512MB is sufficient to do a good P-1.

I only have six 73-bit assignments left at the moment, so I've decided to finish them first. In the meantime, I've converted all four cores of my i7-3770 to P-1. I'll probably do the same for one or two of the Sandy Bridge machines at work tomorrow.

I've been trial factoring with an HD 6410D and I typically reserve 5 exponents at a time, completing them in just under a week at 20GHz Days/Day.

Am I completing this work too slowly? Should I take less work? Five exponents is almost nothing but with this "rush" you've had I would like to make sure.

[QUOTE=Jayder;329061]I've been trial factoring with an HD 6410D and I typically reserve 5 exponents at a time, completing them in just under a week at 20GHz Days/Day.

Am I completing this work too slowly? Should I take less work? Five exponents is almost nothing but with this "rush" you've had I would like to make sure.[/QUOTE]

Your GPU can only do a certain amount, and ALL cycles donated to the project are welcomed. I don't think that 5 assignments a week is a huge amount to have on hand. There are plenty more where they came from, so nobody is starving for work because of you. I would not worry about it. :smile:

Haha, no, I didn't imagine I was hogging the work; only that I might be delaying the release of a range back to PrimeNet, though that may not be how that works. I thought it might be possible that a range is only released when all exponents within that range has been completed.

I believe I have my answer though. :smile: It's unlikely to be of any consequence.

[QUOTE=Jayder;329069]I thought it might be possible that a range is only released when all exponents within that range has been completed.[/QUOTE]

Nope -- it's on a candidate by candidate basis once it's been taken to the appropriate TF level.

And, welcome to the project! :smile:

[QUOTE=chalsall;328977]I'm going to leave the default settings at 72 for at least another day to make sure we've got a few more days of lead time, then I'll move it back up to 73.[/QUOTE]

OK, so everyone knows, the default pledge level is back up to 73.

Also, the change to the system WRT MISFIT users has been made. Anyone who's "Option" setting is "What Makes Sense" will be assigned work to 73 regardless of what the pledge level is.

Again, anyone who instead wants to only go to 72 or 71, simply change the Option to anything else. Human interaction with the form remains the same, including the use of WMS. As in, if a human pledges to, for example, 72 that's what will be assigned for all Options including WMS.

[QUOTE=chalsall;329082]
Also, the change to the system WRT MISFIT users has been made. Anyone who's "Option" setting is "What Makes Sense" will be assigned work to 73 regardless of what the pledge level is.

Again, anyone who instead wants to only go to 72 or 71, simply change the Option to anything else. Human interaction with the form remains the same, including the use of WMS. As in, if a human pledges to, for example, 72 that's what will be assigned for all Options including WMS.[/QUOTE]

This is excellent; I think of it as the "what Chris wants/best for GIMPS" option.

I still can't fathom why there's a difference between the automated WMS option and the manual one. I thik it will only breed confusion for people new to the project who start with manual sttings and move to auomated one. Are there any upsides at all to them being different?

[QUOTE=Aramis Wyler;329087]I still can't fathom why there's a difference between the automated WMS option and the manual one. I thik it will only breed confusion for people new to the project who start with manual sttings and move to auomated one. Are there any upsides at all to them being different?[/QUOTE]

They're not really different.

When a human looks at the assignment form, the default pledge is (now) 73. If they simply enter the number of assignments they want and click "Get Assignments" they will be given "What Makes Sense" (currently the same as "Lowest Exponent"). It is only if they change any of the parameters do things change.

The reason I've made the change to the behavior only for MISFIT is MISFIT simply does a POST to the form requesting assignments -- it doesn't first do a GET for the form to see what the Pledge level should be. Thus until this change I couldn't actually give assignments which "Made Sense".

At the end of the day the results work out to be the same for humans and MISFIT robots.

An important thing to keep in mind is that WMS was built when there was no automation at all. Contrasted to a year ago, MISFIT has completely redefined mfaktX work. Because the one major intention of MISFIT is complete automation, the changes made G72 much more responsive to real-world demands. MISFIT users who use WMS can know that their automated system is doing what's best for GIMPS as a whole, while manual G72 users can still get exactly what they want.

Anyone wanting to move from manual to automated will understand the difference and reasons by just following the threads and development of G72/MISFIT.

Just a [URL="http://www.gpu72.com/reports/workers/p-1/week/"]thing[/URL] on P-1....

[SIZE=1][COLOR=Silver](on another thought... I actually won't be able to do much P-1... the systems that have enough memory? They had memory for another reason I just 'membered. Meh. :[SIZE=1]\[/SIZE] )[/COLOR][/SIZE]

[QUOTE=kracker;329120]Just a [URL="http://www.gpu72.com/reports/workers/p-1/week/"]thing[/URL] on P-1....

