Primenet web design

[cuBerBruce]

Originally Posted by Madpoo

The CPU name shows up if you hover over the account name. I didn't think the CPU name would be of much use to anyone except the account holder, and in that case they're better off looking at their own CPUs report anyway. Or is there some reason people might want to scrape the CPU names of other people? That would be weird if you ask me. I guess if there was some machine that kept returning suspect results, it would be interesting to keep an eye on any other assignments that particular machine was working on, but I don't know... that's the only thing I could think of.

On the path to the milestone to prove M(32582657) is the 44th Mersenne Prime, a recently finished exponent was this one where "For Research" got a mismatch with the original LL test, and immediately did another test to confirm his own result. The "Computer name" column at least shows that the two matching tests were done (supposedly) by two different computers. (Some have argued that the double-check should be done by a different user than the original test. In general, the more independent the two tests are, the better.)

There might be other reasons for seeing the computer names such as detecting if a user seems to be reserving an unusually large number of LL tests on the same machine.

I am not aware if there have been any past discussions regarding privacy concerns over publicizing computer names. It looks to me that there has been at least one ANONYMOUS user that has revealed an actual name through the computer name info, though.

[chalsall]

Originally Posted by James Heinrich

Presumably Chis[sic] who can't spell "cumbersome".

Dislexics of the world, UNTIE! (Note: smelling mistake intentional....)

[chalsall]

Originally Posted by cuBerBruce

On the path to the milestone to prove M(32582657) is the 44th Mersenne Prime, a recently finished exponent was this one where "For Research" got a mismatch with the original LL test, and immediately did another test to confirm his own result.

Independant machines; full run done on each.

Anyone wishing to run a triple check are welcome to do so.

[Madpoo]

Originally Posted by cuBerBruce

I am not aware if there have been any past discussions regarding privacy concerns over publicizing computer names. It looks to me that there has been at least one ANONYMOUS user that has revealed an actual name through the computer name info, though.

In our company for example, they use the Dell model # and service tag to make the computer name, so if Prime95 were using those by default (and no, I do NOT run Prime95 on them, I've learned that lesson) that's the kind of info that probably shouldn't be public. Not that you could really do anything too bad with it, but it would probably be considered a disclosure of sensitive info.

Even my home computers, I name them with my and my wife's name and I don't need that seen by just anyone.

I guess if someone were worried about it they'd just set the CPU name to something besides the computername.

At any rate, it's been this way for however long and I don't really *mean* to rock the boat on that.

For the purposes of Primenet and detecting whether tests are run by the same machine or not, there is a machine specific GUID. It's better than tracking the CPU name at any rate. Not infallible, but it's a better way of tracking activity from a particular machine.

You can copy the entire set of prime files to another system and fire it up, and it'll use a different GUID unless you specify that extra option so it uses the same one... so even though it would be configured with the same cpu name, it'll be treated differently on the server.

Personally I'd feel better knowing double-checks were done by an entirely different team, but there's probably been some past discussion on that and I don't want to rehash old topics. :)

Let's just say that if you need the CPU name for some reason, it's still there, just not front and center.

[LaurV]

@Bruce:
PrimeNet's "computer name" has nothing to do with your real computer name. It is just a "tag" that you type in Prime95's "Test/PrimeNet..." menu, which is optional anyhow, and if you don't type anything there, some blank sh!t is assigned. Now, if you are lazy enough to avoid the effort of typing 5 random letters funny name there, that's your problem.

Related to "the same user testing an exponent", there are users and users. Some users are "trusted", in the sense that they will never report a fake LL result (i.e. one for which no test was done, I am not talking about hardware errors, for which reproducing the same residue, twice, trice, etc, is impossible). I believe Chris is such an user (and I am proud to consider myself such, but that is a different story).

Anyhow, there were discussions in the past that such exponents be marked as a "candidate for a triple check, with a low priority". I don't know if any action was taken, but I am also in the situation where, if I ran DCs and they didn't match, I re-run them till a match is got, and only report the both (good) results in the same time, after. This is to avoid increasing my "bad results count" in case my DC was in the weeds and the initial DC was right. (Yes, some of us care about these stupid metrics, you should all do! )

[LaurV]

Originally Posted by LaurV

and the initial DC LL was right.

