it looks like maybe 41 more days possibly less. at least according to the math I did.[/QUOTE]

It's odd; I would have thought there would be enough GPUto72 history to establish a closer estimation. Maybe there was a recent surge in activity.
And if that surge was a one-time then the average will move back to 200 days-ish and if it is a sign of more consistent activity then it will drop even faster.

[QUOTE=chalsall;295956]No; you are not interpreting the charts correctly. Please look, and come again.

I have already released 11,000 candidates back to PrimeNet to issue for Trial Factoring. Why don't you take some of them?

As an aside rcv, just how much TF work can you do compared to what we can do?[/QUOTE]

1. I had a feeling you would be playing games, so I captured a screen-shot immediately before my last post. See below. If I'm not interpreting it correctly, then please explain. The row marked 53M. The first green column headed "Reserved from PrimeNet / Total" contains the number 6588. The last green column headed "Assigned' contains the number 59.

2. I believe you deliberately released candidates that you *knew* PrimeNet would not hand out for Trial Factoring. [Yes, you gave back several thousand in the 59M range, which PrimeNet is now handing out for P-1 testing.] The Candidates George last stated were made available for Trial Factoring for GPUs (unless I missed an announcement) were under the lower 50M ranges. Exactly how many candidates did you return in the 53M, 54M, and 55M ranges?

3. You arrogant so-and-so. You have no business asking how much I can do or why I want the candidates. You have no business tracking me at all. (You also had no business giving me the complete login credentials, e-mail address, and password of another user, thanks to another of your many bugs.) The answer to your question is "It is none of *your* business".

[If George asks, I'll tell him my plans.]

Meanwhile, all you need to know is that I want candidates in *all* ranges. So, I am formally asking you to release excess candidates in *all* ranges, so that I may obtain some assignments directly from PrimeNet in any or all ranges I choose.

[QUOTE=petrw1;295966]It's odd; I would have thought there would be enough GPUto72 history to establish a closer estimation. Maybe there was a recent surge in activity.
And if that surge was a one-time then the average will move back to 200 days-ish and if it is a sign of more consistent activity then it will drop even faster.[/QUOTE]

yeah I was just basing it on 50 days/18 days = 150 days/54 days and [TEX]\text {ceil}(150^{log_{18}(50)})= 41[/TEX]

[QUOTE=rcv;295967]1. I had a feeling you would be playing games, so I captured a screen-shot immediately before my last post. See below. If I'm not interpreting it correctly, then please explain. The row marked 53M. The first green column headed "Reserved from PrimeNet / Total" contains the number 6588. The last green column headed "Assigned' contains the number 59.

2. I believe you deliberately released candidates that you *knew* PrimeNet would not hand out for Trial Factoring. [Yes, you gave back several thousand in the 59M range, which PrimeNet is now handing out for P-1 testing.] The Candidates George last stated were made available for Trial Factoring for GPUs (unless I missed an announcement) were under the lower 50M ranges. Exactly how many candidates did you return in the 53M, 54M, and 55M ranges?

3. You arrogant so-and-so. You have no business asking how much I can do or why I want the candidates. You have no business tracking me at all. (You also had no business giving me the complete login credentials, e-mail address, and password of another user, thanks to another of your many bugs.) The answer to your question is "It is none of *your* business".

[If George asks, I'll tell him my plans.]

Meanwhile, all you need to know is that I want candidates in *all* ranges. So, I am formally asking you to release excess candidates in *all* ranges, so that I may obtain some assignments directly from PrimeNet in any or all ranges I choose.[/QUOTE]

Grow up already. If you look [URL]http://www.gpu72.com/reports/estimated_completion/[/URL] you will see a more accurate look at what primenet/GPU72 have.

Is the P-1 pool limited to 200?

[QUOTE=Dubslow;295979]Is the P-1 pool limited to 200?[/QUOTE]

It was supposed to be 1000; maybe that is simply all that is available?

Somehow I don't think so. chalsall said that he threw back 600 or so earlier, in anticipation of a large batch of TF to 72 coming in, but it's never gone back up, in fact it's gone down, and it's been stable at exactly 200 for most of the day. OTOH, 200 is only two days' worth, that's why we went with 1000 originally...

[QUOTE=rcv;295967]1. I had a feeling you would be playing games, so I captured a screen-shot immediately before my last post. See below. If I'm not interpreting it correctly, then please explain. The row marked 53M. The first green column headed "Reserved from PrimeNet / Total" contains the number 6588. The last green column headed "Assigned' contains the number 59.[/QUOTE]

I'm not playing games.

In between when you posted and I did, someone reserved several hundred candidates.

[QUOTE=rcv;295967]You arrogant so-and-so. You have no business asking how much I can do or why I want the candidates. You have no business tracking me at all. (You also had no business giving me the complete login credentials, e-mail address, and password of another user, thanks to another of your many bugs.) The answer to your question is "It is none of *your* business".[/QUOTE]

I will speak to this tomorrow.

