[ANN]Bminer: a fast Equihash/Ethash/Cuckaroo29z miner for AMD/NVIDIA GPUs 16.4.9

[realbminer]

The version 3.0.0 has been released:

• Minor performance improvement.
• Disable CPU polling by default.
• Introduce a Web-based dashboard.
• Support slushpool.

CPU polling has been disabled by default. It might lead to lower hashrate in some configuration. You can turn it back on by appending -polling=true in the command line.

[1714848152]

1714848152

[1714848152]

1714848152

[1714848152]

1714848152

[Jsorrentino]

The version 3.0.0 has been released:

• Minor performance improvement.
• Disable CPU polling by default.
• Introduce a Web-based dashboard.
• Support slushpool.

CPU polling has been disabled by default. It might lead to lower hashrate in some configuration. You can turn it back on by appending -polling=true in the command line.

was the watchdog issue fixed?

[Vann]

Awesome update. On 3.0 I'm getting 1080-1090 Sol/s with a GTX 1080 and 1070 @ ~300 W! The high CPU usage is now gone with ~30% usage on a G1840. Thank you!

[TONY_GUN]

First at all - thanks for excellent miner! I have best results with it.
But I have some problem. I have 2 rigs on the old 775 socket with identical ASUS P5K Premium MoBos.
First rig has 4 gtx 1060 and core2quad q8400 cpu. Bminer works like a charm on this rig. (2,5 - 3 % cpu load and about 1300 sol/s).
But the second rig with 2 gtx 1060 and core2duo E6750 cpu fails to run bminer. At the same time DSTM's miner works good with this E6750 cpu.
I've made a commandline screenshot of the error.
http://images2.imagebam.com/e3/96/06/9368a5692349993.png

sry for my bad eng

[sic57005]

Recently tested Bminer: agree, the performance is just amazing. More than 15% faster than EWBF on the same cards. Up to stable 320 Sol/s for 1060 3GB and 600 Sol/s for G1 1080.

But before mass moving to Bminer I have two questions:
1. What the "firmware.bin" file is for? Does Bminer reflash your video cards? Can this affect guarantee?
2. Why it keeps connections not only to the specified pool but two more hosts nowhere mentioned?

[metman]

Recently tested Bminer: agree, the performance is just amazing. More than 15% faster than EWBF on the same cards. Up to stable 320 Sol/s for 1060 3GB and 600 Sol/s for G1 1080.

But before mass moving to Bminer I have two questions:
1. What the "firmware.bin" file is for? Does Bminer reflash your video cards? Can this affect guarantee?
2. Why it keeps connections not only to the specified pool but two more hosts nowhere mentioned?

Good questions. Would also like to know an answer to those before mass moving to bminer.

[NameTaken]

Other connections is probably devfee. Try running with -nofee.

[realbminer]

First at all - thanks for excellent miner! I have best results with it.
But I have some problem. I have 2 rigs on the old 775 socket with identical ASUS P5K Premium MoBos.
First rig has 4 gtx 1060 and core2quad q8400 cpu. Bminer works like a charm on this rig. (2,5 - 3 % cpu load and about 1300 sol/s).
But the second rig with 2 gtx 1060 and core2duo E6750 cpu fails to run bminer. At the same time DSTM's miner works good with this E6750 cpu.
I've made a commandline screenshot of the error.


sry for my bad eng

I didn't have a Core 2 duo handy but will address the issue in the next release.

[Gaglam]

For me it's only 100 Sol/s more than EWBF

[un4given]

For me it's only 100 Sol/s more than EWBF

"ONLY"  for what GPU - 1080ti?

[moozooh]

Can confirm that Slushpool works now. On my GTX 1070 with 8960 MHz mem clock, 105% power limit, and CPU polling disabled in the miner, I'm getting about 508.5 sol/s (3.22 sol/J), or ~6 sol/s more than on the DSTM miner on average. (Note that this setup is optimized for maximum profit rather than maximum power efficiency because power is very cheap here.) I also enjoy the fact that it keeps desktop applications sufficiently responsive by default without dropping too much performance (I see a ~4% drop when e.g. watching a YouTube video, which is perfectly acceptable) and without having to manually tune intensity settings. Good job there!

Although by a small margin, this is, indeed, the fastest miner around. At least until DSTM turns off the dev fee. :p

Could any 1080 Ti owners test this miner on their highest stable clocks (without voltmod) and report the sol/s values they're getting? I'm interested in investment opportunities. If my estimates are correct, it should be possible to push some 810..825 sol/s with this miner.

Awesome update. On 3.0 I'm getting 1080-1090 Sol/s with a GTX 1080 and 1070 @ ~300 W! The high CPU usage is now gone with ~30% usage on a G1840.

Is this with CPU polling off? What are your 1070 clocks? I don't think I've ever managed to go above 500 at 98% power limit (which I assume is your setting).

[stash2coin]

On my linux rig with 12 cards i'm getting a lot of "Duplicate share" messages, on windows i dont see this message. Is this normal? Question for the dev.

