[QUOTE=petrw1;391150]Is that a "BUG" or a "FEATURE" :smile:

For now I changed MisFit to give me specifically DCTF - lowest Exponent and I got 41M 70-71 Bits even though there are still 40M 71-72 Bits available....and 41M 69 - 70 Bits.

Then I unreserved the 78M from GPU72 and deleted from the MisFitWorkToDo. It that the correct procedure?[/QUOTE]

That is a feature. Primenet decided that doing that 78M exponent would save more LL CPU time than working on any other DC TF exponent available. "What makes sense" always translates into "saves the most LL CPU time per unit of GPU work expended".

Primenet does not hand out 40M exponents for DCTF to 72. By default, Primenet uses James' crossover charts for a GTX 670 and switches from bit depth 71 to bit depth 72 when the chart crosses 71.75. See [url]http://www.mersenne.ca/cudalucas.php?model=487[/url]

[QUOTE=Prime95;391168]That is a feature. Primenet decided that doing that 78M exponent would save more LL CPU time than working on any other DC TF exponent available. "What makes sense" always translates into "saves the most LL CPU time per unit of GPU work expended".

Primenet does not hand out 40M exponents for DCTF to 72. By default, Primenet uses James' crossover charts for a GTX 670 and switches from bit depth 71 to bit depth 72 when the chart crosses 71.75. See [url]http://www.mersenne.ca/cudalucas.php?model=487[/url][/QUOTE]

Fair enough.

Would this not always hand out the largest exponents since that would result in the most LL savings?
And if so how do you get the leading edge (lowest exponents) TF'd?

[QUOTE=petrw1;391150]For now I changed MisFit to give me specifically DCTF - lowest Exponent and I got 41M 70-71 Bits even though there are still 40M 71-72 Bits available....and 41M 69 - 70 Bits.[/QUOTE]

Please remember that only LG72D (and, soon, WMS) sets the "Pledge" level. You asked for "Lowest", but still have only pledged to 71. This is why you didn't get anything to take to 72. And, also, you didn't get any assignments from 69 because, again, you asked for lowest exponent. If you had asked for lowest TF level, you would have been given the lowest available candidates at the lowest bit level.

[QUOTE=petrw1;391150]Then I unreserved the 78M from GPU72 and deleted from the MisFitWorkToDo. It that the correct procedure?[/QUOTE]

Scott (Mr. MISFIT :smile:) would have to answer as to how to handle the MisFitWorkToDo side of things. But unreseving from GPU72 is definitely the correct first step.

[QUOTE=petrw1;391170]Would this not always hand out the largest exponents since that would result in the most LL savings? And if so how do you get the leading edge (lowest exponents) TF'd?[/QUOTE]

Please be aware that the GPU72 to Primenet proxying is not yet "live". As in, the 78M assignments you were given were by GPU72.

As I understand George's plans, Primenet will give out TF assignments which are the most profitable to the project AND are likely to be assigned in the next 60 days or so.

[QUOTE=petrw1;391170]Would this not always hand out the largest exponents since that would result in the most LL savings?[/quote]

Depends on how far each has been factored.

[quote]And if so how do you get the leading edge (lowest exponents) TF'd?[/QUOTE]

Unless you explicitly give a range of exponents, primenet is dolling out exponents exponents from the rather small range of those apt to be assigned for LL in the next 90 days. Within that narrow range you are correct that PrimeNet will tend to work downward.

[QUOTE=Prime95;391174]Within that narrow range you are correct that PrimeNet will tend to work downward.[/QUOTE]

This brings up a point we had privately discussed.

Would it not make sense to also assign LL work sorted by "Factored_To desc", in addition to and much like you do already for assignments where P-1 completed assignments are given before no P-1 completed (grouped by 1M range)?

I think this would be a (very slight) more optimal use of resources.

[QUOTE=chalsall;391176]I think this would be a (very slight) more optimal use of resources.[/QUOTE]

Very slight times thousands of machines is significant.

[QUOTE=chalsall;391176]
Would it not make sense to also assign LL work sorted by "Factored_To desc", in addition to and much like you do already for assignments where P-1 completed assignments are given before no P-1 completed (grouped by 1M range)?[/QUOTE]

Yes, it would make sense. There is no index on that column, so currently SQL would have to do a full scan of the 1M range. The new server would likely have no trouble doing that.

Even more optimal would be for LLTF to have enough firepower such that LLs are always optimally factored during the 90 day window the exponent is available for TF.

[QUOTE=Prime95;391200]There is no index on that column, so currently SQL would have to do a full scan of the 1M range. The new server would likely have no trouble doing that.[/QUOTE]

Indeed; adding a new index should only take a few seconds.

[QUOTE=Prime95;391200]Even more optimal would be for LLTF to have enough firepower such that LLs are always optimally factored during the 90 day window the exponent is available for TF.[/QUOTE]

I think we're alread there.

But I would argue that if "working down" for TF'ing is what is done, than "working down" for LL'ing also makes sense.

Hey Chris, could you bring in 42/43M?
There are no DC assignments free to 70 bits for my 7970.

[QUOTE=LaurV;392251]Hey Chris, could you bring in 42/43M?
There are no DC assignments free to 70 bits for my 7970.[/QUOTE]

Oops. I've been doing a breadth-first approach, for fun.

