|
Testing V1.05 presently. If its good to go by tomorrow sometime, Ill post it up.
To prepare: linked on my post where I update the code are all of the miner apps I have used. I took the neoscrypt and lyra2 versions from the nicehash miner /bin folder, and named them ccminerNS.exe and ccminerlyra2.exe respectively.
And as goes the x11evo version and the 1.8.0 versions for the other two as it was before.
I didn't strip out the timer just yet.... but I took some code redundancy out and it looks much cleaner.
I ran each miner app and checked what default intensity was and entered it in all the algos that I haven't trimmed back; just to clean up the code segments.
skein is still removed from the main loop. you will find it's code and normalization values down at the bottom of the code when it's posted.
So far, much simpler than I expected it to be. The old bug of not detecting the active app is gone.
|
|
|
|
|
crackfoo: any chance you can add the ability to show last 24 hour earnings graph in different selectable time ranges? lets say last 2 hours, last 4 hours, last 8, last 12, etc. would be kinda nice. Maybe something like a dropdown menu....
Working at batch updates. Trying to have a complete package with individual batches for all algos so you can normalize yourself easily.
Things on the list:
1) Add optimized lyra2
2) Add optimized neoscrypt
3) Add nvidia NVPSMI P0 state force
4) Assess need for static application location
5) Remove timer, clean up recursiveness and active application detection
Will need feedback if the neoscrypt version doesn't work on non maxwell models.
Trying to add dynamic system path redundancy to be able to force video card p-state to P0 every loop. From this point on, The batch will require the default nvidia folder name in the %programfiles% [or whatever] folder, and more importantly, the batch will need to be run as administrator.
Running it as administrator has caused a slight issue in the past; it causes the default directory to be c:/windows/system32 and not where the batch was run from. So, I may create a specific folder hierarchy, or better yet, maybe copy it directly into the system32 folder and run from there.
I have some tasks to bite off for this. But I want it to smooth out all the issues on my plate in one swoop. Hopefully progress doesn't go slow.
Starting to wonder if learning to code an app in the windows interface might be worth it at this point.... Or maybe whip out borland Turbo C++ to create an app to take functions and pass results. From there the sky would be the limit and the batch would be a bunch of variables and a chain of calls and a simple loop..... hmmm.....
*edit* added links to all necessary files and data to the batch link in my sig. A lot better than searching back and forth. I commend the people whom have done what I just had to do to make it all in one place for you. -JK
|
|
|
|
|
To add perspective my father sells power to pge in the Bay Area at 13c/kWh... They charge like 28c. Insanity.
|
|
|
|
Hopefully the ppl who wanted it will come and mine some as well.
Thanks
Ill do some testing when I get back home tonight. |
|
|
|
So I think you are going to be stuck with higher RPM fan's unless you underclock it. And I would monitor the heck out of it if your messing with fan's. They really were not made to be quiet miner's. There are some good mod's people have done on forum to quiet them some though. Going to S3 level fan... I don't think is one I would do.
If I backed off freq to ~400, and put a S3 type fan on my Batch 6 S7.... I think it would still overheat on a day with ambient > 85*F.... But why do that when you can just turn it down to 30% or less fan speed with the stock fans. I would never consider S3 fans on an S7... The S7 needs the CFM. Do two things: Create a cabinet that is insulated to house the miner and its power supply. Have a ducted inlet to the front, and have the exhaust ducted directly from the fan to an exhausting source. the further away, the quieter it is. you really only need like 6-8Ft if you are ducting directly to a duct vent. |
|
|
|
|
since minercontrol is being used to launch the apps, have you tried running MC under administrator elevation to see if it helps?
It's a strange problem.... I can't think of what would cause HAL to not let your app access the video card... The only time I had a problem with miner apps not accessing the video card myself was when there was a driver updated by windows automatically and it wasn't the nvidia package with CUDA updates.
Joblo is spot on; test all hardware one at a time. I know it seems annoying and it takes a lot of time, but you will find out where your issue resides. If it is on a specific algo; try trimming the intensity back a touch, at least .5 if not 1 from what its at currently. Since Temp is an issue, I would consider this for sure.