[QUOTE=rcv;295967]So, I am formally asking you to release excess candidates in *all* ranges, so that I may obtain some assignments directly from PrimeNet in any or all ranges I choose.[/QUOTE]

And I'm saying if you want them so badly then write your own Spiders which can do what mine can. Or, simply reserve them from GPU72.

[QUOTE=petrw1;295966]It's odd; I would have thought there would be enough GPUto72 history to establish a closer estimation. Maybe there was a recent surge in activity.
And if that surge was a one-time then the average will move back to 200 days-ish and if it is a sign of more consistent activity then it will drop even faster.[/QUOTE]

No... "science_man" is just guessing, as usual.

The estimates reports are based on our current 30 day average production and the exact TF level of all exponents we hold, and seperately, what PrimeNet holds.

[QUOTE=chalsall;295984]I'm not playing games.[/QUOTE]
May I suggest you de-stress your life a bit by ignoring rcv's posts. And thanks again for all of the work you have put into making GPU72 the great project that it is. :tu:

[QUOTE=frmky;295988]May I suggest you de-stress your life a bit by ignoring rcv's posts. And thanks again for all of the work you have put into making GPU72 the great project that it is. :tu:[/QUOTE]

I heartily agree. Since GPU72 only has 38K of the 201K exponents that need TF, which is 1/5 and much smaller than the 1/2 he says you should hold so there are some for him to use. If he does not have the brains to be able to get any of those exponents(of which he could only do 10 or so a day by his track record), oh well.

[QUOTE=chalsall;295984]And I'm saying if you want them so badly then write your own Spiders which can do what mine can. Or, simply reserve them from GPU72.[/QUOTE]
Thank you for releasing some TF candidates back to PrimeNet. :smile:

[QUOTE=rcv;296053]Thank you for releasing some TF candidates back to PrimeNet. :smile:[/QUOTE]

They'll probably go fast (not speaking personally).
Over 50 gone already.

[QUOTE=chalsall;295984]I'm not playing games.

In between when you posted and I did, someone reserved several hundred candidates.
[/QUOTE]

hehe I did. :)


-- Craig

[QUOTE=chalsall;295984]And I'm saying if you want them so badly then write your own Spiders which can do what mine can. Or, simply reserve them from GPU72.[/QUOTE]

Ok, now I'm confused. I just noticed he is using GPUto72.:shock:
He has so far completed 1 DC-TF and 34 LL-TF.
That's not just in the last day is it?

[QUOTE=petrw1;296084]Ok, now I'm confused. I just noticed he is using GPUto72.:shock:
He has so far completed 1 DC-TF and 34 LL-TF.
That's not just in the last day is it?[/QUOTE]

No, he is no longer using GPU272.
Edit: He did that work over the course of 20 days in March.
[url]http://gpu72.com/reports/worker/067ccb9c43297464d2310d6c19589acb/[/url]

[QUOTE=petrw1;296084]Ok, now I'm confused. I just noticed he is using GPUto72.:shock:
He has so far completed 1 DC-TF and 34 LL-TF. That's not just in the last day is it?[/QUOTE]

No... Twenty days last month.

BTW, I released a few hundred in the 53M and 54M ranges today.

Almost all ended up being assigned as LL.... :cry:

[QUOTE=chalsall;296086]No... Twenty days last month.

BTW, I released a few hundred in the 53M and 54M ranges today.

Almost all ended up being assigned as LL.... :cry:[/QUOTE]

Lesson learned....think you had it right the first time.

[QUOTE=petrw1;296084]Ok, now I'm confused. I just noticed he is using GPUto72.:shock:
He has so far completed 1 DC-TF and 34 LL-TF.
That's not just in the last day is it?[/QUOTE]
Since I may be the "he" you are talking about, I thought I might reply.

First of all, I had been under the impression that all contributors to GIMPS were valued, whether they were large or small. So, I'm not sure why this matters.

But since you ask, petrw, this is a recent 28-hour sample of my direct PrimeNet contributions. [None of these assignments were obtained from GPUto72.]