[SIZE=1][COLOR=Silver](on another thought... I actually won't be able to do much P-1... the systems that have enough memory? They had memory for another reason I just 'membered. Meh. :[SIZE=1]\[/SIZE] )[/COLOR][/SIZE][/QUOTE]

Is it the sort of situation where 'LowMemWhileRunning' would be of any use?

[QUOTE=kladner;329123]Is it the sort of situation where 'LowMemWhileRunning' would be of any use?[/QUOTE]

Mike had mentioned it, and I had looked it up in undoc but couldn't really understand, But what does it actually do? Does it reduce memory when a specified program is run etc?

[QUOTE=kracker;329124]Mike had mentioned it, and I had looked it up in undoc but couldn't really understand, But what does it actually do? Does it reduce memory when a specified program is run etc?[/QUOTE]

Personally I find MaxHighMemWorkers=2 much more useful.

[QUOTE=kracker;329124]Mike had mentioned it, and I had looked it up in undoc but couldn't really understand, But what does it actually do? Does it reduce memory when a specified program is run etc?[/QUOTE]

When the specified program(s) run, it switches any Stage 2 P-1 workers to the next available Stage 1 assignment. In line with Chris's comment, I have found that for my Photoshop work I get the best results with PauseWhileRunning on 3 workers, [U]and[/U] LowMemWhileRunning, [U]and[/U] MaxHighMemWorkers=4. LowMem alone was still too sluggish. This means that while Photoshop and its cousins are running I switch from 6 P-1 workers with 4 High Mem workers, to 3 Stage 1 workers.

In prime.txt:[INDENT]PauseWhileRunning=photoshop[3],bridge[3]
LowMemWhileRunning=photoshop,bridge
[/INDENT]In local.txt:[INDENT]MaxHighMemWorkers=4

[/INDENT]

MaxHighMemWorkers, 2, 3, or 4 what is the difference?

[QUOTE=kracker;329152]MaxHighMemWorkers, 2, 3, or 4 what is the difference?[/QUOTE]

The number of workers who can use the maximum allowed amount of RAM.

[QUOTE=kracker;329152]MaxHighMemWorkers, 2, 3, or 4 what is the difference?[/QUOTE]

Sorry. I should have explained my settings more thoroughly. To recap:

In prime.txt:
"PauseWhileRunning=photoshop[3],bridge[3]"
I have a hex core Phenom II, running P-1 on all six. This line causes three workers to pause when either Photoshop or Adobe Bridge starts up, and resume when they quit.


"LowMemWhileRunning=photoshop,bridge"
This makes the remaining workers switch to Stage 1 work if any of them is running Stage 2. When P'shop and/or Bridge shut down, S2 work will resume, though not necessarily on the same workers as before.

In local.txt:
"MaxHighMemWorkers=4"
This limits the number of workers which can be in Stage 2 at the same time. This allows more memory to be allocated to each instance in S2. Stage 1 completes more quickly than Stage 2. If the number of HighMemWorkers is half or less of the total workers, you will build up a backlog of assignments with S1 completed which are waiting for an opening on the S2 front. At some point then, you have to do something to let that backlog get processed. On the other hand, if the number of HighMemWorkers is greater than half of the total there is a good chance that the S2's will keep up with the S1's.

[QUOTE=kladner;329178]Sorry. I should have explained my settings more thoroughly. To recap:

In prime.txt:
"PauseWhileRunning=photoshop[3],bridge[3]"
I have a hex core Phenom II, running P-1 on all six. This line causes three workers to pause when either Photoshop or Adobe Bridge starts up, and resume when they quit.


"LowMemWhileRunning=photoshop,bridge"
This makes the remaining workers switch to Stage 1 work if any of them is running Stage 2. When P'shop and/or Bridge shut down, S2 work will resume, though not necessarily on the same workers as before.

In local.txt:
"MaxHighMemWorkers=4"
This limits the number of workers which can be in Stage 2 at the same time. This allows more memory to be allocated to each instance in S2. Stage 1 completes more quickly than Stage 2. If the number of HighMemWorkers is half or less of the total workers, you will build up a backlog of assignments with S1 completed which are waiting for an opening on the S2 front. At some point then, you have to do something to let that backlog get processed. On the other hand, if the number of HighMemWorkers is greater than half of the total there is a good chance that the S2's will keep up with the S1's.[/QUOTE]

Ahh, I get it now, tnx! :smile:

Just a status update...

At the suggestion of Jerry, I have added a new field to the LLTF and DCTF forms: GHz Days of work.