[NameTaken]

Can confirm that Slushpool works now. On my GTX 1070 with 8960 MHz mem clock, 105% power limit, and CPU polling disabled in the miner, I'm getting about 508.5 sol/s (3.22 sol/J), or ~6 sol/s more than on the DSTM miner on average. (Note that this setup is optimized for maximum profit rather than maximum power efficiency because power is very cheap here.) I also enjoy the fact that it keeps desktop applications sufficiently responsive by default without dropping too much performance (I see a ~4% drop when e.g. watching a YouTube video, which is perfectly acceptable) and without having to manually tune intensity settings. Good job there!

Although by a small margin, this is, indeed, the fastest miner around. At least until DSTM turns off the dev fee. :p

Could any 1080 Ti owners test this miner on their highest stable clocks (without voltmod) and report the sol/s values they're getting? I'm interested in investment opportunities. If my estimates are correct, it should be possible to push some 810..825 sol/s with this miner.

Awesome update. On 3.0 I'm getting 1080-1090 Sol/s with a GTX 1080 and 1070 @ ~300 W! The high CPU usage is now gone with ~30% usage on a G1840.

Is this with CPU polling off? What are your 1070 clocks? I don't think I've ever managed to go above 500 at 98% power limit (which I assume is your setting).

[marco.ruiz]

On my linux rig with 12 cards i'm getting a lot of "Duplicate share" messages, on windows i dont see this message. Is this normal? Question for the dev.

Yeah, me too. At least 30% of my shares are duplicates, that’s why i haven’t switched from dstm. I also run 12 GTX 1070 but on Windows, haven’t tried it in Linux.

[stash2coin]

On my linux rig with 12 cards i'm getting a lot of "Duplicate share" messages, on windows i dont see this message. Is this normal? Question for the dev.

Yeah, me too. At least 30% of my shares are duplicates, that’s why i haven’t switched from dstm. I also run 12 GTX 1070 but on Windows, haven’t tried it in Linux.

For 30min mining almost half of the shares are rejected, this is no go for 12 cards rigs back to the old miner. Hope the dev will address this issue.

[Gaglam]

For me it's only 100 Sol/s more than EWBF

"ONLY"  for what GPU - 1080ti?

"only" for all 5x 1070 together

with the 2% fee, EWBF is still more profitable for me

[moozooh]

"only" for all 5x 1070 together

with the 2% fee, EWBF is still more profitable for me

You might need to check your math. In order for a 2% fee to be a net loss compared to 100 less sol/s your setup must output over 4900 sol/s, which is outright impossible even if you have five 1080 Tis.

More realistically, I'd expect each of your five 1070s to work at around 480 sol/s. That's 5×480=2400 sol/s, or 2400+100=2500 sol/s with bminer. Deducting 2% dev fee, that's 2500×0.98=2450, which is 50 sol/s more than EWBF.

[Vann]

Can confirm that Slushpool works now. On my GTX 1070 with 8960 MHz mem clock, 105% power limit, and CPU polling disabled in the miner, I'm getting about 508.5 sol/s (3.22 sol/J), or ~6 sol/s more than on the DSTM miner on average. (Note that this setup is optimized for maximum profit rather than maximum power efficiency because power is very cheap here.) I also enjoy the fact that it keeps desktop applications sufficiently responsive by default without dropping too much performance (I see a ~4% drop when e.g. watching a YouTube video, which is perfectly acceptable) and without having to manually tune intensity settings. Good job there!

Although by a small margin, this is, indeed, the fastest miner around. At least until DSTM turns off the dev fee. :p

Could any 1080 Ti owners test this miner on their highest stable clocks (without voltmod) and report the sol/s values they're getting? I'm interested in investment opportunities. If my estimates are correct, it should be possible to push some 810..825 sol/s with this miner.

Awesome update. On 3.0 I'm getting 1080-1090 Sol/s with a GTX 1080 and 1070 @ ~300 W! The high CPU usage is now gone with ~30% usage on a G1840.

Is this with CPU polling off? What are your 1070 clocks? I don't think I've ever managed to go above 500 at 98% power limit (which I assume is your setting).

Zotac Amp! Extreme GTX 1070. Core +100 MHz, Memory, +500 MHz, Power Limit 65%, CPU polling off. Running stable almost 2 days. With EWBF 0.34b I get ~460 H/s and with DSTM v0.5.7 ~495 H/s with the same settings.

[moozooh]

Ah, that explains it: your power limit isn't actually in effect, which is demonstrated by the card consuming up to 161W, or 107% (not 65%), performing at 3.15 sol/J. Those are expected values for a well-made 1070. I thought the 503 sol/s at 147W was a sustained figure, which it doesn't appear to be. Got my hopes up for a second there. :D

[Vann]

Ah, that explains it: your power limit isn't actually in effect, which is demonstrated by the card consuming up to 161W, or 107% (not 65%), performing at 3.15 sol/J. Those are expected values for a well-made 1070. I thought the 503 sol/s at 147W was a sustained figure, which it doesn't appear to be. Got my hopes up for a second there.

The 65% power limit is in effect. The max power is 250 W, 250 x 0.65 = 162.5 W. As long as the extra hash is profitable the extra wattage is irrelevant.