[CODE]Manual testing 54738163 NF 2012-04-10 12:39 no factor for M54738163 from 2^71 to 2^72 [mfaktc 0.18...] 8.7371
Manual testing 54815989 NF 2012-04-10 12:39 no factor for M54815989 from 2^71 to 2^72 [mfaktc 0.18...] 8.7247
Manual testing 54711793 NF 2012-04-10 12:39 no factor for M54711793 from 2^71 to 2^72 [mfaktc 0.18...] 8.7414
Manual testing 54588557 NF 2012-04-10 12:39 no factor for M54588557 from 2^71 to 2^72 [mfaktc 0.18...] 8.7611
Manual testing 55102759 NF 2012-04-10 11:02 no factor for M55102759 from 2^71 to 2^72 [mfaktc 0.18...] 8.6793
Manual testing 55089497 NF 2012-04-10 11:02 no factor for M55089497 from 2^71 to 2^72 [mfaktc 0.18...] 8.6814
Manual testing 55078277 NF 2012-04-10 11:02 no factor for M55078277 from 2^71 to 2^72 [mfaktc 0.18...] 8.6832
Manual testing 55102741 NF 2012-04-10 11:02 no factor for M55102741 from 2^71 to 2^72 [mfaktc 0.18...] 8.6793
Manual testing 55083701 NF 2012-04-10 11:02 no factor for M55083701 from 2^71 to 2^72 [mfaktc 0.18...] 8.6823
Manual testing 54855301 NF 2012-04-10 11:02 no factor for M54855301 from 2^71 to 2^72 [mfaktc 0.18...] 8.7185
Manual testing 54854497 NF 2012-04-10 11:02 no factor for M54854497 from 2^71 to 2^72 [mfaktc 0.18...] 8.7186
Manual testing 61193879 NF 2012-04-10 07:50 no factor for M61193879 from 2^69 to 2^70 [mfaktc 0.18...] 1.9539
Manual testing 61199093 NF 2012-04-10 07:50 no factor for M61199093 from 2^69 to 2^70 [mfaktc 0.18...] 1.9537
Manual testing 61197839 NF 2012-04-10 07:50 no factor for M61197839 from 2^69 to 2^70 [mfaktc 0.18...] 1.9537
Manual testing 61178771 NF 2012-04-10 07:50 no factor for M61178771 from 2^69 to 2^70 [mfaktc 0.18...] 1.9543
Manual testing 55082593 NF 2012-04-10 06:59 no factor for M55082593 from 2^71 to 2^72 [mfaktc 0.18...] 8.6825
Manual testing 55049237 NF 2012-04-10 06:59 no factor for M55049237 from 2^71 to 2^72 [mfaktc 0.18...] 8.6878
Manual testing 55034867 NF 2012-04-10 06:59 no factor for M55034867 from 2^71 to 2^72 [mfaktc 0.18...] 8.6900
Manual testing 55014049 NF 2012-04-10 06:59 no factor for M55014049 from 2^71 to 2^72 [mfaktc 0.18...] 8.6933
Manual testing 54566471 NF 2012-04-10 06:59 no factor for M54566471 from 2^71 to 2^72 [mfaktc 0.18...] 8.7646
Manual testing 55081289 NF 2012-04-10 06:59 no factor for M55081289 from 2^71 to 2^72 [mfaktc 0.18...] 8.6827
Manual testing 54849997 NF 2012-04-10 06:59 no factor for M54849997 from 2^71 to 2^72 [mfaktc 0.18...] 8.7193
Manual testing 49787029 C 2012-04-10 06:51 a912aacf8d04c3__ 105.5208
Manual testing 55019477 NF 2012-04-10 03:55 no factor for M55019477 from 2^71 to 2^72 [mfaktc 0.18...] 8.6925
Manual testing 54964529 NF 2012-04-10 03:55 no factor for M54964529 from 2^71 to 2^72 [mfaktc 0.18...] 8.7012
Manual testing 55049387 NF 2012-04-10 03:55 no factor for M55049387 from 2^71 to 2^72 [mfaktc 0.18...] 8.6878
Manual testing 54833197 NF 2012-04-10 03:55 no factor for M54833197 from 2^71 to 2^72 [mfaktc 0.18...] 8.7220
Manual testing 55004603 NF 2012-04-10 01:29 no factor for M55004603 from 2^71 to 2^72 [mfaktc 0.18...] 8.6948
Manual testing 54869831 NF 2012-04-10 01:29 no factor for M54869831 from 2^71 to 2^72 [mfaktc 0.18...] 8.7162
Manual testing 54854537 NF 2012-04-10 01:29 no factor for M54854537 from 2^71 to 2^72 [mfaktc 0.18...] 8.7186
Manual testing 55042817 NF 2012-04-10 01:28 no factor for M55042817 from 2^71 to 2^72 [mfaktc 0.18...] 8.6888
Manual testing 54786341 NF 2012-04-10 01:28 no factor for M54786341 from 2^71 to 2^72 [mfaktc 0.18...] 8.7295
Manual testing 61178009 NF 2012-04-09 22:40 no factor for M61178009 from 2^69 to 2^70 [mfaktc 0.18...] 1.9544