How this works is if a number is entered here (>0), the "Number of assignments" field will be ignored, and instead the requested number of GHz Days of work will be assigned (up to the nominal limit of assignments allowed per request; 100 for most people). Please note that at least one assignment will be issued, even if it's more than the GHzDays field specifies.

This is largely intended for spider use, but if anyone would like to see this implemented on the other forms (e.g. LL P-1) please let me know.

Also, Aramis... You will be pleased to know that (at least here) programmers listen to their users. Scott has come up with an elegant solution which will hopefully avoid any confusion for MISFIT users.

Lastly, I'm continuing to monitor the assignment load on Primenet. Yesterday the assignment count went back up to about 1,000 a day. If needed, I'll lower the Pledge level back down to 72 to ensure we're ready to release at that level. Also, the "What Makes Sense" option has changed to mean "Lowest TF Level" from "Lowest Exponent". The thinking is to ensure that the candidates at 72 bits are not reassigned to go to 73, so they will remain available to release back to Primenet if needed.

GPU72.com out.

I took a break from my usual LMH-TF work and LLTF'd 100 exponents 71-72, found 3 factors. If the demand remains high I can chew through another 100.

[QUOTE=James Heinrich;329365]I took a break from my usual LMH-TF work and LLTF'd 100 exponents 71-72, found 3 factors. If the demand remains high I can chew through another 100.[/QUOTE]

Please.

I've grabbed 200 then.

[QUOTE=James Heinrich;329367]I've grabbed 200 then.[/QUOTE]

Thanks!

What is more useful[SIZE=1](et needed)[/SIZE]? 70 to 71, 71 to 72, etc?
(right at the moment)

[SIZE=1][COLOR=Silver] EDIT: w00t looks like one of my laptops found a P-1 factor a[SIZE=1] hour or so ago.[/SIZE][/COLOR][/SIZE]

[QUOTE=kracker;329394]What is more useful[SIZE=1](et needed)[/SIZE]? 70 to 71, 71 to 72, etc?
(right at the moment)[/QUOTE]

Anything going to 72.

[QUOTE=chalsall;329395]Anything going to 72.[/QUOTE]

Thanks. :max:

It might be a good idea to lower the default pledge level back to 72 bits right now.

[QUOTE=chalsall;329364]I have added a new field to the LLTF and DCTF forms: GHz Days of work.[/QUOTE]

When I read that, I thought "Hey, just what I was about to suggest myself!". Turns out it is not, but the ideas fit together nicely. :grin:

What GPU72 is lacking is a way of telling how much work [in GHzd] of a certain work type you currently have assigned.

The Assignments page will give you the total GHzd you have assigned for (LLTF+DCTF) and (LL+DC) along with average performance since you joined the project and an ETA based thereupon. It does not have any information whatsoever on any of the four work types on its own.

The Overall Statistics page, as well as the individual work type reports will give you the assignment count, but not the GHzd assigned, for any of the four work types on its own.

So I would suggest that

(a) a "GHzd Assigned" column be added to the table on the Overall Statistics page,
(b) a "GHzd Out" column be added after the "Out" column in the work type reports (for all workers),
(c) per work type lines be added to the top table on the Assignments page, eventually, and
(d) the average rates and ETAs on the Assignments page be computed in a more meaningful way. I'd go with the greater of (all time average, 30 day running average), but there may well be better measures.

While we are at it, I'd like to have the overall data per 1M range added to the Factoring Cost/Percentage reports, i.e. simply not splitting the data by bit levels.

Also, on the Individual Factoring Cost report, an "Out" column (and potentially a "GHzd Out" column) may prove helpful in spotting results that have not been submitted or acknowledged yet. :wink:

[QUOTE=ckdo;329432]So I would suggest that...[/QUOTE]

Not asking for much, are you... :wink:

Everything you suggest are good ideas (if I understand them correctly).

The amalgamation of statistics at the top of the individual assignment report is a legacy thing which I can make smarter (read: break out for each individual work type). LaurV has asked for the same thing quite some time ago -- and more than once (:razz:).

I have just finished implementing your suggestion of GHz Days (shortened to GHzD) assigned to the [URL="https://www.gpu72.com/reports/workers/ghzdays/"]Workers' Report[/URL]. As one would expect, it shows the GHz Days assigned -- overall and for each individual work type.

[QUOTE=ixfd64;329400]It might be a good idea to lower the default pledge level back to 72 bits right now.[/QUOTE]

I think we're going to be OK.

"Only" ~330 assignments were issued yesterday. And we have enough of a "pledge" volume to both 72 and 73 that even if we do need to release at 72 we'll have the numbers available.

