[ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX]

[bigjme]

just turned on debugging (-D)

10/10 shares without a single issue, strange how all i did was turn debugging on?

EDIT:
rejected share 1-5: Duplicate share
all happened together

[1714972051]

1714972051

[1714972051]

1714972051

[1714972051]

1714972051

[cbuchner1]

just turned on debugging (-D)

10/10 shares without a single issue, strange how all i did was turn debugging on?

there is an extra switch (no short form) to enable stratum debugging specifically. check the --help

[cbuchner1]

rejected share 1-5: Duplicate share

oh wow. Maybe it is a problem with my nonce scanning code.

you guys could insert this after pdata[19] = result; in maxcoin.cpp

applog(LOG_INFO, "GPU #%d: %s submitting nonce $%08x to server...", device_map[thr_id], device_name[thr_id], result);

if you see duplicate nonce values in short succession I will have to do some more testing and debugging.

[bigjme]

at 21/29 now

all rejects are duplicate shares
they always happen together

just had 3 one after another
then stratum gets a new block and its back to yays

EDIT

little more info, i just had a share rejected, stratum found a new block, and it took 10 seconds from finding the new block before cudaminer did anything else

just got 8 duplicate shares 1 after the other
new block found, and back to yays again

[cbuchner1]

why are we even doing this with maxcoin at 0.0017

on the other hand... this might be the time to BUY

[Stanr010]

almost at 50% rejected now, this is aweful lmao

yeah why is that? why do I have two workers accapting shares at 95% and one at 84%? And why does this stay so consistent when restarting cudaminer?

Christian

What's your config for that 95% acceptance rate? You're talking about your 780 right? I'm getting 60% on the ti and 75% on my 780's.

a mix of 3 GPUs, totaling 300 MHash/s

cudaminer.exe --algo=keccak -d gtx780,gtx660ti,gt640 -i 0 -m 1 -l K2304x32,K2688x32,K768x32 -L 1024 -o  stratum+tcp://max.netcodepool.org:5555 -O secret

O.o you're running 2304x32? gosh...

[bigjme]

yes it is low but im more interested in making sure my compiled cudaminer is properly working :p
and why it has so many issues as if its not fixed we may get this problem again on another coin

[cbuchner1]

O.o you're running 2304x32? gosh...

the latest github allows this. I had twice that value before, but it slowed down my video display on the GTX780
My porn wasn't playing smoothly.

[bigjme]

the latest github allows this. I had double that value before, but it slowed down my video display.
My porn wasn't playing smoothly.

haha use a laptop ;-)

[ManIkWeet]

MAX is better than YAC atm so yeah we should still mine it :p

Code:

[2014-02-09 17:51:20] accepted: 10/15 (66.67%), 47659 khash/s (booooo)
[2014-02-09 17:51:38] GPU #0: GeForce GTX 780, 163178 khash/s
[2014-02-09 17:52:04] GPU #0: GeForce GTX 780, 0.00 khash/s
[2014-02-09 17:52:04] GPU #0: GeForce GTX 780, 46601 khash/s
[2014-02-09 17:52:16] GPU #0: GeForce GTX 780, 162361 khash/s
[2014-02-09 17:52:16] accepted: 11/16 (68.75%), 162361 khash/s (yay!!!)
[2014-02-09 17:52:19] GPU #0: GeForce GTX 780, 161253 khash/s
[2014-02-09 17:52:20] accepted: 12/17 (70.59%), 161253 khash/s (yay!!!)
[2014-02-09 17:52:43] GPU #0: GeForce GTX 780, 162387 khash/s
[2014-02-09 17:52:43] accepted: 12/18 (66.67%), 162387 khash/s (booooo)

This is on 1gh.com; I switched because it was the same terribleness on max.netcodepool.org.
Compiling your print thingy now...

[cbuchner1]

maxcoin soon to rival mincoin on mcxNOW

not sure if this extra if condition around "begin work on next CUDA stream" would help, maybe.
EDIT: updated this once more

Code:

		nonce[nxt] = n+1; n += throughput;
		if ((n-throughput) < max_nonce && !work_restart[thr_id].restart)
		{
			// begin work on next CUDA stream
			cuda_do_keccak256(thr_id, 0, cuda_hash64[nxt], nonce[nxt], throughput, validate);
		}

the same fix could be applied in much the same way to scrypt.cpp and scrypt-jane.cpp by the way.
it prevents unnecessary work from being given to the GPU.

[Stanr010]

When I try that config of yours, it give a full crash of cudaminer when it trys to do the 32 hash warp.

EDIT no wonder you've got the latest build. ETA on new revision for windows X86?

[bigjme]

maxcoin soon to rival mincoin on mcxNOW

lmao thats about right
but when your use to mining doge.... its still not bad haha

[djm34]

maxcoin soon to rival mincoin on mcxNOW

It seems I will cherish during a very long time my 10 remaining max before selling them

[cbuchner1]

EDIT no wonder you've got the latest build. ETA on new revision for windows X86?

why should I release a build that has a very much broken stratum ;-/

[Mirkic7]

Well, it went a bit up to 1.71 atm.
Guess not for long.

Speaking of rejects.. dunno.. still have very low rate. Both Nvidia/ATI. Which is good

->> http://maxpool.sytes.net <<-
Need more workers!
0% fee, stratum
cgminer/cudamier/cpuminer supported
Correct HashRates
Based on MPOS

1st block finder gets 1 MAX bonus!

[Stanr010]

EDIT no wonder you've got the latest build. ETA on new revision for windows X86?

why should I release a build that has a very much broken stratum ;-/

Clearly, you should fix stratum before releasing a new version but soon 

[bigjme]

every nonce value is different on all the rejects....
i should know, 6 one after another :-(

[cbuchner1]

every nonce value is different on all the rejects....
i should know, 6 one after another :-(

Yeah WTF then are the pool operators cheating?

On a different note the fix I posted above has brought my accept rate to 100%.
Explain this. ;-)

[Stanr010]

I'm beginning to wonder if we should start holding onto MAX for the price spike when it hits cryptsy

Selling for under 0.002 feels like just throwing it away but I'd hate to hold onto coins if it's still go down and never come back up.