Manual testing 61181711 NF 2012-04-09 22:40 no factor for M61181711 from 2^69 to 2^70 [mfaktc 0.18...] 1.9542
Manual testing 61174051 NF 2012-04-09 22:40 no factor for M61174051 from 2^69 to 2^70 [mfaktc 0.18...] 1.9545
Manual testing 61177643 NF 2012-04-09 22:40 no factor for M61177643 from 2^69 to 2^70 [mfaktc 0.18...] 1.9544
Manual testing 54852869 NF 2012-04-09 22:37 no factor for M54852869 from 2^71 to 2^72 [mfaktc 0.18...] 8.7189
Manual testing 55032737 NF 2012-04-09 22:37 no factor for M55032737 from 2^71 to 2^72 [mfaktc 0.18...] 8.6904
Manual testing 54781409 NF 2012-04-09 22:37 no factor for M54781409 from 2^71 to 2^72 [mfaktc 0.18...] 8.7302
Manual testing 54833239 NF 2012-04-09 20:55 no factor for M54833239 from 2^71 to 2^72 [mfaktc 0.18...] 8.7220
Manual testing 54833131 NF 2012-04-09 20:55 no factor for M54833131 from 2^71 to 2^72 [mfaktc 0.18...] 8.7220
Manual testing 54794617 NF 2012-04-09 20:55 no factor for M54794617 from 2^71 to 2^72 [mfaktc 0.18...] 8.7281
Manual testing 55017073 NF 2012-04-09 20:55 no factor for M55017073 from 2^71 to 2^72 [mfaktc 0.18...] 8.6929
Manual testing 55009709 NF 2012-04-09 20:55 no factor for M55009709 from 2^71 to 2^72 [mfaktc 0.18...] 8.6940
Manual testing 54736081 NF 2012-04-09 20:55 no factor for M54736081 from 2^71 to 2^72 [mfaktc 0.18...] 8.7375
Manual testing 54735487 NF 2012-04-09 20:55 no factor for M54735487 from 2^71 to 2^72 [mfaktc 0.18...] 8.7376
Manual testing 54711301 NF 2012-04-09 20:55 no factor for M54711301 from 2^71 to 2^72 [mfaktc 0.18...] 8.7414
Manual testing 54786311 NF 2012-04-09 13:27 no factor for M54786311 from 2^71 to 2^72 [mfaktc 0.18...] 8.7295
Manual testing 54998473 NF 2012-04-09 13:27 no factor for M54998473 from 2^71 to 2^72 [mfaktc 0.18...] 8.6958
Manual testing 54686647 NF 2012-04-09 13:27 no factor for M54686647 from 2^71 to 2^72 [mfaktc 0.18...] 8.7454
Manual testing 54739403 NF 2012-04-09 11:15 no factor for M54739403 from 2^71 to 2^72 [mfaktc 0.18...] 8.7369
Manual testing 61176767 NF 2012-04-09 09:41 no factor for M61176767 from 2^69 to 2^70 [mfaktc 0.18...] 1.9544
Manual testing 61174037 NF 2012-04-09 09:21 no factor for M61174037 from 2^69 to 2^70 [mfaktc 0.18...] 1.9545
Manual testing 61173577 NF 2012-04-09 09:21 no factor for M61173577 from 2^69 to 2^70 [mfaktc 0.18...] 1.9545
Manual testing 61178989 NF 2012-04-09 09:21 no factor for M61178989 from 2^69 to 2^70 [mfaktc 0.18...] 1.9543
Manual testing 61199023 NF 2012-04-09 09:21 no factor for M61199023 from 2^69 to 2^70 [mfaktc 0.18...] 1.9537
Manual testing 61177117 NF 2012-04-09 09:21 no factor for M61177117 from 2^69 to 2^70 [mfaktc 0.18...] 1.9544
Manual testing 61193789 NF 2012-04-09 09:21 no factor for M61193789 from 2^69 to 2^70 [mfaktc 0.18...] 1.9539
Manual testing 61227763 NF 2012-04-09 09:21 no factor for M61227763 from 2^69 to 2^70 [mfaktc 0.18...] 1.9528
Manual testing 61173979 NF 2012-04-09 09:21 no factor for M61173979 from 2^69 to 2^70 [mfaktc 0.18...] 1.9545
Manual testing 61177157 NF 2012-04-09 09:21 no factor for M61177157 from 2^69 to 2^70 [mfaktc 0.18...] 1.9544
Manual testing 61180541 NF 2012-04-09 09:21 no factor for M61180541 from 2^69 to 2^70 [mfaktc 0.18...] 1.9543
Manual testing 61191049 NF 2012-04-09 09:21 no factor for M61191049 from 2^69 to 2^70 [mfaktc 0.18...] 1.9539
Manual testing 61176769 NF 2012-04-09 09:21 no factor for M61176769 from 2^69 to 2^70 [mfaktc 0.18...] 1.9544
Manual testing 61179023 NF 2012-04-09 09:21 no factor for M61179023 from 2^69 to 2^70 [mfaktc 0.18...] 1.9543
Manual testing 54736109 NF 2012-04-09 09:18 no factor for M54736109 from 2^71 to 2^72 [mfaktc 0.18...] 8.7375