I know you definitely need the individual driver packages for each model of card you are running to be installed. In the past I have heard of people running into bugs and issues when they are mixing and matching types/generations of cards, and they were always able to resolve by having each card type on a separate rig. This way drivers are unified and the system OS is simpler.
|
|
|
|
sounds good! I am trying to cut the DB's(NOISE) down so trying different things.
Corsair fan is super quiet. I don't get why they went with these super loud 4400rpm fans when the S1/S3's 3800rpm fans move enough air.
Put 2 s3 fans onto an s7 and watch how quick it overheats.... |
|
|
|
My first thought is its an antivirus program. I know mine wrecked havok on my mining rig until I put exceptions for the directories my miners were in. I had to resort to disabling my antivirus just to get the files downloaded. On the weird lyra2v2 issue, everything caught up eventually. Don't know what caused it, but I got paid so won't complain too loudly lol
I personally don't even have a firewall or anything like antivirus on my standalone miners.... Useless fluff. I'm always behind a router. If they get past that it's my own damned fault.... Lol. But Tbearhere: It sounds like something changed on that system between the point of it working properly and when it started happening. An update; something. So far I have only had those types of issues when the video driver wasn't a true nvidia released driver. |
|
|
|
|
Also hashrate right now is half of what it has been in the past week....
*edit*
And crackfoo, with a hashrate of 24.9Mh solid, and pool report at ~20-22Mh on lyra2, a diff of 10 smoothed out my hashrate. It can be seen via my link above. Idk if this helps you at all if u can trim the stratum auto diff back a bit.
|
|
|
|
|
Try being in California in Silicon Valley.....
My parents pay over 11C/KWH at night for their house.....
And if you have a commercial hookup, be ready to see charges of ~0.32/Kwh..... Cause that's what we pay PG&E for the power service in our aircraft hanger... It's insane.
I am lucky to have a power line to my rental space in the countryside that comes with my rental. I squeeze every last watt I can safely run on my line at all times..... I am working on upgrading to a 220 source and expanding my capabilities at home soon.
The S7 is now makes less than a single GTX980 strictly running Lyra2; That says a LOT about what the halving did to the older equipment. In other terms I can squeeze 3 of those video cards into the same power draw the S7 is currently pulling at a reduced hashrate (960w).... And the only reason its kept online is because I dont pay a dime for the power it draws since its already included with my obsurdly cheap rent.
I am seriously thinking of ditching them all together and not running any new bitmain units until they are fairly priced.... well under the 1K mark.... no way that I would pay more at this point... not a fat chance in hell.
The halving killed the profitability of my S7, and my video card is just a tad less profitable than it was before the halving. That's some serious food for thought to people wanting to buy a new antminer unit.
|
|
|
|
|
stratum seems to put lyra2 diff at double what it should be... I have been able to get my hashrate consistent at the pool by lowering it from 26 to 10.
|
|
|
|
I have a good feeling that there will be a stable python port of the stratum servers available eventually which I'll put putting in BTC to help push the dev along to get it done. It's a tricky project so I hope it makes it to finish.
I myself ponder why altcoin devs on the SHA port don't add a way to have merge mine support with BTC. That way we can mine BTC, and send all lesser hashes to altcoins. Would be a win-win for long-term miner support and for the miners as well. We can only hope for someone to unify the coding so it happens and its easy to add to different coins. |
|
|
|
This is good to hear. We've made changes to the lowest level socket connections which seems to cut out a lot of connection problems. Mainly problems were coming from miners disconnecting from the stratum server when we were in the middle of handing out jobs for one of the coins that were being mined. We now have 5s socket timeouts.
I hoped it would help with all the problems but there there is still problems when I enabled AUX coins so I've dropped them all on sha again and just left DOGE on scrypt. the other algos don't seem to have problems with them.