So everyone knows, "What Makes Sense" is back to being Lowest Exponent.

Just noticed the [URL="https://www.gpu72.com/reports/workers/ghzdays/"]GhzD graph[/URL]. Looks great! :smile:

I didn't think it was possible, but someone (Jerry) is [URL="https://www.gpu72.com/reports/workers/graph/1-10/"]about to pass Craig[/URL] in overall GHzDays Done. Craig, we understand the reasons, but we miss your GPUs.... :smile:

(This is an indirect way of also saying that I've changed the Workers' graphs to cover the last 16 months, rather than only one month.)

That sure gives an interestingly different perspective.

EDIT: Congrats, Jerry! Enjoy it while the southern summer lasts!

It's an honor :surprised, but really only because the leader has all but flat-lined since late October. We will be moving sooner or later, so I'll have some flat-line time too :yucky:

[QUOTE=chalsall;329507]but someone (Jerry)
[/QUOTE]
we think he bought the 4 gtx570 from xyzzy... :whistle:
somehow we expected then didn't we?

[QUOTE] the last 16 months[/QUOTE]
long waited for! it was about the time... [STRIKE]god[/STRIKE] science bless you!

[QUOTE=LaurV;329558]we think he bought the 4 gtx570 from xyzzy... :whistle:
somehow we expected then didn't we?


long waited for! it was about the time... [STRIKE]god[/STRIKE] science bless you![/QUOTE]
I passed xyzzy but didn't cause nearly this much fuss lol

[QUOTE=bcp19;329597]I passed xyzzy but didn't cause nearly this much fuss lol[/QUOTE]

I guess Place and Show aren't nearly as exciting unless you've got bets on them.

[QUOTE=Uncwilly;328761]Have seen a bump up of ~1/3 increase in the number of 332M LL assigned expos. We were running around 1500 for quite a while. Now it is 2200. I hope that they all do TF to at least 77 first.[/QUOTE]It is up to 2400 as of today. The TF assignments are only around 59, which is in the typical range. The number of 100M digit assignments that have yet to yield a result after a year or better seems quite high >97%

I've sent a report bug to the Contact Us on the home page several days ago. Is that address still monitored?

[QUOTE=RichD;330035]I've sent a report bug to the Contact Us on the home page several days ago. Is that address still monitored?[/QUOTE]

Yes. And thanks for the report. I just haven't had time to drill down on it yet.

[QUOTE=Uncwilly;329931]It is up to 2400 as of today. The TF assignments are only around 59, which is in the typical range. The number of 100M digit assignments that have yet to yield a result after a year or better seems quite high >97%[/QUOTE]
I don't see why this would be considered unusual, it took me right around a year to complete one and that was after moving it to a bit of a monster machine and devoting 2 cores (though it did have to have P-1 and a fair amount of TF done)

[QUOTE=bcp19;330138]I don't see why this would be considered unusual, it took me right around a year to complete one and that was after moving it to a bit of a monster machine and devoting 2 cores (though it did have to have P-1 and a fair amount of TF done)[/QUOTE]Most of them have not done a single bit level of TF during that year. Even on a slow machine with a single core, one can take the expo from 74 to 75 in a reasonable time period. That is part of yielding a result. If the folks had taken the exponents that they got to 77, the current factoring report would be quite different.

Is the site currently down?

[QUOTE=KyleAskine;330453]Is the site currently down?[/QUOTE]

It seems up right now.

(btw, long time no see! :smile:)

Been switching my minions to P-1.... currently about [SIZE=3]¾ [/SIZE]are running P-1 at around ~10-12 I believe a day. 'need to catch up to that "Chris Halsall" guy there :bump2:

The ranking on the reports is rather broken, e.g. [URL]http://en.gpu72.com/reports/workers/[/URL]:

[code]
87 ramgeis 999
Stocker33 914

115 Anonymous 46
Anonymous 31

128 mdettweiler 4
Jack Sandor 3
[/code]Those shouldn't share a rank.

Also on [URL]http://en.gpu72.com/reports/workers/ll/[/URL]

[code]
10 Chuck 4,249
11 Sid & Andy 3,303
15 Carsten Kossendey 3,158
12 Pete 2,936
13 markr 2,691
14 delta_t 1,849
[/code]Where does the #15 part come from?

There's probably more, those just caught my eye.

Those ranks were always broken (i.e. they were never right, that is why people don't really pay attention to the numbers anymore). If you sort by, say, assigned ghzd, you see what I mean. Even in the main sorting (by "done") I was on position 20, but my number still was 18, and in front of me was 19 and 20 (to what I complained in the past, long ago).

