BW 14nm Miners Update

[jackg]

There appears to be a new competition for the smallest ASIC chip that can be produced.

The BitFury chips have hardly been on the market and there is already BW that is overtaking the powerfullness of the BitFury Chips.

The new development is quite interesting from BW as they are a similar company to BitFury and are one of its biggest rivals.

[IanQuinn]

There appears to be a new competition for the smallest ASIC chip that can be produced.

The BitFury chips have hardly been on the market and there is already BW that is overtaking the powerfullness of the BitFury Chips.

The new development is quite interesting from BW as they are a similar company to BitFury and are one of its biggest rivals.

BFD...  They are not selling them.

[notlist3d]

There appears to be a new competition for the smallest ASIC chip that can be produced.

The BitFury chips have hardly been on the market and there is already BW that is overtaking the powerfullness of the BitFury Chips.

The new development is quite interesting from BW as they are a similar company to BitFury and are one of its biggest rivals.

I don't think BW hit specs though from what we saw on the few that made it to resellers it was close to A6/S7 on efficiency.   But it was a early batch of chips... what they can do now hard to say.

We just know they have not shown one that really breaks current gen efficiency to general public at least.  I don't know what they have internal ... so hard to say.

[QuintLeo]

The B-Eleven may or may not have met it's early specs - hard to tell when there have been almost none seen in public if any at all, and the only info we have about the actual final hardware is a few pictures.

 They never announced ground-breaking specs for it though, even the EARLY specs were very close to the specs on the S7/BM1385 - and there is no indication that they did NOT meet or achieve very close to those specs (other than the 3/5/10TH miner part which DOES appear to have changed) on the actual miner.


 The LK1402 is what's going to make things interesting though, if they can achieve THOSE specs or very very close in the timeframe they're talking about - *IF* they actually start selling the things to someone other than "the big farms" and using them internally.

[RichBC]

I think it's important to give BW credit for what they have done. The B11 is the first real production 14nm miner it may not be ground breaking but as  QuintLeo says is close to the advertised spec, has been deployed in significant volumes for the Cloud hash customers which paid good interest during the presale stage and is now paying good cloud mining returns.

The jury is still out on it's successor based on the improved BW-LK1402, this being a string based design with an expected efficiency of 0.149GH at 195W. Now due in June with sales to individuals.

Rich

[philipma1957]

I think it's important to give BW credit for what they have done. The B11 is the first real production 14nm miner it may not be ground breaking but as  QuintLeo says is close to the advertised spec, has been deployed in significant volumes for the Cloud hash customers which paid good interest during the presale stage and is now paying good cloud mining returns.

The jury is still out on it's successor based on the improved BW-LK1402, this being a string based design with an expected efficiency of 0.149GH at 195W. Now due in June with sales to individuals.

Rich

Would be nice to see a new efficient miner say .14 or so.

Frankly I wonder if these will  be like the s5 compared to the s3

A drop from. 0.8 to 0.52

Rather the the s-5 to the s-7. Which was .52 to .25

I have a batch 2 s-7 doing  .25 watts a gh on an underclock

[RedRockMining]

Ok, so what's the bottom line on efficiency speculation at this point for BW.com's new miners vs. Bitmain S7?

The Bitmain S7 is 0.249w/GHS (per Hashnest's website) and is supposed to be the most efficient miner currently available to the public.

Bw's "Proclamation" on 2/15 says "BW-LK1401 chip is 160W/T, which makes it the lowest in the industry. Building on this technology the second generation, BW-LK1402 chips are currently in production with a lower power consumption of 95 W/T.BW"

Which translates into .160w/GHS and .095w/GHS which are HUGE jumps in efficiency.

I presume the former is the one currently promised for June.

Do we actually believe that these specs will be delivered. Has anyone actually gotten their hands on one or seen it running?

[notlist3d]

Ok, so what's the bottom line on efficiency speculation at this point for BW.com's new miners vs. Bitmain S7?