Hi Chris, could you please check what GPU72 expects to happen for [B][URL="http://www.mersenne.org/report_exponent/default.php?exp_lo=75685481&full=1"]75685481[/URL], [/B]for example? I have a couple of old assignments that I was about to add to my worktodo again. This time I checked upfront and could avoid the "result not needed" ... anyway, these assignments are long done, but not picked up by GPU72.

[QUOTE=Bdot;392266]Hi Chris, could you please check what GPU72 expects to happen for [B][URL="http://www.mersenne.org/report_exponent/default.php?exp_lo=75685481&full=1"]75685481[/URL], [/B]for example? I have a couple of old assignments that I was about to add to my worktodo again. This time I checked upfront and could avoid the "result not needed" ... anyway, these assignments are long done, but not picked up by GPU72.[/QUOTE]

GPU72 has no thoughts as to 75685481.

It doesn't "own" it, nor has it "worked" it.

[QUOTE]GPU72 has [U]no thoughts[/U] as to 75685481[/QUOTE]

heh! :smile:

[QUOTE=LaurV;392251]Hey Chris, could you bring in 42/43M?
There are no DC assignments free to 70 bits for my 7970.[/QUOTE]

While Chris is healing, you can use [url]http://mersenne.org/manual_gpu_assignment[/url]

[QUOTE=Prime95;392299]While Chris is healing, you can use [url]http://mersenne.org/manual_gpu_assignment[/url][/QUOTE]

This form required specifying an exponent range when choosing "TF for double-checks" and "Lowest exponents" with a bit level of 70. Is that a bug?

[QUOTE=Prime95;392299]While Chris is healing, you can use [url]http://mersenne.org/manual_gpu_assignment[/url][/QUOTE]

No disrespect intended, but choosing "TF for double-checks" with "What Makes sense" results in "Error code: 40
Error text: No assignment available for GPU trial factoring, cpu_id: 248633, user_id = 802"

Very possible I'm being stupid.

[QUOTE=Mark Rose;392303]This form required specifying an exponent range when choosing "TF for double-checks" and "Lowest exponents" with a bit level of 70. Is that a bug?[/QUOTE]

Not really, but we are open to improving the form.

What makes sense hands out the most priftable TF assignments on exponents that are expected to be handed out for DC or first-time LL in the next 90 days. That 90-day range is listed at the bottom of the page. One can override this by specifying there own range, bit levels, etc.

At this point in time, all the DC exponents in the 90-day range are either already assigned for TF or reserved by GPU72.

Short-term memory loss?
 

[QUOTE=chalsall;392289]GPU72 has no thoughts as to 75685481.

It doesn't "own" it, nor has it "worked" it.[/QUOTE]

Maybe. But it assigned it to me. I just want it to accept that it's done.

[QUOTE=Bdot;392321]Maybe. But it assigned it to me. I just want it to accept that it's done.[/QUOTE]

Hmmm... Weird...

The "GPU" table had it listed as "Returned without having been worked", but you are correct, the assignment table indeed had it assigned to you. Fixed this, and seven others.

Let me know if you, or anyone else, sees this situation again.

Is Spidee Stuck?
 

All DCTF Assignments I finished since 14:00GMT Jan 14 (about 60 or so) are still on my assignments page.

spot instances
 

[QUOTE=Mark Rose;388023]The g2.2xlarge is not such a good value. You only get a single GPU at 240 GHz-d/d instead of two GPUs at 280 GHz-d/d. For mfaktc, the g2.2xlarge's fair value in comparison is $0.05/hr.

The cg1.4xlarge is only available in us-east-1 and eu-west-1. Of those, only in one available zone in us-east-1 has the spot price been hovering around $0.135.[/QUOTE]
This discussion helped me a lot to cut down on the curve. Thanks, Chris, Mark!

Hmmmmm... However, along with GPU(s), you get a 3-yrs younger Xeon E5-2670 [I]with AVX [/I](and larger cache and faster memory interface). For a typical prime (or PRP) search project, E5-2670 is as as fast as the two quad-cores that come with the cg1.4xlarge. Also, lesser availability leads to more frequent node drop-outs (even if you set your bid to $0.20, the history shows that you will lose nodes, maybe as frequently as once a week).

So, ~$0.135/hr pays for:
a) 2x M2050 + 8 cores of X5570 CPU (cg1.4xlarge), or
b) 2x K520 + 8 cores of E5-2670 AVX CPU (2 x g2.2xlarge).
I've run both variants for some time. A little more management for variant b.

I think I'll pick up 10-20 of g2.2xlarge spots for my pet project (need to run some [I]Cyclo[/I]; and will pick up some slack on my G.-M. search and/or k=5 Proth that I had dormant for a while). For Cyclo, M2050 = K520, but the CPU power in variant b is twice as much:
[CODE]@cg1.4xlarge:
86990^262144 - 86990^131072 + 1 is composite [Res=11411658520]. ([COLOR=Blue]2132.3 sec.,[/COLOR] err=5.74e-03)
85217^262144 - 85217^131072 + 1 is composite [Res=11150025403]. ([COLOR=Blue]2128.4 sec.[/COLOR], err=5.69e-03)
10^249010-21 is base 3-Strong Fermat PRP! (249010 decimal digits) Time : [B][COLOR=DarkRed]591.348 sec.[/COLOR][/B]
Starting Lucas sequence
10^249010-21 is strong-Fermat and Lucas PRP, Starting Frobenius test sequence
10^249010-21 is strong-Fermat, Lucas and Frobenius PRP! (P = 4, Q = 2, D = 8) Time : [B][COLOR=DarkRed]2303.341 sec.[/COLOR][/B]
(Factored part = 100.00%)