523.1508
[/CODE]

Well, at least we know it's possible to get fully TFd LL assignments :smile:

[QUOTE=rcv;296095]Since I may be the "he" you are talking about, I thought I might reply.... [None of these assignments were obtained from GPUto72.]

[CODE]... [mfaktc 0.18...]
... [mfaktc 0.18...] [/CODE][/QUOTE]
This "doesn't compute". What's the size of the chip on your shoulder, dude?
GPU72 was created to help people with GPUs to get reasonable assignments that otherwise usually went straight to LL with less than desired/now-possible/ TF done. The point of all of your fight for freedom was ... to get them from PrimeNet to do mfaktc?? now, that's funny.

[QUOTE=KyleAskine;295907]I would again like to point out on behalf of BDot, Jerry, myself, and all of the other AMD owners that AMD's are around 10-15% faster going to 70 and below than to 71 and 72 due to how the optimal kernel is constructed.

I won't click the slow card button, and I will grab some 72's if the project thinks that is what is better, but I just hope you have this in your mind as a factor to think about. The performance drop is very significant for us above 70.[/QUOTE]

Sorry there's a slight delay in getting the next preview to you guys (AMD owners), but it will contain some remedy. At least on my HD5770, a new kernel's speed is about in the middle between the fast "up-to-70" and the next slower kernel. And it is good for up to 2^73 - our usual workload should be covered by this one. And it holds my great hope for significantly improving Cayman's performance. I'll send it to you later today (if nothing changes my plans, again).

Chalsall, thanks for the "slow card" option, I also have one of those CC1.2 nvidia cards that right now chews on one of the 2^71 -> 2^72 assignments at about 66 secs per class (>17hrs total) ...

[QUOTE=Batalov;296097]This "doesn't compute". What's the size of the chip on your shoulder, dude?
GPU72 was created to help people with GPUs to get reasonable assignments that otherwise usually went straight to LL with less than desired/now-possible/ TF done. The point of all of your fight for freedom was ... to get them from PrimeNet to do mfaktc?? now, that's funny.[/QUOTE]

What a pyrrhic victory as well, between nucleon, xyzzy and me, those released exponents would have taken 24 hours to complete. He's also upgraded, as a single 560 could not output the ~335GHzD/day that printout shows, which was a nice touch, claiming '28 hours of work' when it should be obvious that counting work completed prior to that time frame does not count(Maybe he expects us to belive he could actually complete a 49M LL in 21 hours, when a 580 would take over 44 hours).

There are phrases that aptly apply to his distrust of GPU72: Never assume malice when ignorance will suffice and Get off your high horse. challsall is obviously not an expert web designer, so errors are possible, but making a mountain out of a molehill? Relax. The potential 'compromise' of information here is rather small, it's not like he has credit card information on the site, plus I must ask, once notified of the problem, wasn't it taken care of in a timely fashion?

Not to mention that primenet strikes me as a zillion times more insecure than gpu272.

[QUOTE=rcv;296095]Since I may be the "he" you are talking about, I thought I might reply.[/QUOTE]

You are "he". My surprise came from noticing that you were already on GPUto72 though I had the impression that you were opposed to the effort or need to sign up to GPUto72.

[QUOTE]First of all, I had been under the impression that all contributors to GIMPS were valued, whether they were large or small. So, I'm not sure why this matters.[/QUOTE]

And what would make you imply they are not?
You are obviously contributing a lot more than I am and I have NEVER felt under valued.

[QUOTE=KyleAskine;296125]Not to mention that primenet strikes me as a zillion times more insecure than gpu272.[/QUOTE]

I had promised to speak to this "security exposure" rcv discovered, and keeps bringing up...

I made a mistake on the Sign Up form. The Perl/Pseudo code was:

[CODE]my ($UN, $PW, $PWC, $EMail) = ("", "", "", "");

if ($ENV{'REQUEST_METHOD'} eq "POST") {
[COLOR="Red"][Extract $UN, $PW, $PWC, $EMail from POST object][/COLOR]

[COLOR="Red"][Do sanity checks on submitted data][/COLOR]

if ($Err eq "") {
[Insert user into database with unconfirmed state]
[EMail user with a link to activate account]
exit;
}
}

[COLOR="Red"][Render form with the $UN, $PW, $PWC and $EMail fields if an error occured
(i.e. e-mail address not correct format, $PW and $PWC (confirm) don't match,
Username already in use, $PW too short, etc.)][/COLOR][/CODE]

The bug rcv discovered (and, to his credit, reported to me) was that because this code is running under mod_perl, the $UN, $PW, $PWC and $EMail variables are only initiated when the code is run for the first time under each Apache server context, not every time the code is run.

Thus, there was a small temporal window after someone has signed up [B][I][U]for the first time[/U][/I][/B] when these variables would be exposed to another user if they happened to visit the Sign Up page shortly after, and happened to be being served by the same server context as the new user.

Immediately after rcv brought this to my attention I fixed the code and e-mailed the new user explaining what had happened.

As it has always said on the sign-up form, the passwords submitted are one-way encrypted before being stored. The only way someone could extract someone else's password is if they had access to their web-browser, or "sniffed the wire" in between the client's browser and my server.

Like bcp19 said, I think this is a bit of a molehill being turned into a mountain. But, at the same time, I do appreciate having this very embarrassing bug pointed out to me, as I expect to use the framework I've built for GPU72 for other projects, and it did mean my system broke the stated privacy policy.

:chalsall:

:batalov:[ATTACH]7889[/ATTACH]

[QUOTE=Xyzzy;296195]:chalsall:[/QUOTE]

Wow. I'm humbled. Truly.