The Bitmain S7 is 0.249w/GHS (per Hashnest's website) and is supposed to be the most efficient miner currently available to the public.

Bw's "Proclamation" on 2/15 says "BW-LK1401 chip is 160W/T, which makes it the lowest in the industry. Building on this technology the second generation, BW-LK1402 chips are currently in production with a lower power consumption of 95 W/T.BW"

Which translates into .160w/GHS and .095w/GHS which are HUGE jumps in efficiency.

I presume the former is the one currently promised for June.

Do we actually believe that these specs will be delivered. Has anyone actually gotten their hands on one or seen it running?

Your missing part where S7 is one of most consumer efficient miners for THIS generation.   Next generation that goes out the  window S7 does not stay at top spot forever.... likely bitmain has a S9 or something in works.  They are just good at keeping it secret.

On specs above it was mentioned no one has seen a BW chip that met later on specs.  But they were EARLY in generation  and only ones we saw were a few that somehow made it to re-sellers.  It was not really meant to be sold to general public.  But no doubt they will refine it... what it ends up being no one can really say  for sure.

[QuintLeo]

You're also comparing CHIP efficiency to MINER efficiency.
The actual MINER is specified at .26W/GH

 https://bitcointalk.org/index.php?topic=1340363.0

The BM1385 in the S7 is CAPABLE of getting down to something like .15w/GH on a CHIP level - but an actual miner will never see that due to the power needs for the bucks/power distribution, the controller electronics, the mining comp to RUN it, the FANS, etc....

The efficiency of the S7 also varies with the batch - the more recent lower chip count batches are LESS efficient than some of the early batches, as they're running fewer chips per string = higher operating voltage = lower efficiency (but higher hashrate capability abet at higher power use and heat level).

[RedRockMining]

You're also comparing CHIP efficiency to MINER efficiency.
The actual MINER is specified at .26W/GH

 https://bitcointalk.org/index.php?topic=1340363.0

The BM1385 in the S7 is CAPABLE of getting down to something like .15w/GH on a CHIP level - but an actual miner will never see that due to the power needs for the bucks/power distribution, the controller electronics, the mining comp to RUN it, the FANS, etc....

The efficiency of the S7 also varies with the batch - the more recent lower chip count batches are LESS efficient than some of the early batches, as they're running fewer chips per string = higher operating voltage = lower efficiency (but higher hashrate capability abet at higher power use and heat level).

Ah, thanks. That clarifies things quite a lot for me. Could explain why Bitmain hasn't responded with their next model yet...

So my next question then is assuming BW's process is a good representation of the state of the art in processor fabrication, is die shrink now a dead-end for increasing efficiency?

[QuintLeo]

BW's current chip is not - their announced LK-1402 chip should be a lot closer, as it's specced in the same ballpark as the "on video" Bitfury chip.

 I suspect there can be one more round of optimisation past those (and any similar chips showing up from BitMain and Avalon and Innosilicon etc) as folks get better aquainted with the 14/16nm FinFet type node and process, but I doubt it would be a massive huge improvement.

 At this point, Bitcoin mining chips are about to reach the current semiconductor "state of the art" past which there won't be any major improvements for a few years (when the NEXT node reaches production stage).



 Might not happen then, given the "almost no improvement" that happened at the 20/22nm node vs 28nm.

[Efficiency V]

The efficiency of the S7 also varies with the batch - the more recent lower chip count batches are LESS efficient than some of the early batches, as they're running fewer chips per string = higher operating voltage = lower efficiency (but higher hashrate capability abet at higher power use and heat level).

What S7 batch is currently the most efficient?  Also what are some of the other batches that are at least good runner ups in efficiency / most profitable to be running?  I want to know what used ones I should look for if I am going to buy some.