@g2.2xlarge:
82517^262144 - 82517^131072 + 1 is composite [Res=10786891370]. ([COLOR=Blue]2109.0 sec.[/COLOR], err=5.14e-03)
82512^262144 - 82512^131072 + 1 is composite [Res=10823198411]. ([COLOR=Blue]2108.6 sec.[/COLOR], err=5.10e-03)
Starting probable prime test of 10^249010-21
10^249010-21 is base 3-Strong Fermat PRP! (249010 decimal digits) Time : [B][COLOR=DarkGreen]295.374 sec.[/COLOR][/B]
Starting Lucas sequence
10^249010-21 is strong-Fermat and Lucas PRP, Starting Frobenius test sequence
10^249010-21 is strong-Fermat, Lucas and Frobenius PRP! (P = 4, Q = 2, D = 8) Time : [B][COLOR=DarkGreen]1262.299 sec.[/COLOR][/B]
(Factored part = 100.00%)[/CODE]Here is how can prep any of these nodes, in a few lines:
[CODE] sudo yum erase nvidia cudatoolkit
sudo yum install gcc
wget http://us.download.nvidia.com/XFree86/Linux-x86_64/340.46/NVIDIA-Linux-x86_64-340.46.run
sudo yum install kernel-devel-`uname -r`
sudo /bin/bash ./NVIDIA-Linux-x86_64-340.46.run
# answer prompts
nvidia-smi
# ... and ready to run, either CUDA or OpenCL[/CODE]

The GTX 660 Ti is going offline for now. I am moving it to the new box I made. Until I get it up and running, the GTX 660 Ti won't be doing anything.

Now I get to fiddle with drivers and all that for a good little while.

Trying not to be a nag....
 

I know typing is still an issue ... so for some future reference ...

We seem to be losing more and more ranges from GPU Visualization.
Looking at this report ...

[url]http://www.mersenne.info/trial_factored_tabular_delta_7/0/0/[/url]

Everywhere that you see a -20,000 or more in the unfactored column (and there are several) a drill down will soon lead to a missing range.
Obviously we have NOT factored 20,000 in any range in the last week but rather we have lost entire ranges ... they appear as blank lines on the report.

Those are the ranges were the "not yet known" mersenne primes are, so they are hidden on purpose, Chris is doing everything possible to impede us finding them... He might be cooperating with curtisc for that, behind of our backs.

I was especially pissed off about disparition (why is this redlined? disappearance? aha!) of the first million exponents from those tables, but I gave up after a while, trying to explain to myself that getting rid of the 3-4 columns on the left of the table was useful, as it relieved other 3 more important columns on the right, so after some initial cry, I kept quiet. Well, the same effect would be achieved using columns labeled "<63" and ">74" for example, and everything would be more accurate, and the first mio won't be pushed out, but well, that would require some programming, and Chris [STRIKE]was grumpy and didn't want to do it[/STRIKE] was busy/not available.

Now that you mentioned it, I want my first mio exponents back! Grrrr... But Chris said already that gimps-visu site was never intended to be something serious, so I think we can frame our hopes for now, and put them on a wall..

[QUOTE=petrw1;393368]

Obviously we have NOT factored 20,000 in any range in the last week but rather we have lost entire ranges ... they appear as blank lines on the report.[/QUOTE]

They appear as blank lines on the Factored Depth report because they are counted as "Factored". If you go to the "Exponent Status" report and drill down to e.g. < 1M range you´ll see them all standing there as Factored.

I hear what you say about these reports. They were also a strong point of interest for me, and I still check them, although I know they are incorrect. Most ranges are still OK, though. It would be very nice to have them back in shape, but there are many things in life we would like to have and at the end of the day we have to do without... :whistle:

All in all, I guess LaurV´s explanation is the right one: heaps of Mersenne Primes are being hidden from us mere mortals, to keep us focused on finding the next one, with the wrong belief they are rare :smile:.

[QUOTE=petrw1;393368]I know typing is still an issue ... so for some future reference ...

We seem to be losing more and more ranges from GPU Visualization.
Looking at this report ...

[url]http://www.mersenne.info/trial_factored_tabular_delta_7/0/0/[/url]

Everywhere that you see a -20,000 or more in the unfactored column (and there are several) a drill down will soon lead to a missing range.
Obviously we have NOT factored 20,000 in any range in the last week but rather we have lost entire ranges ... they appear as blank lines on the report.[/QUOTE]

Well actually, if you look at 299m-300m you will see a movement of 20,227 OUT of the factored to 66 bits and 19,931 moved INTO the factored to 67 bits column.

I am working on that range and testing one exponent every 24 seconds.

If you drill right down until you get to the GIMPS factoring page you will find my name against all of them.

No conspiracy here :smile:

OOPS Partial TF results?
 

When I moved to LLTF I picked the wrong BIT Level by mistake and got 25 assignments for 58M from 73-75 bits.
I only want to do 74 as that is the recommended GPU72 level.

So I changed my MisFitworktodo.txt and changed all the 73,75 to 73,74.
They are completing to 74 but these results are NOT showing up on my RESUTLS page
... and the assignments are staying on my assignments page.

Apparently I should have done something different.

