Topic: [ANN]Bminer: a fast Equihash/Ethash/Cuckaroo29z miner for AMD/NVIDIA GPUs 16.4.9 - page 161. (Read 148347 times)

Would really love to hear something from the author on topic of Slushpool problems and the high CPU usage. Their explanation regarding performance smells like bullshit, considering the fact that for people who intend to use this piece of software profitability trumps performance any day. Bminer doesn't even break even with dstm's when it comes to full system sol/J ratio.

So I decided to switch over to dstm 0.5.6 starting at 6:25pm yesterday to see how it compares to bminer. After 12+ hours of operation here are my observations:

bminer is about 3% faster, based on reported hashrate.

However, dstm results in much less power draw based on the watt display on my UPS: 220W for bminer vs. 190W for dstm. That 30W difference will prove critical.

While dstm does tax the CPU much less - task manager reports 0.8-1.0% - the 15% usage by bminer didn't really affect the computer's operation so not really a penalty.

As far as the UI goes, I like that dstm reports average Sols/s and GPU temperature whereas I like that bminer reports # of rejected vs. accepted shares as well as just being more verbose in general about what it is doing (e.g. - "Share submitted", "Share accepted", etc...). Another possible advantage for dstm is that it reports Sols/W as well, but with my card it only says "infinite" so that function doesn't seem to be working, at least with my card/setup.

Both have been equally stable with no crashes or weird network errors (connected to the us-east server on Nanopool).

When I plug the relevant numbers into this calculator - https://minethecoin.com/coin/zcash - to compare profitability I find that dstm edges out bminer by 7.5% on a monthly basis ($17.00 profit vs. $15.82 per month). So dstm is the clear winner here, despite that bminer mines 3% faster.

How much sol/s?

Quick update - bminer has been running for ~30 hours averaging about 193 Sols/s mining Zcash to Nanopool with my GTX 1050 Ti. I've racked up 6 rejected shares out of 766 and counting; not sure if that has more to do with ping latency or the miner itself, but thought I'd relay the datum anyway.

I plan on running this until I've collected the minimum 0.01 ZEC payout from Nanopool (ie, another day or so... lol) then I'll be switching back to either Vertcoin (trying One Click Miner) or Ubiq (using ethminer 0.12.0, unless I find something interesting to try).

Does this support suprnova?

That’s ok, you have to know your hardware just don’t make the same mistake i did. Don’t waste a lot of time benchmarking with just one card, it’s a completely different game multi GPU, especially in a 10+ GPU rig.

FWIW, I just started mining a week ago and am still in the "trying different things out before I commit to buying any specialized hardware" phase, so I am not even remotely an expert. I also just started running bminer earlier today - currently at about 7 hours of mining Zcash on Nanopool...

That said, I haven't had any crashes, CPU usage (AMD FX-8300) is around 15% (according to Windows 10 task manager), and I am getting a software-reported hashrate of 190-193 Sols/s with a rinky-dink Zotac GTX 1050 Ti Mini that has been mildly overclocked (see details if interested below) with MSI Afterburner. I will be trying out a couple of other different miners but I want to let this one run for a full 24 hours first. I did run Excavator for about 15 minutes but it was reporting a terrible hashrate of around 100 Sols/s so I didn't investigate it further. I also have EWBF 0.3.4b and Claymore 10.2 to play with, but I likely won't bother with Claymore since the way it does the dev fee mining actually costs you over 1 minute per hour vs. 36 seconds if you aren't mining Ethereum because of the time it takes to change pools, generate the DAG for Ether, etc. Not to go off on a tangent in my own post, but my personal opinion is that dev fees should be paid  in whatever coin is being mined and by switching the payment address.

EDIT: one feature request so far - display the cumulative run time.

MSI Afterburner overclocking details: +86 on the core clock to 1795 (max possible is 1797); +196 on the mem clock to 3700. Lowering TDP resulted in a lower hashrate so no point in doing that while excessively increasing mem clock did nothing but make the card run hotter so it seems that both core and mem clock need to go up together for the most benefit.

Rock solid with EWBF, dstm can run for hours but sometimes just 10 minutes then restarts. That’s why i wanted to try bminer but looks i’m stuck with EWBF at 5K Sols/s.

You are able to run 12 Nvidia cards on windows? Your system is stable?

just a hint for users having problem starting-you're not using notepad probably,use notepad and replace lf with cr/lf manually,or convert in npp edit/eol conversion
or use your cmd/shortcut
edit:this is for case where the miner does not output to console at all

and question/warning,there seems to be some autoupdate,or at least version checking,that might be dangerous,seems to be for firmware.bin,check is output to console only if the file is missing,so i'm guessing some low level instructions for quicker update without miner update maybe?might be a problem for advanced firewall users/farms with proxy

aaaand edit2:,on 1080ti a wee bit oc(gigabyte stock water),same settings,dstm 430 i/s,bminer 440 nonce/s (i think this has the same meaning)

For me the miner starts and appears to be working but GPU's don't kick in... they stay cool as hell, fans even off and the miner software reports GPU0: 0.00 Sols/s  GPU1: 0.00 Sols/s..... It can stay this way for as long as i leave it like that. The miner keeps showing "Got new job..." but the pool never gets a share nor an active worker.

Rig:
12x Asus 1070 ROG Strix OC
Asrock H110 Pro BTC
Intel G4400
8GB RAM
120 SSD
2400W server PSU

OC:
Power Limit: 65
GPU Core: 95
Memory Core: 450
Target Temp: 70 (but the normally run at 56 - 62)
These settings works like a charm on dstm getting aprox. 4.24 Sols/W at a total of 5.1 KSols/s.

Fastest miner for me, getting 2241 s/s with 3 x 3GB 1060s and 3 x 1070s and runs stable for 24 hours+

CPU usage is at 89% for a FX 8320e though which adds another 30w to the system power... Hopefully an update will fix this

Can you fix it for slushpool? I can not get it to work there.

 Connection failed due to End of file, retrying in 5 seconds

i overlooked the cpu usage it is high and yes it is not stable soooo  I'll pass..

seems not virus... but it have very high CPU usage which make it seems abnormal....

how can it use over 40% on a 4790K CPU for just 3 X 1080TI??

the power consumption on the CPU cannot compensate the gain ..... and it is not stable too....

A newby and a new miner ...... mmm...

I won't be using this for now, to much risk for a few more hash....

Getting this error when I run it:

panic: interface conversion: interface {} is []interface {}, not string
[WARN] [2017-12-07T10:31:27-05:00] Miner died! It will be restarted soon...

me too... it is crashed on 3 X 1080ti Z170 system within 5 minutes....
and it used over 40% cpu power on my 4790K CPU with 32GB RAM....
While ZWBF used 2-3% only....

I do worry it have some hidden CPU mining program.... and I giveup and use back ZWBF already...

at home I have better performance than ewbf on my 1080.

but when I use it on my rig of 7x1070, the miner crashes after a few minutes and restarts, did you have this problem?

it seems faster than ewbf but see the stability.

it seems a bit faster and a bit cooler

typically cooler cards means more efficient, less power consumed for a bit more hash..

now observing for stability.