Thanks!!! :smile:

[QUOTE=Bdot;296112]Chalsall, thanks for the "slow card" option, I also have one of those CC1.2 nvidia cards that right now chews on one of the 2^71 -> 2^72 assignments at about 66 secs per class (>17hrs total) ...[/QUOTE]

And thank you for your continuing efforts making mfakto as optimal as it can be.

I'm afraid I have been a little distracted lately dealing with a certain individual (who, for some reason, doesn't want anyone to know the real size of his hardware :wink:), but for those who want to continue working with GPU72 but have older cards, I've changed the assignment policy slightly, and have brought in some additional candidates to be processed by them.

There is now a stepped limit on the availability of candidates based on TF pledge level.

Those below 56M must be pledged to 72 bits.

Those below 58M must be pledged to 71 bits.

There are currently 149 candidates in the 58M range available to take to 70 bits, and more than 2,000 in th 61M range.

There is no longer a need to choose the "I have a slow card" option, "What makes sense" works as well.

[QUOTE=chalsall;293785]We currently hold about 42 days of LL Trial Factoring work.[/QUOTE]

So...what's next?

Keep working ahead of LL?
More focus on the DC line?
GPUto73/74?

[QUOTE=petrw1;296328]So...what's next?

Keep working ahead of LL?[/QUOTE]

We're not there yet. Although we're making good progress.

[QUOTE=petrw1;296328]More focus on the DC line?[/QUOTE]

Please, no... We're approximately 1M ahead of the wavefront in the DC range, and at ~200 a day there we're steady.

[QUOTE=petrw1;296328]GPUto73/74?[/QUOTE]

Some are doing that. And I thank them for doing so.

But at this point in time the most important thing is to go to 72 bits.

Finished my last assigned 25M DC on GPUto72

[QUOTE=davieddy;295312]... that you GPUto72 guys are a tiny and atypical fraction of all participants. ... [/QUOTE]Tiny? Rank 3 of the PrimeNet - Top Teams, Rank 1 of Top TF Teams (more than twice of number 2)

Keep in mind though that not every major TF player is on a team necessarily. GPU272 is also second place in P-1 throughput, but as a team, we would be third of fourth in the [i]individual[/i] P-1 list. I still agree that GPU272 is a large amount of total TF throughput; however, if I recall correctly, in context of what he was saying, I [i]believe[/i] he meant that 1) The number of participants is a small fraction of the whole (not that our throughput is tiny); we're 50 people, but there must be thousands of people doing TF work on PrimeNet; and that we are 50 doing a large majority of the work makes us atypical. In other words, I don't think he meant that we're insignificant, but rather that we are so few and yet significant makes us "weird compared to the average GIMPSter (who's never even heard of this forum).

Had some trouble getting exponents this morning. The available assignments page showed many exponents at 71 or less in the 53M range but I did not get anything less than 56M no matter how hard I tried. Does this have something to do with your new policy?

[QUOTE=garo;296393]Does this have something to do with your new policy?[/QUOTE]

Sort of... I added a conditional such that only people who had a certain daily rate could get exponents below 56M. This was intended to prevent new users (who we occationally see show up, reserve a few exponents in the lowest ranges, and then never report back) from holding up clearing out the lower ranges.

I've changed the cut off value to be only 20 GHzDays / Day. This will allow almost all of our users to get any range they want, except (ironcially) me.

My rate is a modest yet not insignificant 65 GHzDays/Day. Was the bar too high?

[QUOTE=garo;296407]My rate is a modest yet not insignificant 65 GHzDays/Day. Was the bar too high?[/QUOTE]

Yes, it was. 80 GD/D to be exact. It's now 20 GD/D.

Sorry about that.

Feel free to unreserve what you were assigned, and claim lower candidates.

I just turned in my very first piece of work going to 72 :smile:

[QUOTE=KyleAskine;296423]I just turned in my very first piece of work going to 72 :smile:[/QUOTE]

Congrats! How long did 71-72 take, in what exponent range?

[QUOTE=kladner;296424]Congrats! How long did 71-72 take, in what exponent range?[/QUOTE]

Not really sure for the time, but something pretty low, in the 48M range or so. I just did 'whatever the system thinks is best' or whatever.

It is much faster now that BDot has sped everything up for us AMD folk!

[QUOTE=chalsall;296410]Yes, it was. 80 GD/D to be exact. It's now 20 GD/D.

Sorry about that.

Feel free to unreserve what you were assigned, and claim lower candidates.[/QUOTE]

No probs. I'll finish what I have.

[QUOTE=chalsall;296410]80 GD/D to be exact. It's now 20 GD/D.[/QUOTE]Can this value be displayed on each user's [url=http://www.gpu72.com/reports/worker/56f1b7572536a14513b08c88b2ba9578/]Overall Statistics[/url] table? There's currently a total GHz-days column, but no per-day average. Actually overall-average and last-30-days-average would both be interesting, if they're not hard to calculate.