So.....now what should I do to:
- Get these results recorded
- Get the assignments marked as complete

Thanks.

[QUOTE=Gordon;393414]Well actually, if you look at 299m-300m you will see a movement of 20,227 OUT of the factored to 66 bits and 19,931 moved INTO the factored to 67 bits column.

I am working on that range and testing one exponent every 24 seconds.

If you drill right down until you get to the GIMPS factoring page you will find my name against all of them.

No conspiracy here :smile:[/QUOTE]

These are good.....I am only commenting on the ranges that report all 20,000 as Factored.

Hmm but at the present time I do not see any....maybe some magic has happened.

Though there are still ranges completely missing.
Here 973000000 is missing ... at least when I looked it was:
[url]http://www.mersenne.info/trial_factored_tabular_delta_1/2/970000000/[/url]

Um, go on and take them to 75? If this is impractical, perhaps Chris can fix it for you. That you got assignments to 75 is not a mistake, though. There was discussion before "the goal posts got moved."

PM to Chris is the best.

Just go to your gpu72 page and unreserved those exponents. You may not get gpu72 credit for those and may need to contact Chris if you care.

[QUOTE=petrw1;393479]When I moved to LLTF I picked the wrong BIT Level by mistake and got 25 assignments for 58M from 73-75 bits.
I only want to do 74 as that is the recommended GPU72 level.

So I changed my MisFitworktodo.txt and changed all the 73,75 to 73,74.
They are completing to 74 but these results are NOT showing up on my RESUTLS page
... and the assignments are staying on my assignments page.

Apparently I should have done something different.

So.....now what should I do to:
- Get these results recorded
- Get the assignments marked as complete

Thanks.[/QUOTE]

Actually its only missing one small thing:

Prime95 has them marked as complete and has credited me.
I am only missing GPU72 credit....nothing I will lose sleep over.
I will un-assign them from GPU72.

Here is the list if Chris is ever bored and wants to look into it
They were all taken from 73-74 bits

[CODE]58778063
58778053
58777981
58777787
58777009
58776943
58776409
58776343
58776307
58776233
58775491
58775461
58775081
58775011
58774927
58774433
58774307
58774267
58774091
58773167
58773151
58773007
58772887
58772849
58772237[/CODE]

[QUOTE=petrw1;393480]These are good.....I am only commenting on the ranges that report all 20,000 as Factored.

Hmm but at the present time I do not see any....maybe some magic has happened.

[/QUOTE]

Here's one...at least in this instant in time

[url]http://www.mersenne.info/trial_factored_tabular_delta_1/2/130000000/[/url]

I see 135M missing but a blank row at the end showing 22,853 factored.

[QUOTE=petrw1;393523]Here is the list if Chris is ever bored and wants to look into it
They were all taken from 73-74 bits[/QUOTE]

Yeah... That's the first thing I'll pay attention to when I'm allowed to type again...

That's meant to be funny, and serious, at the same time.

Saw my doctor again yesterday; back on the sterioids for a brief period.

[QUOTE=chalsall;393533]Yeah... That's the first thing I'll pay attention to when I'm allowed to type again...

That's meant to be funny, and serious, at the same time.

Saw my doctor again yesterday; back on the sterioids for a brief period.[/QUOTE]

No big deal....I mean my thing....just get better soon.

[QUOTE=petrw1;393539]No big deal....I mean my thing....just get better soon.[/QUOTE]

Thanks.

[QUOTE=chalsall;393667]Thanks.[/QUOTE]

Stop typing already :smile:

P.S. How's your better half's recovery going?

Now how are you going to answer this without typing?

[QUOTE=petrw1;393669]Stop typing already :smile:

P.S. How's your better half's recovery going?

Now how are you going to answer this without typing?[/QUOTE]

LOL... As always, women are stronger than men.

[QUOTE]Now how are you going to answer this without typing? [/QUOTE]

On screen keyboard?

Am I getting this???
 

When I see 10's of thousands of Available LL assignments in the 66-70M ranges here:
[url]http://www.mersenne.org/primenet/[/url]

But only hundreds TF'd to the optimal 75 bits here in that same range:
[url]http://www.gpu72.com/reports/current_level/[/url]

Is that telling me we GPU72'ers are very much NOT keeping up in LL-TF?

OR

Am I way off base?

[QUOTE=petrw1;393865]Is that telling me we GPU72'ers are very much NOT keeping up in LL-TF? Am I way off base?[/QUOTE]

It's a little complicated...

In the short form:

1. Many are doing DCTF'ing when it doesn't actually make much sense.

2. LLTF'ing is where the effort is most profitable.

3. Every keystroke I do costs me my recovery.

4. Once a day I examine the situation, and instruct my spiders.

5. Just in case it isn't clear, please do LLTF'ing unless it amuses you to do otherwise.

[QUOTE=chalsall;393866]
3. Every keystroke I do costs me my recovery.


5. Just in case it isn't clear, please do LLTF'ing unless it amuses you to do otherwise.[/QUOTE]

3. I will take a NULL response as a silent "GPU72 Thanks You".

5. OK....I'm convinced....starting tonight I will move my 515 Ghz back to LL (it was there for a few days... then got moved back to DC due to torn loyalties) :smile: And I will guess that "Let GPU72 Decide" is the preferred option

AND...

and give Die DC-TF Die!!! an extension on its Death Sentence.