Upon further inspection, that GHzd total is way off, and the #15 ranking would be correct based on the actual GHzd total. My bad. :confused:

Is the server down? I can't get the page to load more than partially.

It appears back up now.

On the left menu, under "My Account", when you first visit gpu72.com and you're not logged in, two items do not appear on the menu:
* Individual Overall Statistics
* Help Translate

Is there any particular reason these two are excluded from the not-logged-in menu? I would especially appreciate the first one, since that's the first page I'm (personally) likely to look at, and I always get confused when I can't find the link.

[QUOTE=James Heinrich;331196]On the left menu, under "My Account", when you first visit gpu72.com and you're not logged in, two items do not appear on the menu:
* Individual Overall Statistics
* Help Translate

Is there any particular reason these two are excluded from the not-logged-in menu? I would especially appreciate the first one, since that's the first page I'm (personally) likely to look at, and I always get confused when I can't find the link.[/QUOTE]

Still works here without login.

[QUOTE=James Heinrich;331196]On the left menu,...[/QUOTE]
click on workers progress and then on your name (any other name) in the table. There are the same statistics.

Now I'm puzzled. Until I posted yesterday, whenever I'd visit any part of My Account on the site I'd be prompted to authenticate. No big deal since the authentication is saved in my browser. But now, after posting, I've tried accessing the site in 5 browsers, including 2 I'm certain I've never accessed gpu72.com from before, [i]and[/i] from another computer from which I've not logged in to gpu72.com

Chris: are you playing tricks on me? :smile:
I know there was discussion a long while back about changing the authentication system, has that come into play now? Are logins stored by IP now?

[QUOTE=James Heinrich;331230]Chris: are you playing tricks on me? :smile:[/QUOTE]

Nope. Are you me? :wink:

[QUOTE=James Heinrich;331230]I know there was discussion a long while back about changing the authentication system, has that come into play now? Are logins stored by IP now?[/QUOTE]

Absolutely not. The site is simply using HTTP Basic AUTH -- no IP; no cookies (with one exception detailed below). I want to migrate to Digest AUTH, but it turns out many people have their browsers remember their passwords, and the browsers don't know how to handle an AUTH type change.

I have no idea why you're not being prompted for a log-in. Are you [I][U]sure[/U][/I] you never logged in before with the other browsers/machines?

And, to answer your question about why you don't see the Individual Overall Stats link until you log in, it's because I exposed that with a bit of a hack. It's actually a publicly available link, but with your UID as part of it. The system doesn't know what the UID is until you've logged in.

[QUOTE=chalsall;331240]Are you [I][U]sure[/U][/I] you never logged in before with the other browsers/machines?[/QUOTE]As reasonably certain as I can be.

Just for some added certainty, I fired up my netbook (which I rarely use) and installed Safari (to be [U]certain[/U] it's a browser installation I have never [I]ever[/I] logged in with), went to [url]www.gpu72.com[/url] and there's my (with user hash) Individual Overall Statistics link on the left.

I will hazard one guess: my building switched ISPs yesterday, so it's possible this new behaviour I'm seeing is related: possibly the ISP is caching something? Although if I go to pages I've not been to from this ISP (e.g. Contact or What's New) I still see the Individual Overall Statistics link.

:unsure:

[QUOTE=James Heinrich;331242]I will hazard one guess: my building switched ISPs yesterday, so it's possible this new behaviour I'm seeing is related: possibly the ISP is caching something? Although if I go to pages I've not been to from this ISP (e.g. Contact or What's New) I still see the Individual Overall Statistics link.

:unsure:[/QUOTE]

Unsure as well...

Is anyone else seeing this behavior? I have not made any changes to the code base which would explain this.

I just connected (with permission :smile:) to my neighbour's wifi and verified I was on a different IP (wildly different, actually, despite being on the same ISP installed on the same day -- 216.221.*.* vs 72.38.*.*)

Using my netbook on Chrome, I loaded [url]www.gpu72.com[/url] on "my" IP and I appear to be logged in. I switched wifi networks to my neighbour's, loaded gpu72.com in a new tab, and I'm not logged in. So it's not computer or browser related, but somewhere in my ISP's chain of caching.

[QUOTE=James Heinrich;331253]I just connected (with permission :smile:) to my neighbour's wifi and verified I was on a different IP (wildly different, actually, despite being on the same ISP installed on the same day -- 216.221.*.* vs 72.38.*.*)

Using my netbook on Chrome, I loaded [URL="http://www.gpu72.com"]www.gpu72.com[/URL] on "my" IP and I appear to be logged in. I switched wifi networks to my neighbour's, loaded gpu72.com in a new tab, and I'm not logged in. So it's not computer or browser related, but somewhere in my ISP's chain of caching.[/QUOTE]
Try clearing your cookies. GPU72 lays down about 4 or 5 of them.