[QUOTE=James Heinrich;296445]Can this value be displayed on each user's [URL="http://www.gpu72.com/reports/worker/56f1b7572536a14513b08c88b2ba9578/"]Overall Statistics[/URL] table? There's currently a total GHz-days column, but no per-day average. Actually overall-average and last-30-days-average would both be interesting, if they're not hard to calculate.[/QUOTE]

Overall average can be found if you goto View Assignments. Since I had gradually added cards to GPU72, my average is just over 1/2 my current level.

[QUOTE=bcp19;296451]Overall average can be found if you goto View Assignments.[/QUOTE]It can, but I still think it would be relevant to put on the individual Overall stats page.

[QUOTE=chalsall;293785]We currently hold about 42 days of LL Trial Factoring work. [/QUOTE]

Curiously, about 25 days later we are still at 42.
At the time of this original copletion estimate did GPUto72 NOT have exponents all the way to 59M counted in this range?

In the same period the primenet range dropped from 200 to 150.
Does this just mean PrimeNet is releasing assignments to us as quickly as we can complete them (by the way, I am using the royal "we").

[QUOTE=petrw1;296555]Curiously, about 25 days later we are still at 42.
At the time of this original copletion estimate did GPUto72 NOT have exponents all the way to 59M counted in this range?

In the same period the primenet range dropped from 200 to 150.
Does this just mean PrimeNet is releasing assignments to us as quickly as we can complete them (by the way, I am using the royal "we").[/QUOTE]
The first number varies widely by what GPU272 actually has reserved (30K, 40K, 60K, etc.) and the second overall/PrimeNet stat depends on if the 60M-61M or 61-62 ranges are counted, which also depends entirely on whether or not GPU272 has any expos in those ranges.

[QUOTE=petrw1;296555]Curiously, about 25 days later we are still at 42.[/QUOTE]

According to Douglas Adams, the answer to Life, the Universe, and Everything!!!

What was the question again? :wink:

[QUOTE=petrw1;296555]At the time of this original copletion estimate did GPUto72 NOT have exponents all the way to 59M counted in this range?[/QUOTE]

It did. And does.

[QUOTE=petrw1;296555]Does this just mean PrimeNet is releasing assignments to us as quickly as we can complete them (by the way, I am using the royal "we").[/QUOTE]

More or less...

But please do note that we're being as "polite" as possible, and there are always TF assignments available below 60M for those who don't like playing with us...

To answer your question a bit more seriously, when the Estimated Completion report was written we didn't hold nearly as many candidates in the 58M and 59M ranges. If you take those out of the summary going to 73 bits, we hold just over 31.5 days of work.

I operate one computer that's slow enough that I let it get P-1 from PrimeNet; it was just assigned a 52M at 72 bits (which is good, in the longrun) but why wasn't it kept by GPU272?

[QUOTE=Dubslow;296569]I operate one computer that's slow enough that I let it get P-1 from PrimeNet; it was just assigned a 52M at 72 bits (which is good, in the longrun) but why wasn't it kept by GPU272?[/QUOTE]

Could have been someone who TF'd it outside of GPUto72 (you can check actually)....or one of the aforementioned in the previous post.

'Twas Xyzzy from 70->72 in December I believe; yes.
[url]http://mersenne.org/report_exponent/?exp_lo=52714939[/url]

[QUOTE=Dubslow;296573]'Twas Xyzzy from 70->72 in December I believe; yes.
[url]http://mersenne.org/report_exponent/?exp_lo=52714939[/url][/QUOTE]

Then could be one chalsall graciously released...for one reason or another.

Well... some mostly good news. One of my friends from high school says the recent nVidia drivers aren't playing nice with Folding@Home, so he asked me to help him get setup with mfaktc; I asked him about his assignments, and he got 54M from 71->72, so PrimeNet is at least partially doing something right with all the expos we released.

@[i]chalsall[/i]: Could you please put a "check all" feature on the assignments page, it's painful to regenerate worktodo if you have to click each one.

I'll even give you some sample code :smile:[code]<input type="checkbox" onchange="for (var i = 0; i < this.form.elements.length; i++) { var e = this.form.elements[i]; if ((e.type == 'checkbox') && e.name.match(/^Exp[0-9]+$/)) { e.checked = this.checked; } }">[/code]

[QUOTE=James Heinrich;296768]@[i]chalsall[/i]: Could you please put a "check all" feature on the assignments page, it's painful to regenerate worktodo if you have to click each one.

I'll even give you some sample code :smile:[/QUOTE]

Thanks for the nudge, and the code...

I took a slightly different approach, with the intention of also being able to select by work types, expiring exponents, etc in the near future. But the code was a good inspiration, and what you've asked for is now implemented.

[QUOTE=chalsall;296786]what you've asked for is now implemented.[/QUOTE]Thanks!

I seem to have gotten a nice bunch of factor-rich exponents lately, 23 factors in the last 4 days. My work saved graph is all skewed now from finding 11 in a single day.

[QUOTE=bcp19;296797]I seem to have gotten a nice bunch of factor-rich exponents lately, 23 factors in the last 4 days. My work saved graph is all skewed now from finding 11 in a single day.[/QUOTE]

DANG! :rant: :max:

What range?