[QUOTE=petrw1;393868]3. I will take a NULL response as a silent "GPU72 Thanks You".

5. OK....I'm convinced....starting tonight I will move my 515 Ghz back to LL (it was there for a few days... then got moved back to DC due to torn loyalties) :smile: And I will guess that "Let GPU72 Decide" is the preferred option

AND...

and give Die DC-TF Die!!! an extension on its Death Sentence.[/QUOTE]

I hadn't realized how valuable even that amount of firepower was. I'll shift to LLTF.

Shifting 1.1 THz-d/d to LLTF.

[QUOTE=Mark Rose;393880]Shifting 1.1 THz-d/d to LLTF.[/QUOTE]

Cool.

I just added an $84 +$10 shipping eBay 580. EVGA board, with a Zalman cooler. The latter adds as much slot-blocking thickness as an Asus DirectCU cooler for such boards. Fortunately the board itself is not as long as the Asus. It's not as noisy as the Asus, flat out; but it needs to run flat out (2 fans, 3000 RPM max) to stay in the 70s C. Factory OC is 797 MHz, and while I have run it up to 848 MHz, it definitely starts feeling 80ish at that point.

I still think the best GPU coolers that I have used are the Gigabyte dual and triple fan models. At top speed they are noticeable, but not like the DirectCU ASUS 580. Still, I would say it did a better job of cooling for a decibel price.

[QUOTE=kladner;393898]

I still think the best GPU coolers that I have used are the Gigabyte dual and triple fan models. At top speed they are noticeable, but not like the DirectCU ASUS 580. Still, I would say it [B]**did a better job of cooling**[/B] for a decibel price.[/QUOTE]


**Better than the Zalman cooler.**

[QUOTE=chalsall;393533]Yeah... That's the first thing I'll pay attention to when I'm allowed to type again...
[/QUOTE]
Hope that time will come soon... We tolerated one bug per page, but they [URL="http://www.mersenne.info/trial_factored_tabular_delta_30/2/420000000/"]start spreading[/URL], Soon, we will see more blanks than correct lines, per page... :razz:

.

[QUOTE=LaurV;395087]Hope that time will come soon... We tolerated one bug per page, but they [URL="http://www.mersenne.info/trial_factored_tabular_delta_30/2/420000000/"]start spreading[/URL], Soon, we will see more blanks than correct lines, per page... :razz:[/QUOTE]

Hmmmm... F' me...

Please do note that mersenne.info is completely independant from GPU72.

The errors being shown on mersenne.info is because of bad Cable and Wireless connectivity here in Bim, and (without question) stupid programming errors by yours truly.

It seems [url=http://www.mersenne.info/trial_factored_tabular_delta_1/3/41000000/]41M[/url] isn't updating daily at mersenneinfo, but work [url=http://www.gpu72.com/graphs/dctf/week/]is being done[/url] at gpu72.

[QUOTE=Mark Rose;395332]It seems [url=http://www.mersenne.info/trial_factored_tabular_delta_1/3/41000000/]41M[/url] isn't updating daily at mersenneinfo, but work [url=http://www.gpu72.com/graphs/dctf/week/]is being done[/url] at gpu72.[/QUOTE]

This can happen if there have been no changes over the day in the range queired, or if the query failed bacause of bad connectivity.

The machine which does the spidering for Mersenne.INFO is on a flaky Clueless and Worthless DSL here in Barbados. When I have some time (probably early March now; I'm able (and allowed) to type again, although slowly) I'm going to migrate the spidering to the US based machine which hosts GPU72 -- I have some "real" work to catch up on which pays for all of this.

[QUOTE=chalsall;395333]flaky Clueless and Worthless DSL here in Barbados.[/QUOTE]Some people love it, but I've never in 20 years had a good experience with DSL. In my house it stands for Dis Service is Lousy.

And yes, where I'm staying for this year I'm stuck on DSL. :davieddy:

[QUOTE=chalsall;395333]This can happen if there have been no changes over the day in the range queired, or if the query failed bacause of bad connectivity.

The machine which does the spidering for Mersenne.INFO is on a flaky Clueless and Worthless DSL here in Barbados. When I have some time (probably early March now; I'm able (and allowed) to type again, although slowly) I'm going to migrate the spidering to the US based machine which hosts GPU72 -- I have some "real" work to catch up on which pays for all of this.[/QUOTE]

"Clueless and Worthless" LOL

I'm happy to hear you're feeling better!

No rush, of course. I'm glad this issue is a symptom of the same issue and not another problem.

GPU to 75!
 

Just a brief update...

The TF'ing to 75 for the "Cat 3" range (read: 66M and above) is going fairly well. We're not /quite/ keeping up, but are mostly.

Interestingly, since George moved the "churners" down to the DC range, very few Cat 4 assignments are issued any more. All of these are being given at 75 bits.

So people know, in addition to trying to keep up with the Cat 3 LL'ers, we're also having to try to keep up with the P-1'ers (mostly in the Cat 4 range) to at least 74.

Thus, those who choose "Let GPU72 Decide" will sometimes see assignments in the 66M range to 75, and sometimes in the >71M range to 74 or 75.

As always, thanks for all the cycles everyone (and a special "call out" for those who have brought resources back to the project and/or have recently joined). If anyone has anything more to bring, it would of course be appreciated... :smile:

P.S. Please do note that soon Primenet will be in control of TF assignments. But the work (read: code) I need to do for that is a bit more than I can handle at the moment (I'm still down at less than 10 words per minute); and I have some important "paying work" to catch up on. Implementing the above was simple -- just instructing my current code what to do.

