PrimeNet 5.0 Upgrade

[Old man PrimeNet]

We have initiated the migration to PrimeNet 5.0. Please be patient. It will probably require several more days to complete the staging.

The computers will hover just fine until we switch on the 4.0 API support, and though some will coast to idle the vast majority will keep crunching productively. Once the 4.0 API is online they will auto-sync to the new server.

We'll post further upgrade instructions soon.

[jinydu]

Here is a screenshot of my v5 account page:

Not sure what "entile by stats type" means.

[Kevin]

Originally Posted by jinydu

Here is a screenshot of my v5 account page:

Not sure what "entile by stats type" means.

It's supposed to say "Percentile by stats type", but some of it gets cut off depending on how it happens to render in your particular browser.

[S485122]

On a 8GB quad core PrimeNet V5 does not accept the max memory usage set :

[Main thread Oct 26 11:02] Mersenne number primality test program version 25.7
[Main thread Oct 26 11:02] Starting workers.
[Comm thread Oct 26 11:03] Exchanging program options with server
[Comm thread Oct 26 11:03] URL: http://v5.mersenne.org/v5server/?v=0.95&px=GIMPS&t=po&g=22665033b6fd3fc74416c1a76061e0fa&DayMemory=7307&NightMemory=7307&ss=184
[Comm thread Oct 26 11:03] RESPONSE:
[Comm thread Oct 26 11:03] pnErrorResult=7
[Comm thread Oct 26 11:03] pnErrorDetail=parameter DayMemory: Invalid int value/precision '7307'
[Comm thread Oct 26 11:03] ==END==
[Comm thread Oct 26 11:03] 
[Comm thread Oct 26 11:03] PrimeNet error 7: Invalid parameter
[Comm thread Oct 26 11:03] parameter DayMemory: Invalid int value/precision '7307'
[Comm thread Oct 26 11:03] Visit http://mersenneforum.org for help.

The program (Prime95 64 v25.7) has no problems with that value (untill now... I am not yet busy in a P-1 stage 2)

Jacob

[Aexoden]

Originally Posted by uigrad

Once we upgrade to v25, is there a way to continue running separate assignments on each core? On both of my systems, I have 1 core running first-time LL, and the other cores are each running double check assignments, and I'd like to keep it this way if possible.

Yes, you can individually select what what you want each core to do, either via the program itself or via the web interface.

[Old man PrimeNet]

The PrimeNet V4.0 APIs are now enabled and running on PrimeNet 5.0. You can link your v4 account(s) to your v5 account.

All the new assignments and results are flowing from/to the 5.0 server. The traffic looks stable and busy catching up. When the v4 API is too busy to queue your call it will continue to return error 2250.

Credits are accumulating for work received from the v4_computers virtual CPU for each v4 account. You can claim your V4 credits by linking your old v4 account. This re-associates your v4 work credit (parked under ANONYMOUS) back to you.

[NBtarheel_33]

Hi all,

Anyone know if our linked v4 stats will show up under our lifetime v5 stats? I just linked to my v4 account, and got the huge credit to my "Last 365 Days Results", but right now, my "Lifetime Completed Results" only reflect the work I've done since joining v5 in July 2008.

Great job George, Scott, et al. on the new server. The look and feel of everything is great. Considering that this is all done on a volunteer basis, we really owe a tip of the hat to these guys for all their hard work.

[RMAC9.5]

My version 4 user id on my home PCs and all of the PCs that I have successfully "borged/recruited" for Prime95 since 1999 is RMAC9.5. My RMAC9.5 version 5 registration attempt was rejected by the server because of the "special character/period" in my user id. Please (at least unofficially) add a "." to the list of valid characters for version 5 user ids. Thanks in advance.

[cheesehead]

rudi m has reported in another thread ("v5 can't login" at http://mersenneforum.org/showthread.php?t=10844) that the W3C Markup Validation Service (http://validator.w3.org/) finds lots of errors at http://v5www.mersenne.org/.

Use

Originally Posted by rudi_m

http://validator.w3.org/check?uri=http%3A%2F%2Fv5www.mersenne.org%2F&charset=%28detect+automatically%29&doctype=Inline&group=0

... as rudi did, to see the report.

I suspect that some of the 197 errors and 70 warnings it reported just now when I tried it are inconsequential, while many others are related to "under construction" status and would be cleared up as migration/upgrade work proceeds anyway, but this ought to be on the to-do list for later attention.

[diamonddave]

I found some strange result in the "Recent Result" report on PrimeNet v5

Member Name          Computer Name    Exponent  Type     UTC Time Received    Days GHz-days  Result               
-------------------- ---------------- --------- -------- ------------------- ----- --------  -------------------------------------------------
ANONYMOUS            v4_computers     43593527  NF-PM1   Oct 27 2008  4:57PM   0.0   2.7374                                                                                                                                                                                                          
KriZp                Opt              272141603 NF       Oct 27 2008  4:57PM   0.0   0.0010  no factor from 2^58 to 2^61                                                                                                                                                                             
ANONYMOUS            v4_computers     43593527  NF       Oct 27 2008  4:57PM   0.0  86.3950

What scenario would give 86 GHz-Days of credit for factoring work when PM1 Factoring on the same exponent gives 2.73. and a LL test in that range give about 75 Days of credit

[Phantomas]

I'm quite sure that there exist some problem with casesensitivity.
When i request LL-Results for "phantomas" i got a 2 entry list

Unverified test resultsExponent,User name,Computer name,Residue,Error code (if any),Date found
41076619,Phantomas,JTc2D2333,24536CDD911947__,,2008-10-26 01:54
41077073,Phantomas,JTc2D2333,D5238604D19D79__,,2008-10-26 01:54
Last 2 hex digits of unverified residues are masked.

for uppercase user Phantomas. And that are new entries. I never used small phantomas except in the linkage (because Phantomas didn't worked)

When I request the list for Phantomas i get a long list for "Jörg Thomsen" (my name) with the old V4 Workunits

Verified test resultsExponent,User name,Computer name,Residue,Date found
1444481,Jörg Thomsen,florida_P200,0219B7B6C7D9B179,
1447379,Jörg Thomsen,florida_P200,2B91E1A20529DEC0,
1463947,Jörg Thomsen,SYSFE_P133,F2381D48F6FCB49B,
:
:
:

and between the two entries from above

41076619,Phantomas,JTc2D2333,24536CDD911947__,,2008-10-26 01:54
41077073,Phantomas,JTc2D2333,D5238604D19D79__,,2008-10-26 01:54