[QUOTE=bcp19;296797]I seem to have gotten a nice bunch of factor-rich exponents lately, 23 factors in the last 4 days. My work saved graph is all skewed now from finding 11 in a single day.[/QUOTE]

:shock::ouch1::ouch2:





:ignore::ignore:


:rant::rant::rant::rant::rant::rant::rant::rant::rant::rant::rant::rant:
:raman::raman::raman:


:help:
:surrender


:no::cry:




...:cool:


([url]http://psychcentral.com/lib/2006/the-5-stages-of-loss-and-grief/[/url])

That graph actually makes it look like Pete had a very long dry spell. Now, I know that statistics have no memory, but that long flat line makes it look like he deserved some good times.

EDIT: Or does this recent burst just put the past several months in deep shade? I see that that is approximately what he said in the first post. In any case, congrats bcp.

[QUOTE=kladner;296806]That graph actually makes it look like Pete had a very long dry spell. Now, I know that statistics have no memory, but that long flat line makes it look like he deserved some good times.

EDIT: Or does this recent burst just put the past several months in deep shade? I see that that is approximately what he said in the first post. In any case, congrats bcp.[/QUOTE]
Look at his work complete graph; you'll notice that he didn't start doing any major quantity of LL work until a few days ago anyways.

Also, @chalsall: Xyzzy's graph appears to be broken?
[url]http://gpu72.com/reports/worker/7e6a2e592a37a719fac4f765eb0f6ca8/[/url]
The "Worked Saved" reports a bunch of factors spread through April, but the "Work Done" graphs don't show anything since March.

[QUOTE=chalsall;296786]Thanks for the nudge, and the code...

I took a slightly different approach, with the intention of also being able to select by work types, expiring exponents, etc in the near future. But the code was a good inspiration, and what you've asked for is now implemented.[/QUOTE]

Awesome! Thanks ;)

[QUOTE=Dubslow;296807]Also, @chalsall: Xyzzy's graph appears to be broken?
[url]http://gpu72.com/reports/worker/7e6a2e592a37a719fac4f765eb0f6ca8/[/url]
The "Worked Saved" reports a bunch of factors spread through April, but the "Work Done" graphs don't show anything since March.[/QUOTE]

Yeah... Xyzzy seems to enjoy messing with my head/code (don't know why; not a lot of sport in it)...

He's submitted nothing but factors (60 of them) since 2012.04.04....

[QUOTE=chalsall;296811]Yeah... Xyzzy seems to enjoy messing with my head/code (don't know why; not a lot of sport in it)...

He's submitted nothing but factors (60 of them) since 2012.04.04....[/QUOTE]

...
...
...

...why? (42?)

My sympathies :P

Edit: Does that mean he unreserves assignments just before submitting them to PrimeNet? Or doesn't report them to PrimeNet? But then how do you know when to unreserve them? What about the throughput estimates? Wow, that's actually very problematic. My actual sympathies -- that's a lot of crap to handle.

[QUOTE=chalsall;296811]Yeah... Xyzzy seems to enjoy messing with my head/code (don't know why; not a lot of sport in it)...

He's submitted nothing but factors (60 of them) since 2012.04.04....[/QUOTE]

If he doesn't report everything, then if the exponents expire someone else will have to run them with a guarantee of finding no factors.

[QUOTE=kladner;296806]That graph actually makes it look like Pete had a very long dry spell. Now, I know that statistics have no memory, but that long flat line makes it look like he deserved some good times.

EDIT: Or does this recent burst just put the past several months in deep shade? I see that that is approximately what he said in the first post. In any case, congrats bcp.[/QUOTE]

As Dubslow noted, most of my work has been in the DC range, which has very low 'saved' work for found factors. Most of these factors were from exponents that needed 69-72 work done, but it was a lucky streak.

[QUOTE=bcp19;296815]As Dubslow noted, most of my work has been in the DC range, which has very low 'saved' work for found factors. Most of these factors were from exponents that needed 69-72 work done, but it was a lucky streak.[/QUOTE]

Actually, we've all forgotten about chalsall's new metric: you've found 38 factors to 34.6 expected, so not too lucky.

[url]http://gpu72.com/reports/worker/b30b0ed22bbd92716af40ebf12fbf217/[/url]

[QUOTE=flashjh;296814]If he doesn't report everything, then if the exponents expire someone else will have to run them with a guarantee of finding no factors.[/QUOTE]

I don't think he's being malicious; just playful...

But, yeah... It does mess with the WLRs, and thus the predicted factors found for everyone. Plus it means the predicted completion reports are under-estimating by a few days.

When you think about it, it's a little scary just how much influence Xyzzy has!!! :smile:

[QUOTE=Dubslow;296812]...

Edit: Does that mean he unreserves assignments just before submitting them to PrimeNet? Or doesn't report them to PrimeNet? [/QUOTE]

He's always submitted everything before...eventually

He actually used to only submit NF results, and then would submit the F results in a huge wave. Who knows what he is up to now.