Topic: [XPM] [ANN] Primecoin High Performance | HP14 released! - page 85. (Read 397616 times)

Here's the .wer file (is there somewhere else to look?) ...

Version=1
EventType=APPCRASH
EventTime=130188002101076763
ReportType=2
Consent=1
UploadTime=130188002101476786
ReportIdentifier=90e6d4ed-f13f-11e2-b1be-001cc0fa6641
IntegratorReportIdentifier=90e6d4ec-f13f-11e2-b1be-001cc0fa6641
Response.BucketId=119483045
Response.BucketTable=4
Response.type=4
Sig[0].Name=Application Name
Sig[0].Value=primecoin-qt.exe
Sig[1].Name=Application Version
Sig[1].Value=0.8.3.0
Sig[2].Name=Application Timestamp
Sig[2].Value=51e81bc8
Sig[3].Name=Fault Module Name
Sig[3].Value=primecoin-qt.exe
Sig[4].Name=Fault Module Version
Sig[4].Value=0.8.3.0
Sig[5].Name=Fault Module Timestamp
Sig[5].Value=51e81bc8
Sig[6].Name=Exception Code
Sig[6].Value=c00000fd
Sig[7].Name=Exception Offset
Sig[7].Value=0000000000b63876
DynamicSig[1].Name=OS Version
DynamicSig[1].Value=6.1.7600.2.0.0.256.48
DynamicSig[2].Name=Locale ID
DynamicSig[2].Value=1033
UI[2]=C:\Program Files (x86)\Primecoin\primecoin-qt.exe
UI[3]=primecoin-qt.exe has stopped working
UI[4]=Windows can check online for a solution to the problem.
UI[5]=Check online for a solution and close the program
UI[6]=Check online for a solution later and close the program
UI[7]=Close the program
State[0].Key=Transport.DoneStage1
State[0].Value=1
State[1].Key=DataRequest
State[1].Value=iData=1/nDumpFile=//Upload//iCab//5e9e45b16d7a4d7494b6a009e3ffccbd-374eb1ba359e4434c1816b58e399719f-4-119483045-AppCrash64-6-1-7600-2.cab/nDumpServer=watson.microsoft.com/nResponseServer=watson.microsoft.com/nResponseURL=//dw//StageFour64.asp?iBucket=119483045&szCab=5e9e45b16d7a4d7494b6a009e3ffccbd.cab&EventType=AppCrash64&BucketHash=374eb1ba359e4434c1816b58e399719f&MID=BD26F0E7-CAC8-4180-81CD-D19C0A18337C/nBucket=119483045/nBucketTable=4/nResponse=1/n
FriendlyEventName=Stopped working
ConsentKey=APPCRASH
AppName=primecoin-qt.exe



thanks for the assistance!

UPDATE : Well this is all moot now. I found that I can run hp3 just fine as high as 6553600 (with close to same testnet results as hp5 @ 6144000 <-- but is horrible setting for hp3) and able to mine live. it's a considerable amount less on PPS, but block wise it was running @ 20 every 10 minutes on testnet (vs 29 every 10 min w/hp5 - no other sievesize topped 15).

He probably means primecoin's debug.log
I suspect there's an abort() call somewhere when the client is out of memory.
Doesn't look like the OS is killing it.

"The" log file? As in, like /var/log/messages or the output of dmesg?

If you only mean the program's own log, all that would be likely to show would be asserts and such where it deliberately shut itself down rather than go into an unexpected or untenable state...

If on the other hand main net needs more RAM than test net for a given sieve size, and the operating system shut you down for using too much RAM, the operating system's own log should say so somewhere...

In general if the operating system shut you down it should, somewhere, have said why.

-MarkM-

Sorry, I found nothing noteworthy in the log file (which is huge).

The client gives no error message, just shuts down immediately after "setgenerate true" is entered, nice and clean.

I did discover I'm getting orphans like crazy since 7/18 though.

Crashes is such a vague term. Aren't there any clues in logs or core-dumps or running under a debugger as to what exactly it is really doing that to an uninformed eye might appear as just generically a "crash" ?

-MarkM-

On testnet, I found that a sievesize=6144000 outperformed every other sievesize I tried (incl. 6000000 & 6246400) by 30% using an i7.  

However, when i try to use this sievesize mining live, h5 crashes as soon as mining starts (every time).  It's unfortunate because it showed such promise on testnet.  

Blocks were coming in 4-5 chunks at a time steadily, even hit 6 blocks in 1 minute (no other setting hit above 4 blocks/min).

Core count means absolutely nothing. It's the speed of each core that counts. Each core has a MUCH slower speed than each core of the one on your desktop.

Price also means nothing. It just means it's a really, really bad deal.

That's just the way it is.

http://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+E5-2420+%40+1.90GHz

he has -1... as in 1 thread per core

Been running for 72 hrs with 4000-4500 pps and haven't found a block yet. I think there is something wrong with my settings rather than just bad luck. Most people I see are hitting multiple times over a few days with similar pps.

how do you even change sievesize from the debugconsole, i dont use a conf file, mainly because i dont know how to make them

The only correct answer that I have right now is that this IS the question !
I usually find one block per 24h with sievesize=1M (i7 3770k @ 4400)
Did a 24h test run with 4M and did find a block quite fast, around 4 hours and nothing for the rest of the time.

So 4M works but I have no way to say if it is better or not than 1M :/ I just got the same result, one block a day. Once again randomness is the big problem.

The only way to progress on this subject would be to centralize data regarding this parameter for enough time to get reliable data.

I'm trying 2M for 24h just to see...

I like to keep my sievesize at 1M just to have one less variable to worry about. At this point I think is best to experience with different processors, VPS configurations, etc.

Ivy bridge i3 3225 3.3Ghz  2x32kb 8 way L1 data, 2x32kb 8 way L1 inst., Level 2 = 2x256kb 8 way, Level 3 cahche = 3MB 12 way.

Sieve size recommended anyone?

i've tried straight downloading the gmp libs, building from source, tried MIPS or whatever that other variation of the library was, no dice.

I've gotten passed the gmp.h failure, its now failing at gmpxx.h

EDIT: Got it fixed, had to frankenstein the source package and the static libs together. seems to be building fine, whats left is to check to see if its successful at all when running

Some of the other coins have the gmp.h file with them.  You can find the file by iteself and add it to the folder where it is looking for it.

I would check and let you know which ones, but I just reinstalled windows and am setting things back up.

we need some windows compiling instructions, the gmp dependency is screwing with me.

That's a good rate as far as I can tell.

I don't see much support on this thread for our situation owning these AMD 8-core chips, I have seen others with similar queries. I have an older bulldozer FX-8120 ran it up to 4.2Ghz similar pps and did not see a block for several days. I still think at this rate it is still a bingo game.

Too hard to say-again because of the random nature and me not doing the best to keep track.   Running everything at 2M now but no blocks in 24 hours..

Perfect, Thank you.