Fixed. Hehe.
(getting stupid again, replying to myself)

BTW Madpoo:


Yarrrr...

[Madpoo]

Originally Posted by LaurV

BTW Madpoo:
Attachment 11844

Well at this point I think if I moved up in the rankings, y'all would be like "well, he has access to the database, so..." LOL

I'm happy to drive in the slow lane and let everyone else pass me by.

I could share a recent experience... my new laptop has an Nvidia Quadro FX 3700M (Thinkpad W700ds). I thought, hey, this thing can do GPU factoring, so I tried that out. It must have overheated and it froze the display after about 1 minute. :) Totally locked up. I should have known...it's just a laptop but the GPU does have it's own fan so I thought I'd give it a shot. Lesson learned.

So yeah, I'll leave just my one desktop doing LL work, plus the handful of servers I've done benchmarks on.

[tha]

The recent results table is displayed for a moment on my iPad without the black lines but with the alternating blue white background before being redrawn with the frame. I like it a lot better without the black raster because twice as many lines fit on a screen and because patterns are much more easy to spot. So, if we can have the tables without lines around the table cells, I would like that.

[Madpoo]

Originally Posted by tha

The recent results table is displayed for a moment on my iPad without the black lines but with the alternating blue white background before being redrawn with the frame. I like it a lot better without the black raster because twice as many lines fit on a screen and because patterns are much more easy to spot. So, if we can have the tables without lines around the table cells, I would like that.

There's a "flash of unstyled content" on that table due to the large size... the table sorting plugin kicks in once the table itself is all there so it renders using some default stuff before the table sorting skin is applied. Yeah, it's not perfect.

On some other page I "fixed" that by hiding the entire table until the page is loaded, then used Javascript to unhide, but it's kind of a lot of work and then there's always that one guy with Javascript disabled... :) (Just for Retina I put some CSS styling in a no code so the table will show up anyway... you're welcome).

I guess if you like the look without all that, you could disable Javascript for that page. LOL

[retina]

Originally Posted by Madpoo

There's a "flash of unstyled content" on that table due to the large size... the table sorting plugin kicks in once the table itself is all there so it renders using some default stuff before the table sorting skin is applied. Yeah, it's not perfect.

On some other page I "fixed" that by hiding the entire table until the page is loaded, then used Javascript to unhide, but it's kind of a lot of work and then there's always that one guy with Javascript disabled... :) (Just for Retina I put some CSS styling in a no code so the table will show up anyway... you're welcome).

I guess if you like the look without all that, you could disable Javascript for that page. LOL

If you were using CSS "properly" then you wouldn't need all these hacks with JS to "correct" things. Hiding stuff until JS has fixed everything is doing-it-wrong™. Anyhow as least I can still see things at the moment so your hacks are working fine.

[Madpoo]

Originally Posted by retina

If you were using CSS "properly" then you wouldn't need all these hacks with JS to "correct" things. Hiding stuff until JS has fixed everything is doing-it-wrong™. Anyhow as least I can still see things at the moment so your hacks are working fine.

I would posit that disabling Javascript in your browser is doing-it-wrong™
What you're telling me is that showing a spinner while content loads is hacky. Oh... you!

I only have that *feature* on one page, and now I've forgotten which one. It was more to see how well that approach would work... it was something else where it's just dumping a lot of data.

Oh, it's in the "Found factors" report. Sometimes a really large range is requested and it just takes a long time to load. It's not the server's fault, honest.

I challenge you to (with Javascript *enabled*) fail to appreciate the beauty of showing a brief spinner while the table loads on a page like this:
http://www.mersenne.org/report_facto...B1=Get+Factors

There's no way in pure CSS to do that. Javascript can detect when the document is loaded and remove a class... try doing that in just CSS. :)

One of these days we're going to finally convince you that Javascript is not a bad thing. One feature at a time, we'll wear you down.