[QUOTE=LaurV;392251]Hey Chris, could you bring in 42/43M?
There are no DC assignments free to 70 bits for my 7970.[/QUOTE]

Since the 70 bit assignments will be exhausted in six months, I am curious about what kind performance drop you experience beyond 70 bits. How much does GHz-d/d drop?

[QUOTE=James Heinrich;395335]Some people love it, but I've never in 20 years had a good experience with DSL. In my house it stands for Dis Service is Lousy.

And yes, where I'm staying for this year I'm stuck on DSL. :davieddy:[/QUOTE]

Luckily for me I live in a cabled area and can get 152 down, only 12 up though...

[QUOTE=Gordon;395865]Luckily for me I live in a cabled area and can get 152 down, only 12 up though...[/QUOTE]

bps?

Anything under 10mbit/s I would consider slow, but maybe I'm spoiled with cable (90 mbit/s down - 9mbit/s up) and countrywide 4G (27 mbit/s down - 18 mbit/s up at my house). Yes, the uploadspeed on my phone is faster than the squeezed cable connection. If only the uptime would be better, we have had two days with no connection in the last two months and my ISP (Ziggo) is going to do maintenance tomorrow at 10:00 AM, could they pick a worst time?

I recently downgraded my internet connection from 175/175 Mbps to 50/50 Mbps to save some money. For day to day internet usage it's fine, but I really notice the 70% slower speed when doing offsite backups, copying things to and from work, and the like.

Having only 10 Mbps upload would cramp how I use the internet. I'm so glad those days are gone.

[QUOTE=Mark Rose;395869]Having only 10 Mbps upload would cramp how I use the internet. I'm so glad those days are gone.[/QUOTE]

I did my first year at university at 300 baud (dial-up). I was thrilled when I upgraded to 1200 baud for my second year.

Then I dropped out, because I was bored....

[QUOTE=chalsall;395874]I did my first year at university at 300 baud (dial-up). I was thrilled when I upgraded to 1200 baud for my second year.

Then I dropped out, because I was bored....[/QUOTE]

When I was young I was kicked off the computer because I "didn't go outside enough". I managed to cobble together a 16 MB, 40 MHz 386 AMD clone and got myself back online with.... a Hayes 2400B modem. Browsing the web was an exercise in futility even with images off, but at least I could chat (with considerable latency) and do text email.

It's interesting to think I had a home connection six orders of magnitude faster in only 15 years.

[QUOTE=Mark Rose;395869]Having only 10 Mbps upload would cramp how I use the internet.[/QUOTE]And the one time I was able to get the incredible amazing 2 Mbps up I was absolutely thrilled. Right now I'm on 15down/0.5up and consider it at least equal to (if not better than) the average I've used over the past 3 years.

I just speed tested at 31 Mbps down, 80 up -- though speedtest is known to not be a great metric. I can pretty consistently achieve 65+ Mbps down for hours on end with Torrent.

[QUOTE=chalsall;395866]bps?[/QUOTE]

152 Mbps

[QUOTE=Mark Rose;395869]I recently downgraded my internet connection from 175/175 Mbps to 50/50 Mbps to save some money. For day to day internet usage it's fine, but I really notice the 70% slower speed when doing offsite backups, copying things to and from work, and the like.

Having only 10 Mbps upload would cramp how I use the internet. I'm so glad those days are gone.[/QUOTE]

I'd love a symmetric connection, how much was the 175/175 ?

[QUOTE=Mark Rose;395834]Since the 70 bit assignments will be exhausted in six months, I am curious about what kind performance drop you experience beyond 70 bits. How much does GHz-d/d drop?[/QUOTE]
About 10-12% in the worst case (that is me): mfakto, cgn card, and Bdot does not release his long promised new kernels for mfakto :razz: I do DC to 70 only, and only with a 7970 card. It averages[URL="http://www.gpu72.com/reports/worker/2423ae6e8f696d5e7d1447de91ca35a6/"] 550GHzD/D[/URL]. If I move to 71/72 bits, I lose about 55G. That is why I bothered Chris. Of course, I won't move to 71, because, if I consider the "time unit" (tu) the time needed to find a factor at current bitlevel, then 71 would be ~2tu, and 72 would be about 4tu, but doing DCLL for a 41M expo take between 1.6tu and 1.8tu. That is why I was bothering Chris. For Nvidia cards, with the new mfaktc, it does not seem to be any loss, or well... it is an insignificant loss (on 580s and Titans).

In average (as people use different cards) we talk about few percents, like 2-3, max 5, maybe.

[QUOTE=Gordon;395900]I'd love a symmetric connection, how much was the 175/175 ?[/QUOTE]

Well I'm in Canada, so it's still expensive (unlike Europe or Asia where such connections have been cheap for years). I was paying $90/month plus $30 more for unlimited traffic. The 50/50 connection is $70/month.