[QUOTE=swl551;331254]Try clearing your cookies. GPU72 lays down about 4 or 5 of them.[/QUOTE]

All for Google Analytics. Not for authentication.

[QUOTE=swl551;331254]Try clearing your cookies. GPU72 lays down about 4 or 5 of them.[/QUOTE]Remember, I already tried a brand new browser install which guaranteed precludes the possibility of cookie issues:[QUOTE=James Heinrich;331242]Just for some added certainty, I fired up my netbook (which I rarely use) and installed Safari (to be [U]certain[/U] it's a browser installation I have never [I]ever[/I] logged in with), went to [url]www.gpu72.com[/url] and there's my (with user hash) Individual Overall Statistics link on the left.[/QUOTE]

[QUOTE=chalsall;331243]Unsure as well...

Is anyone else seeing this behavior? I have not made any changes to the code base which would explain this.[/QUOTE]

I have always been able to see my individual user stats but always had to login to see assignments or request work. I never thought about it till now.

[QUOTE=bcp19;331310]I have always been able to see my individual user stats but always had to login to see assignments or request work. I never thought about it till now.[/QUOTE]
Me too. Almost never accessed "individual bla bla", I always went through "workers progress" (to which I have bookmark) and click my name on the list. It never asks me to login for this. When I request work or I want to see the work done, etc, I need to login. Can't do this from job, don't know the password (stored in computers at home).

[QUOTE=LaurV;331338]... Can't do this from job, don't know the password (stored in computers at home).[/QUOTE]

Change jobs... problem solved!

Wow, that was easy..

[QUOTE=LaurV;331338]Me too. Almost never accessed "individual bla bla", I always went through "workers progress" (to which I have bookmark) and click my name on the list. It never asks me to login for this. When I request work or I want to see the work done, etc, I need to login. [/QUOTE]

Same here. It´s the only link under "My Account" that works without me logging in.

[QUOTE=lycorn;331509]Same here. It´s the only link under "My Account" that works without me logging in.[/QUOTE]

As it's supposed to. As I said above, it's actually a publicly available page.

If you see that link available it means your IP hasn't changed since the last time you logged in.

19 P-1 done today(so far), 3 factors! Is that normal or am I lucky?:max:

[QUOTE=kracker;332300]19 P-1 done today(so far), 3 factors! Is that normal or am I lucky?:max:[/QUOTE]

You're lucky.

But please keep P-1'ing!!! :smile:

[QUOTE=chalsall;332304]You're lucky.

But please keep P-1'ing!!! :smile:[/QUOTE]

I will. But I might swap back to DC if [URL="http://www.mersenneforum.org/showthread.php?t=17835"]this[/URL] makes it useless on CPU...

[QUOTE=kracker;332313]I will. But I might swap back to DC if [URL="http://www.mersenneforum.org/showthread.php?t=17835"]this[/URL] makes it useless on CPU...[/QUOTE]

Yes. When both stages of P-1 becomes viable on GPUs, you should switch your CPUs back to DCing.

And I'll have to hand over the second half of my unborn first born. And, yes, I'll provide a full night in a hotel in the St. Lawrence Gap. I'll even buy dinner.

I'm happy to fulfill my pledge if this can be accomplished.

As to the trip here, I hope container shipping is OK...

[QUOTE=chalsall;332318]Yes. When both stages of P-1 becomes viable on GPUs, you should switch your CPUs back to DCing.

And I'll have to hand over the second half of my unborn first born. And, yes, I'll provide a full night in a hotel in the St. Lawrence Gap. I'll even buy dinner.

I'm happy to fulfill my pledge if this can be accomplished.

As to the trip here, I hope container shipping is OK...[/QUOTE]

If you mean for the dinner, money will suffice.. Don't ship it, please :smile:

...Just kidding.:razz:

EDIT: Wait... how do I get to Barbados for that night? ....What did you just say?

Personal LL total <> top LL total

I think it is related to me unassigning being counted as completed.

Thought I did it right this time. I unassigned from the client only.

It is. I had dropped the remaineder of what I had picked up when we were emergency rolling to only 72, and gained about 10,000 GhzDays that day. :)

[QUOTE=Aramis Wyler;332451]It is. I had dropped the remaineder of what I had picked up when we were emergency rolling to only 72, and gained about 10,000 GhzDays that day. :)[/QUOTE]

OK. I obviously am a Stupid Programmer, and made an Error... :smile:

I'll have some cycles to look into this this weekend.