I can see how the server would get kinda pissed and lock up/reset from being overloaded with open sockets.... I have stability, so I am in for the long haul with that S7. I refuse to let people lowball me on nicehash since they can't change the profitability flag to be something other than what it is which is broken on >S3 models. It was amazing watching people immideately pay almost half for the same sized order; then in the next day watch diff on the altcoins all rise and drop their value per hashrate because they arent pointing it at BTC as much as they were...... The ripple of the halving's second signs... |
|
|
|
In looking at my hash rate graph against the pool's, and my dropouts line up fairly precisely with the pool's. How can that be expected behavior?
Im running a late batch 6 600M model of S7. So, its 4.2Th on a good day. I have it backed off to 30% fan speed and 3.2Th. This is my password string: Henry.S7-1,c=BTC,sha256=3000 My next change will be in another 24 hours where ill bump the diff up by 1K to 4000. My issues is faster coins need a lower diff cause of my internet connection.. I tend to get stale shares and they don't count.... When the share spacing is great, I dont get shares for some time and my delta hashrate at the pool drops. Finding the point where it drops off, and then backing it back off a bit, seems to do the trick.... PRoblem is it needs to stay online long enough to be sure its trimmed properly. What a bugger that is so far.... |
|
|
|
Here's another useful cut and paste. I was able to kick my live mining machine from P2 into P0, no interruptions. I am thinking of building this command into my batch every time it re-launches the miner apps to make sure P-state is zero. Run this in cmd in admin mode:
c:\progra~1\nvidia~1\NVSMI\
nvidia-smi -q -d SUPPORTED_CLOCKS | more
(that's a pipe before more)
Take the top number for memory, something like 3505 (GTX970), and the first number for graphics, something like 1531. Again in admin mode, enter your numbers in this format:
nvidia-smi -ac 3505,1531
You're card's memory will now run in compute mode (P0). nvidia-smi.exe -ac 3505,1443 For my ASUS GTX 980. "nvidia-smi.exe -acp UNRESTRICTED" removes the admin requirements. Ive also noticed X11evo is a power-hungry algo. One machine used to sit at 80-84*C when running only X11EVO. Now Running Lyra2 full-time and it sits @65*C... |
|
|
|
|
I have again started pointed a S7 at zpool. It is consistently reconnecting but, is not getting kicked off needing a hard-reset like it was before.
I do see the results of the disconnects (pool hashrate drops massively is one clue) and my miner hashrate resets then climbs back up.
So it seems as whatever garbage that was keeping the miner from reconnecting or getting hung waiting to reconnect... is no longer happening. My only issues now are the normal power issues associated with living out of the city in the country.
The lyra2 port seems to waver greatly on my hashrate it is seeing. Its very odd.... So sometimes I get a full reward (~25Mh) and sometimes a partial reward (~13Mh)... I am wondering if I force a diff that this will change my hashrate consistency.... telling me that the diff is going just a little too high for the speed of the coins its mining.
|
|
|
|
Best share from a few days solo mining my SP20s was 801,248,735. It was fun trying. They are now profit switching between NiceHash and Kano.is until the rental rates calm down, then they go into retirement (likely sometime later today).
Yep, I pointed the last of my SHA gear at the solo pool for the last ~13Hr just to see... I think only a share of 20M... lol Was worth a chance though =) |
|
|
|
|
ANyone else having issues with the nvidia driver crashing and causing a major hit in hashrates?
I may have a solution I can integrate into my batch if the NVPManager app allows a seamless pre-configured run with no user interactions......
interested on opinions. Its not recommended for machines you use while mining; it causes odd issues with video display apps such as flash, HTML5 containers, and media players....
|
|
|
|
|
Recently I have been using "EVGA PrecisionX 16" (available on steam at steampowered.com) to fix all of my Nvidia clock issues.
My Asus GTX980 only mines in P2 state, but oddly it sits in P0 state when in windows with no apps running. I was never able to get it to be in P0 state while mining.
After enabling kboost, it overrides the P state values for the ones you currently have selected no matter what P state the card switches into.
I have verified this because I get identical hashrates in P8, P5, and P2 power states, which are all higher than the hashrate I get in P2 state with kboost off.
Hope it helps others. Keeps you out of having to screw with the bios manually. Note this works with many makes and models of nvidia cards.
|
|
|
|
|
Show Posts
