OFFICIAL "SERVER PROBLEMS" THREAD

[xx005fs]

Originally Posted by chalsall

Short answer: No. Longer answer:

The closest thing available at the moment is the MISFIT Application for Winblows, or the mfloop script for Linux. These are quite reliable; many people use them in a "fire and forget" scenario.

Importantly, these won't use a "GUID", but if you only have one GPU this shouldn't be a huge problem. They will be provided with an "AID" (Assignment ID), if you fetch from Primenet. These will also be tied to your particular Primenet and/or GPU72 Account (whichever you choose to use), and so you'll see them on your assignments page, and receive the credit upon completion.

Hope that helps.

Thank you so much for the explanation.

[GP2]

By using this query, we can find all LL tests done by Simon Cunningham. He's the guy with the buggy machines in this thread.

There are two problems with this report page:

First, you have to enter a value for the upper limit (999999999). If you leave it blank, there is no output. That is not the behavior of most pages on mersenne.org

Second, the query sometimes misses some records.

In the bottom, under "Bad test results", I currently get 35741141, 47148053, 47185343, 49353571, 49369027, 51103477, 51794089, and new ones may be added as further double checks are done.

However I could have sworn the exponent 35741141 did not appear in the report a few days ago, even though it was tripled-checked years ago. And the very old exponents 7477103, 10519273, 14786729 still don't appear at all.

[ric]

Originally Posted by GP2

And the very old exponents 7477103, 10519273, 14786729 still don't appear at all.

Here they are...

[James Heinrich]

Originally Posted by GP2

There are two problems with this report page

@Aaron: while fixing that page, perhaps you could wrap some <label> tags around all the checkbox items, it makes it a bit user-friendlier. Or tell me to fix it.

[Madpoo]

Originally Posted by GP2

There are two problems with this report page:

First, you have to enter a value for the upper limit (999999999). If you leave it blank, there is no output. That is not the behavior of most pages on mersenne.org

If the upper limit is left blank, the default is to set it to the same as the lower limit (i.e. to show a single exponent). I think that's the same on other pages... basically acting as a simple landing page that won't spit out a bunch of data you don't want.

Second, the query sometimes misses some records.

In the bottom, under "Bad test results", I currently get 35741141, 47148053, 47185343, 49353571, 49369027, 51103477, 51794089, and new ones may be added as further double checks are done.

However I could have sworn the exponent 35741141 did not appear in the report a few days ago, even though it was tripled-checked years ago. And the very old exponents 7477103, 10519273, 14786729 still don't appear at all.

There is a reason for this, but it eluded me for a bit... It's because results from the user's old Primenet v4 account have the same display name, but it's different than their v5 account ID.

So... on the surface it may seem like the same account because the name is the same (and it is the same person in this case), but under the covers, a v4 and v5 account are different.

Even for people who linked their v5 account to their old v4 account, all that's doing is letting your account info and credits "roll up", but the older results are still stored using the version 4 info.

This has caused me a good deal of trouble when tracking bad results from particular users and also my side project to independently all of the self-verified work (I had to do extra lookups in the v4/v5 stuff to match things up...slows things down, made my queries hard to follow at times)

Speaking of v4/v5 stuff, we did recently remove the links to that page. George figured if they haven't linked their old account by now, they never will... It's been something like 13 years... something like that.

[ckdo]

Looking at my workload page today (which I do every couple weeks at minimum), I was rather surprised to see 5 TF-LMH assignments there assigned to "v4_computers" (192300497, 192441523, 192442409, 192512357, 192526459).

I wonder where these may come from. I have switched all my machines to PrimeNet v5 shortly after it went live. It's certainly not a machine that's under my control, or has been in the past decade or so.

I have never thrown away a hard drive; I still have all of them, functional or not (starting with sizes in the lower megabyte range). The only thing I am missing is a laptop that was stolen when my flat was broken into close to a decade ago.

But why would that thing now be talking to PrimeNet again when it remained silent for years? I'm absolutely clueless.

I guess since it's a v4 machine, there's no info on the hardware and whatnot to be gained from its communication?

[GP2]

Originally Posted by ckdo

Looking at my workload page today (which I do every couple weeks at minimum), I was rather surprised to see 5 TF-LMH assignments there assigned to "v4_computers" (192300497, 192441523, 192442409, 192512357, 192526459).

This is probably unrelated, but I noticed some recent LL tests by Anonymous using "v4_computers", which seemed kind of unusual. For exponents 47567353, 49615603, 52882421.

Until then, I think in recent years there was only one, non-anonymous user ("spica") still turning in fresh v4_computers results, for example 75759427.

[ckdo]

So the five exponents have evidently all been finished by "hkt" already. The question remains: Why do his v4 assignments show up in my account, and why is this phenomenon seemingly new?

[Prime95]

Originally Posted by ckdo

So the five exponents have evidently all been finished by "hkt" already. The question remains: Why do his v4 assignments show up in my account, and why is this phenomenon seemingly new?

Maybe you and your mystery helper have similar user ids and he mistyped his.

[retina]

Originally Posted by Prime95

Maybe you and your mystery helper have similar user ids and he mistyped his.

You mean I can submit results for other people just by knowing their uid?

That could turn out really good for them, or really bad for them, depending upon the motivation.

[Madpoo]

Originally Posted by retina

You mean I can submit results for other people just by knowing their uid?

That could turn out really good for them, or really bad for them, depending upon the motivation.

Well... yeah. Nothing new there. The only thing in your prime.txt file is your user id.

If someone wants to turn in results using your uid, more power to 'em. It'll be separated by the computer name/GUID.

User ids are forced to be distinct, but if some were close to each other, then, yeah... I guess someone could typo something and showing up in someone else's account.