(No data was permanently harmed during the results of this error.)

[QUOTE=chalsall;332471]I'll have some cycles to look into this this weekend.
[/QUOTE]
Oh no! Please don't look into it this weekend, I just need 3-4 days more to be able to point you to [URL="http://www.gpu72.com/reports/workers/saved/"]this link[/URL] and say "see what happens when you mix the saved time of LLTF with the saved time of LMHTF? you come on the second place!" :razz:

[QUOTE=LaurV;332477]Oh no! Please don't look into it this weekend, I just need 3-4 days more to be able to point you to [URL="http://www.gpu72.com/reports/workers/saved/"]this link[/URL] and say "see what happens when you mix the saved time of LLTF with the saved time of LMHTF? you come on the second place!" :razz:[/QUOTE]

LOL... James has already given me an elegant solution to this "problem".

Thanks for reminding me.... :razz:

The "GHz Days per Day of Work Saved by x" doesn't seem to be that accurate, as [URL="https://www.gpu72.com/reports/worker/factors/d0a0a9610115a8227c686f0d8951998b/bydate/"]here[/URL] shows I found 2 factors, one at 03.08, and 03.09, but those two factors have appeared BEFORE 03.07 in the "saved" graph [URL="https://www.gpu72.com/reports/worker/d0a0a9610115a8227c686f0d8951998b/"]here[/URL].

(I remember the graph before, those two bars were NOT there at 03.05 and 06 before the two factors I found today and yesterday)

[QUOTE=kracker;332571]The "GHz Days per Day of Work Saved by x" doesn't seem to be that accurate, as [URL="https://www.gpu72.com/reports/worker/factors/d0a0a9610115a8227c686f0d8951998b/bydate/"]here[/URL] shows I found 2 factors, one at 03.08, and 03.09, but those two factors have appeared BEFORE 03.07 in the "saved" graph [URL="https://www.gpu72.com/reports/worker/d0a0a9610115a8227c686f0d8951998b/"]here[/URL].

(I remember the graph before, those two bars were NOT there at 03.05 and 06 before the two factors I found today and yesterday)[/QUOTE]

None of the charts are an accurate representation of what and when people have reported anymore, ever since 'somebody' decided it would be a good idea to compensate for those workers who only report results occasionally. From what I've seen all results get 'randomly' spread over several days, if you want an accurate representation of what you have done, you will have to keep your own records and create your own graphs :sad:

I want a GPU like Carsten has — he did 17,000+ GHz days yesterday according to the statistics.

[QUOTE=Antonio;332626]From what I've seen all results get 'randomly' spread over several days, if you want an accurate representation of what you have done, you will have to keep your own records and create your own graphs :sad:[/QUOTE]

When I implemented this, I did say that no information is lost. There's an "adjusted" field in the table which is used to spread the results out over the time they were out.

This has had a very useful effect of removing huge spikes in the overall graphs. However, if people want (and I get the impression they do) I could make available graphs for individuals without the adjustments.

[QUOTE=chalsall;332647]When I implemented this, I did say that no information is lost. There's an "adjusted" field in the table which is used to spread the results out over the time they were out.

This has had a very useful effect of removing huge spikes in the overall graphs. However, if people want (and I get the impression they do) I could make available graphs for individuals without the adjustments.[/QUOTE]

Ah, for like people who don't upload daily?

EDIT: Can/may I have them w/o adjustments? thanks!

[QUOTE=kracker;332650]
EDIT: Can/may I have them w/o adjustments? thanks![/QUOTE]

It would be interesting to see, at least.

I grappled with this chart logic for some time before realising that it doesn't randomly space things over random days. It randomly spaces them over the time between when you got the assignment and when you finish it.

So, if it takes you 10 days to finish an assignment because it sat in your queue for 9 days before you started on it, it will appear in the last 10 days (once your current day has caught up to the work of the previous day).

I'm not going to pretend that chart doesn't annoy the hell out of me, but now that I actually understand exactly what it's doing, and I now that I can control how accurate it is myself by increasing or decreasing the amount of data I keep in my queue, it is considerably less frustrating and more elegant feeling than it was in my mind before.

[QUOTE=chalsall;332647]When I implemented this, I did say that no information is lost. There's an "adjusted" field in the table which is used to spread the results out over the time they were out.

This has had a very useful effect of removing huge spikes in the overall graphs. However, if people want (and I get the impression they do) I could make available graphs for individuals without the adjustments.[/QUOTE]

Unadjusted graphs for me pretty please with bells on. :smile:

I'm aware that nothing is lost. Its just that the redistribution seems to be somewhat arbitrary at times!