[notlist3d]

The efficiency of the S7 also varies with the batch - the more recent lower chip count batches are LESS efficient than some of the early batches, as they're running fewer chips per string = higher operating voltage = lower efficiency (but higher hashrate capability abet at higher power use and heat level).

What S7 batch is currently the most efficient?  Also what are some of the other batches that are at least good runner ups in efficiency / most profitable to be running?  I want to know what used ones I should look for if I am going to buy some.

Look at the s7 page it has a good graph at bottom: https://bitmaintech.com/productDetail.htm?pid=00020160510034707221GxQ3Yr280661

At batch 8 they perfected the lower chip and higher freq design.  This makes it lower cost to produce and should mean lower cost to us (in theory).  But it lost efficiency going with less chips and higher freq. 

[philipma1957]

The efficiency of the S7 also varies with the batch - the more recent lower chip count batches are LESS efficient than some of the early batches, as they're running fewer chips per string = higher operating voltage = lower efficiency (but higher hashrate capability abet at higher power use and heat level).

What S7 batch is currently the most efficient?  Also what are some of the other batches that are at least good runner ups in efficiency / most profitable to be running?  I want to know what used ones I should look for if I am going to buy some.

my batch 2 is really good it can do .245 watts per gh with a quality platinum psu

it can run with .0001 error rate for weeks on end.

newer s-7's are about .291 watts a gh

[JaredKaragen]

my batch 2 is really good it can do .245 watts per gh with a quality platinum psu

it can run with .0001 error rate for weeks on end.

newer s-7's are about .291 watts a gh

That sounds about right....  I have one of the later batch 600M with fewer chips;  It stays near ~965w @ 500Mhz doing ~3.4Th which comes to 0.283w/Gh
This is with a Coolmax 1600 Silver rated PSU.

[Sierra8561]

Looks like we won't be seeing this miner in june. I was really looking forward to it.

[KeezAZA]

https://www.bw.com/pool/miner
So pretty! HOW MUCH?
WHERE CAN ONE GET THEM?

[HagssFIN]

BW.com B11-PLUS
https://www.bw.com/pool/miner


I'm surprised that with the 14nm process it is so less efficient than the S9.
It is more like a competitor for the S7 and the Avalon6.

Of course we must keep in mind that this is not the newest miner from BW.com.
The latest one that I've heard of is BW.com B16, but I haven't seen any pics or detailed specs about it.
All I know that the advertised efficiency for the B16 is 0.149J/GHs.
https://www.bw.com/bsixteen

[Biodom]

BW.com B11-PLUS
https://www.bw.com/pool/miner


I'm surprised that with the 14nm process it is so less efficient than the S9.
It is more like a competitor for the S7 and the Avalon6.

Of course we must keep in mind that this is not the newest miner from BW.com.
The latest one that I've heard of is BW.com B16, but I haven't seen any pics or detailed specs about it.
All I know that the advertised efficiency for the B16 is 0.149J/GHs.
https://www.bw.com/bsixteen

could be competitive if they price it below $440. Do they sell internationally?
If so, instructions have to be in English.

[HagssFIN]

If it would be possible to get efficient results with undervolting and underclocking, it would be tempting  

In March 22 I asked BW.com if they are going for a public release with their miner and this the answer I got, I posted it in this thread: https://bitcointalk.org/index.php?topic=1206056.msg14280444#msg14280444

I was curious about the new miner, so I asked BW about it via e-mail.
And I even got a response pretty quickly.  

  (click on thumbnail)

Response:
Yes of course. the general release of the new miner is sure to be, but the timing is not yet known exactly.
Our staffer, who deals with these issues now is not in the office.
as will appear more detailed information about it we will tell you, the main thing that you reminded about it in advance

Original message:
Hi! Is BW going to make a general release of the newest (14nm?) miner like
Bitmain has made with Antminer S7 and Canaan Creative has done with Avalon6?
Seems like it would perform more efficiently than these miners above, which
would be great news for miners.