[QUOTE=LaurV;395902]About 10-12% in the worst case (that is me): mfakto, cgn card, and Bdot does not release his long promised new kernels for mfakto :razz: I do DC to 70 only, and only with a 7970 card. It averages[URL="http://www.gpu72.com/reports/worker/2423ae6e8f696d5e7d1447de91ca35a6/"] 550GHzD/D[/URL]. If I move to 71/72 bits, I lose about 55G. That is why I bothered Chris. Of course, I won't move to 71, because, if I consider the "time unit" (tu) the time needed to find a factor at current bitlevel, then 71 would be ~2tu, and 72 would be about 4tu, but doing DCLL for a 41M expo take between 1.6tu and 1.8tu. That is why I was bothering Chris. For Nvidia cards, with the new mfaktc, it does not seem to be any loss, or well... it is an insignificant loss (on 580s and Titans).

In average (as people use different cards) we talk about few percents, like 2-3, max 5, maybe.[/QUOTE]

Ahhh, okay. Makes sense :)

[QUOTE=Mark Rose;395911]Well I'm in Canada[/QUOTE]You live in the privileged part of Canada. The fastest internet I could buy here (if internet where I'm staying wasn't landlord-controlled) is 60up/10down/300GB for $75/mo, and I'm happy to have that available (I've never had internet that fast in my life, either up or down).

50 Up / 10 Down is about the fastest a person can get around here. Saskatchewan is very controlling of this kind of stuff so someone can't just come in and say "We'd like to offer 500 Mb/s Up and Down but only to residents of Regina and Saskatoon!" They need to offer roughly equal service to everyone within an area and it just becomes so unprofitable to provide that kind of bandwidth to even the #5 - #15 largest [STRIKE]colonies[/STRIKE] communities of Saskatchewan.

I know SaskTel supposedly offers cell reception to every square inch of the province and that's why the rates are so horrible even in the two big cities. The northern 2/3 are just so remote.


(For anyone interested, look up Prince Albert, Saskatchewan on a map. That is our third biggest city at about 43,000 and is the last relevant civilization going North.

We had some exchange students from Quebec which is quite close to Montreal and even Ottawa and Toronto, and then some US cities as well. Saskatchewan is deserted in comparison. We brought them to our Capital, Regina, which is a 2.5 hour drive south from Saskatoon, and then to Waskesiu Lake, which is a 2.5 hour drive north from Saskatoon with the second half being a windy trek through an endless (beautiful) pine forest. The best moment was when we told them that they'd reach the halfway point to SK's north border after an additional couple of hours)


:direction: x2

[QUOTE=James Heinrich;395914]The fastest internet I could buy here (if internet where I'm staying wasn't landlord-controlled) is 60up/10down/300GB for $75/mo, and I'm happy to have that available (I've never had internet that fast in my life, either up or down).[/QUOTE]

Are you sure you don't have the up/down backwards? Usually down is faster than up (but it's been a long time since I lived in Canada).

Edit: Oh, and it's been a long time since I had >1 Mb/s in either direction, until recently. Yay FLOW (boo proposed merger with LIME...).

[QUOTE=TheMawn;395940]Saskatchewan... The northern 2/3 are just so remote.[/QUOTE]Just like everywhere in Canada. You can drive 2350+km (most direct route) and not leave Ontario (from north of Red Lake to Quebec border) -- about the same distance as Los Angeles to Dallas for the Americans lurking -- and yet you still can't drive within 600km of Hudson Bay (except maybe in the winter on ice roads, but I digress). The [url=http://en.wikipedia.org/wiki/Golden_Horseshoe]Golden Horseshoe[/url] (aka Toronto++) has 7 million people in 10,000km², and the whole province has 12.8 million in 1,000,000km² -- that is 54.5% of the population in 0.938% of the area. No wonder that some can get 100+Mbps and others can't even get basic cell phone coverage.

... back on topic :blush:

[QUOTE=James Heinrich;395953]you still can't drive within 600km of Hudson Bay (except maybe in the winter on ice roads, but I digress)... others can't even get basic cell phone coverage.[/QUOTE]

Actually, I made phone calls while standing in Hudson's Bay last year.

It's true you can't drive to the west coast, but you can drive to [url=http://en.wikipedia.org/wiki/Chisasibi,_Quebec]Chisasibi[/url] on the east coast, where 3G cell service arrived 3 or 4 years ago. So of course I had to make phone calls while standing in the Arctic ocean, out from [url=https://www.google.ca/maps/place/53%C2%B047'10.4%22N+79%C2%B004'41.7%22W/@53.786216,-79.078248,298m/data=!3m2!1e3!4b1!4m2!3m1!1s0x0:0x0]the boat launch[/url] south of town, simply for the novelty factor. They were having a heatwave, 34°C the day I waded in, so the ~4° water was pleasant. It's a [url=http://i.imgur.com/mVlk7ij.jpg]pretty area[/url], but there's little to do besides tour a couple of gigawatt hydro-electric dams, and the bugs are ferocious.

Also, we need more 70->71 DCTF assignments at GPU72.

[QUOTE=chalsall;395945]Are you sure you don't have the up/down backwards?[/QUOTE]

I absolutely did. James started it though.

oops... :redface:

I don't know if this pertains here or to the Official ServerErrors thread.

Yesterday I requested a bunch of exponents to Gpu72 to test my new GTX 980.
While uploading the results.txt file to Primenet through the mnual results upload, I got this messsage in yellow:

[code]
processing: TF no-factor for M71279717 (2[sup]74[/sup]-2[sup]75[/sup])
Result type inappropriate for the assignment type. Processing result but not deleting assignment.
CPU credit is 53.6765 GHz-days.
[/code]

Any hints?