How easy would it be to detect the people who return results daily (say over the previous 7 days) and provide them with unadjusted graphs automatically?

[QUOTE=Antonio;332687]
How easy would it be to detect the people who return results daily (say over the previous 7 days) and provide them with unadjusted graphs automatically?[/QUOTE]

Maybe just give them the choice of which.

[QUOTE=kracker;332688]Maybe just give them the choice of which.[/QUOTE]

OK, I don't have much time at the moment, so I've just done a quick-and-dirty change to the "Worker's" GHz Days Work Saved and Work Done graphs to use the real "Completed" date rather than the "Adjusted" date. You can see that for many people, the graphs become much less readable because of "spikes".

What I'll do (when I have more time) is make both types available. I'll have the "Adjusted" be the default, and the "Un-adjusted" graphs available as a click-through.

Thanks! It has a different perspective on almost everyone, hehe :smile:

[URL="https://www.gpu72.com/reports/worker/6e67460a77a11a707a665a6270df1a82/"]like..[/URL]

I pulled the plug on my TF and P-1 systems yesterday. I have one back up running for a while and my LL systems will run for a while longer. First time in 16 months the house is quiet.

[QUOTE=flashjh;332721]I pulled the plug on my TF and P-1 systems yesterday. I have one back up running for a while and my LL systems will run for a while longer. First time in 16 months the house is quiet.[/QUOTE]

Thank's for all your cycles Jerry!!! :smile:

[QUOTE=flashjh;332721]I pulled the plug on my TF and P-1 systems yesterday. I have one back up running for a while and my LL systems will run for a while longer. First time in 16 months the house is quiet.[/QUOTE]

:smile: Good wishes to you from now! When are you moving?







[SIZE="1"][COLOR="Silvr"](P.S.: There goes our 2nd largest P-1 user)[/COLOR][/SIZE]

We leave Cheyenne on Apr 19.

Thanks everyone. I'll be around once I get settled.

[QUOTE=Aramis Wyler;332670]I'm not going to pretend that chart doesn't annoy the hell out of me, but now that I actually understand exactly what it's doing, and I now that I can control how accurate it is myself by increasing or decreasing the amount of data I keep in my queue, it is considerably less frustrating and more elegant feeling than it was in my mind before.[/QUOTE]
Indeed. Thanks for objectivity. I did not catch this discussion when it started, otherwise I would say that I felt better with the "averaging" things. Some of the people newer here may not know that it was a time when "big guns" like Xyzzy or nucleon used to report their result once per month, and when you put on the graphic a bar of few hundred thousands GhzDays, everything near it looks like dust, so the graphics were looking like a white sheet with a single bar somewhere. That was the reason the "averaging" was introduced, to make possible for "mortals" like us to see other things on the graphics, beside a single tall bar in the middle. Luckily now few "big guns" left, they use misfit or other batches to report their work more frequently, so honestly I don't mind if we keep the averaging or just display the results as they are reported. But just to know the history, this was not done because someone wanted to cheat you on the scores, but because it was necessary. It is a bit to late now, as it was already changed, but if up to me, I would vote to keep the averaging.

[QUOTE=LaurV;332760]It is a bit to late now, as it was already changed, but if up to me, I would vote to keep the averaging.[/QUOTE]

The averaging will be re-introduced soon, with the option of viewing the "true" graphs for those who want to ensure their graphs exactly match their tabular reports.

Also, the "Overall" graphs will remain averaged. We still have too many people submitting their results only every week or so not to.

[QUOTE=chalsall;332802]The averaging will be re-introduced soon, with the option of viewing the "true" graphs for those who want to ensure their graphs exactly match their tabular reports.

Also, the "Overall" graphs will remain averaged. We still have too many people submitting their results only every week or so not to.[/QUOTE]

Ah nice, I liked the old averaged graph :)

BTW, the ordering of the [URL="http://www.gpu72.com/reports/workers/"]Workers' Overall Progress[/URL] table seems to be wrong, #12 comes before #9, #10 and #11, and there are many more errors.

[QUOTE=BigBrother;332804]BTW, the ordering of the [URL="http://www.gpu72.com/reports/workers/"]Workers' Overall Progress[/URL] table seems to be wrong, #12 comes before #9, #10 and #11, and there are many more errors.[/QUOTE]

Yeah, I know. This was a Stupid Programmer Error (SPE) (I am the programmer) when I added a requested feature without appropriate testing which caused a regression.

Unfortunately I'm *really* busy at the moment. Please trust me when I say that no information has been lost -- it's just a bug in my SQL (or maybe the Perl) which produces that report.