Topic: [ANN] Signatum - New Algorithm - Fair Launch - No Premine - page 300. (Read 823878 times)

What is it with you and goats?

Interesting findings with ccminer versions.  Very interesting.

I had been using various versions of the 'krnlx' ccminer for the last 5 days or so and when I plugged in my MINER-reported hashrate into the profitability calculators, my actual results were immediately proven to be significantly LESS than what the calculators were saying.  Using 'whattomine', and the krnlx_170810 version (32 bit Cuda 7.5) I was getting a MINER reported overall hashrate of 255 Mh/s.  But, my rewards were coming at a 190 Mh/s rate.  Ie., if I adjusted hashrate in the calculator to match my rewards, I had to go from my reported rate of 255 all the way down to 190.

So, I have these 12 GPU's getting 9.09 coins per hour while the calculator says I should be getting 13.5. 

Reading a post that I initially through was born of paranoia, insinuating that it was possible that some of these 'faster' modded ccminer versions might be skimming hashrate and delivering it elsewhere, I decided to go with the version that poster recommended, the palingmod versions.

So, I fired that up using the exact same settings for cards and the exact same .bat files.  And, this is what happened...

My MINER reported rate went from 255 to 233. 


At 233, whattomine tells me to look for 12 coins per hour.  That is one coin every 5 minutes.  I have been running on palingmod for 138 minutes.  I am told by whattomine, adjusted for difficulty mind you, that I should have 27 coins.  I have 25.  At my previous 'rate' of 255 I was losing 33% of what I was 'supposed to get' when using the krnlx version.  I am now off by about 8%.

So, I am running SLOWER and getting MORE. 

I know this is not a static environment, but I can tell you that since day one of mining this, I have been consistently OFF by 30-35% on the 'expected' numbers. 

If someone has the expertise to sniff out all the network traffic that the krnlx versions are making, it might be interesting to see if anything suspicious is taking place.

For the last 30 hours I have been within 5 minutes every 50 coin payout.  I know the difficulty shifts around.  I know this is not scientific.  But I am still curious.  For some reason the 'what to mine' sites seem to be giving inflated numbers.  Maybe the inflation IS coming from elsewhere.

So far, my hashrate is down from 255 Mh/s to 233 Mh/s but my average payout for the last 2 hours is 10% HIGHER than the average of the previous 30.

One would not expect that.

Difficulty is not stable, so you won't get a meaningful result.

Only way to test it alone is to perhaps to make a fork of the entire coin and premine a set amount, then reset and test again. Might be way too much effort.

Otherwise have identical setups in the same pool, mine for X amount of time using different miners at the exact same time, repeat results 3-10 times and then average results to compare.

Here's what I am going to do...

With my setup it is taking me a consistent 5:45 to mine 50 coins.  That is using the ccminer_krnlx_170810 build I mentioned earlier.  I am 3 coins away from another 50 and when that is awarded I am going to change everything over to palginmod.  If the established ratio of miner reported hashrate and awards suddenly diverges when we will have some evidence that something nefarious might me up...

Interesting nice work.

As I mentioned before spmod cannot be trusted as the source is not available. It could be redirecting hash rate somewhere else (which you can in theory prove with network analysis) or more likely it's reporting false hashrate.

Ofcourse a single point of data is not enough.

I don't recommend using spmod, since it's technically a leaked private build (which should not be distributed at all) and because we cannot trust it due to not having any source available (potentially a GPL3 violation as well).

Instead use:

palginmod - https://github.com/palginpav/ccminer/releases/

or

krnlx - https://github.com/krnlx/ccminer-skunk-krnlx/

For me palginmod is the most stable on a GTX 1060 6GB.

Where are you getting your 'reported' hashrate?  From your pool or from your rig?  The reason I ask is that my pool (suprnova) never shows anything close to a static figure (or maybe I do not know where to find it...).  Also, my rig that has 5 1060's and just 1 1070 is getting 123.5 reported from the miner.  Various mods have gone from 108 to 123.5 which is a 'settled in' hashrate.  I would think 6 1070's should be topping 145.

My guess is that you might run a bit faster with the ccminer_krnlx_170810 versions that you should be able to find.  There were three versions in the zip file and the Cuda 7.5 32bit version did best for me.

Strongly agree, thank you for helping the community.

Nice review, true crypto OG right there

Recently this has been the only coin for miners in out there, and this one definitely has market maker. But we gotta wait till new blood gets shaken out.

How long does nova exchange take

so what I did:

• 2 identical rigs (6x undervolted gtx 1070 both at same clocks / hashrate when using an identical miner)
• both running on the same pool with static diff set to 1.0
• started miners exactly at the same time
• stopped the miners after one hour

apparently the skunk miner seems to work a bit better on my rig, but this comparison is more for checking the reported hashrate vs accepted shares. Since people feel the reported hashrate is not fair on the skunk miner.
Actually this test is running for a too short time to draw conclusions, but it gives an idea already. Next update will be after 12 hours of mining

I will do more testing. I will test some more miners on request. Please PM me when you want me to test a private miner (ie. spmod 6 or something else) or give me a link to a public miner if you want it to be tested.




my mistake: skunk = krnlx miner I will update the graphics on the next update

Agree , just switched all my rigs to CWI miner it is faster than cc and all my cards 85% power (I can save couple pennies with my 1080 Ti-s).

I like the activity of the dev and the community, the coin has a future, those who say that the coin is dying are newbee, even the legendary members are often mistaken, but here some noobs write nonsense

Most of developers want to get 10-20% developers fees and have full control over their coins. This coin has "No premine. No ICO. No Dev Fee."
May be this is the answer why some exchanges so slow to accept SIGT, but they promptly accept some questionable ICOs(pyramids).

Ok thank you for the suggestions. I will do a few more tests after I finished the skunk vs spmod 5
As I only have 2 identical rigs I have to do a sort of knockout test between each miner

update on skunk vs spmod 5 is here:

https://bitcointalksearch.org/topic/m.20825849

In my cfg, CwiGm is fast as krnlx, both are faster than palgin mod.
CwiGm is hardcoded to CWI pools, while you can use ccminer krnlx with any pool.
I hope CWI will unlock its miner soon.

Its headed back up.

you should also add the CWI Miner to your test, that one might be a surprise for you (and you don't have to pay 0.05 BTC for it)

Aho'! :-)

SIGT will die when it will die.... but I've never seen so much vultures flying.  

Everyone will die... but not today.
Today, if profitable, mine.
And stop mining if not profitable, not because someone tells you it is sinking.

But it is not a pet: if it's sinking, sell; profitable, hodl and/or mine.

Come on!

for Nvidia miners: please let me know if you want me to do this skunk vs spmod miner comparison:

https://bitcointalksearch.org/topic/m.20825849