Luigi

[QUOTE=ET_;396062]processing: TF no-factor for M71279717 (2[sup]74[/sup]-2[sup]75[/sup])
Result type inappropriate for the assignment type. Processing result but not deleting assignment.[/QUOTE]According to the [url=http://www.mersenne.org/report_exponent/?exp_lo=71279717&full=1]exponent history[/url], You were assigned that exponent for P-1 and you submitted the result for that on 21-Feb-2015. The exponent was almost-immediately reassigned to someone else for LL, and after that you submitted TF results for 74-75, even though the exponent was no longer assigned to you. Unintentional "poaching", but no harm done.

[QUOTE=James Heinrich;396067]According to the [url=http://www.mersenne.org/report_exponent/?exp_lo=71279717&full=1]exponent history[/url], You were assigned that exponent for P-1 and you submitted the result for that on 21-Feb-2015. The exponent was almost-immediately reassigned to someone else for LL, and after that you submitted TF results for 74-75, even though the exponent was no longer assigned to you. Unintentional "poaching", but no harm done.[/QUOTE]

I see... I request P-1 from Primenet and TF on GPU72, I guess I took the exponent before the syncing, sorry for that.

Luigi

What is the overall success rate with TF to 75? I have still not found a factor and I'm pushing 300 attempts...

I've found 5, but I have no idea how many attempts

We turn in our first batch tomorrow. So far there are five 74→75 factors!

:max:

[QUOTE=TheMawn;397724]What is the overall success rate with TF to 75? I have still not found a factor and I'm pushing 300 attempts...[/QUOTE]

3/299 in the 66M range

1/50 in the 74M range

Would any of you mind giving me a couple of the exponents that you've found factors for in 74 -> 75? 0/300 is a bit concerning and I'd like to check that my hardware isn't rejecting 75 bits for some reason.

[CODE]
66545503
66428893
66422129
66362921
66341423
71641907
[/CODE]

Thanks.

[QUOTE=TheMawn;397724]What is the overall success rate with TF to 75? I have still not found a factor and I'm pushing 300 attempts...[/QUOTE]

8796 runs from 74 to 75; 106 factors found -- ~1/83.

Edit: Deeper query to include those assignments from below 74 and/or beyond 75: 12713 runs covering 74 to 75; 147 factors found within 74 to 75 -- ~1/86.5.

Updated Workers' Overall Progress report.
 

Just so everyone knows, I've updated the [URL="https://www.gpu72.com/reports/workers/"]Workers' Overall Progress[/URL] report to include TF'ing to 74 and 75 (and >75). As before, you can click on the column headers to see who's doing what to where. For [URL="https://www.gpu72.com/reports/workers/75/"]example, those going to 75[/URL].

I'm afraid it makes the table rather wide, but I know people have been wanting this data exposed.

I looked at your link, and clicked my name out of curiosity, and it took me to [url=https://www.gpu72.com/reports/worker/111,347/]here[/url], which gives me blank graphs for Anonymous. Not just me, all the links under the name are broken.

[QUOTE=James Heinrich;397774]I looked at your link, and clicked my name out of curiosity, and it took me to [url=https://www.gpu72.com/reports/worker/111,347/]here[/url], which gives me blank graphs for Anonymous. Not just me, all the links under the name are broken.[/QUOTE]

Yeah, thanks... SPE. Fixed.

[QUOTE=chalsall;397773]I'm afraid it makes the table rather wide, but I know people have been wanting this data exposed.[/QUOTE]

It still displays fine in a 1200 pixel wide window. Most screens can display it just fine. :)

I tried four 74 to 75 with already known factors and all four were successful so I'm now quite confident that this boils down to bad luck. Thanks again.

[QUOTE=chalsall;397773]Just so everyone knows, I've updated the [URL="https://www.gpu72.com/reports/workers/"]Workers' Overall Progress[/URL] report to include TF'ing to 74 and 75 (and >75). As before, you can click on the column headers to see who's doing what to where. For [URL="https://www.gpu72.com/reports/workers/75/"]example, those going to 75[/URL].

I'm afraid it makes the table rather wide, but I know people have been wanting this data exposed.[/QUOTE]

It's all cool! Thanks, Chris.

[QUOTE=kladner;397812]It's all cool! Thanks, Chris.[/QUOTE]

Thanks for double checking. :smile:

Only a few years ago economic policy was (wrongly) set because of a simple mistake in an Excel spreadsheet....

[QUOTE=chalsall;397817]Thanks for double checking. :smile:

Only a few years ago economic policy was (wrongly) set because of a simple mistake in an Excel spreadsheet....[/QUOTE]

And then there's a telescope mis-ground, missions (including commercial flights) failing based on confusion between metric and English systems, and wah dee doo dah. :davieddy:

[QUOTE=kladner;397821]And then there's a telescope mis-ground,[/QUOTE] Not so much. It had more to do with that mirror (and the smaller mirrors) and instruments being out of whack with each other because the company that delivered the mirrors was used to delivering [URL="http://en.wikipedia.org/wiki/KH-11_Kennan#Design"]spy satellites[/URL] that the telescope was derived from. The same grinding program was used for it as the sat's.

Years ago when I mentioned it and the fact the shuttle's cargo bay was design to fit a KH sat, to a friend that had a high level security clearance and had done early work on the STS, they were surprised that I knew about the KH sats. I talked to them again recently and mentioned that the Hexagon (KH-9) program was declassified, this surprised them again.