GMO B3 33th/s Asic

[minefarmbuy]

https://gmominer.z.com/en/spec/

Thoughts? This thing looks like it will would eventually become a flame thrower in my garage set up. 3.4kw.  .. .


edit for typo

[1714776898]

1714776898

[1714776898]

1714776898

[1714776898]

1714776898

[1714776898]

1714776898

[philipma1957]

https://gmominer.z.com/en/spec/

Thoughts? This thing looks like it will would eventually become a flame thrower in my garage set up. 3.4kw.  .. .

it appears to be able to be down clocked to   81 watts a th.

once again  if i pay 1999 today  I wait until nov.

so July August Sept Oct all lost

for 1999  I can get 3 s9i's in about 10 days.  that is 40.5th  doing 100 watts per th.

same as this unit with super fans.

so I could mine for 110days at the same power rate of 100 watts a th.

[minefarmbuy]

so I could mine for 110days at the same power rate of 100 watts a th.

Agreed, the efficiency is terrible. Also this would practically take up all of my 240/30A line with one miner. If arrived today is just $24 in profits at my power rate and 3k days to ROI at the price mfb would set. Not sure if this worth listing. If I was able to run this and one s9 I'd still be short at least $10/mo in profits compared to whats running.

I suppose the only silver lining is if this arrived just in time for a large bull swing Q4, where efficiencies are still a factor but less of an issue. Not sure if we'll list this one or not, we've been hoping for new tech to be more efficient than this.

PS. I really would be scared this would flame in my garage.

[NotFuzzyWarm]

Does no one have a problem with the miners phoning home? From the sales copy:

Surveillance API
Operation Detection (Anti-theft Measures)

As was the case with Bitmain's s7 remote shutdown (Antbleed?) had it ever been activated it sounds like the miners report home either at boot or periodically when running. The anti-theft measures means GMO can brick the miners at will. Thing is, I'd assume the miners need to phone home and get a permission key before they will run.

I for one will never allow a miner or any other device to do that. Even Bitmain's s7 allowed phone-home (had it ever been activated) to be disabled.

[KaydenC]

GMO's btc miners are disappointing. The 1950w box barely has +20% efficiency to the S9, and this has even worse efficiency than the S9.
The fans must eat 200w, trying to cool 3400w in that box.

[minefarmbuy]

GMO's btc miners are disappointing. The 1950w box barely has +20% efficiency to the S9, and this has even worse efficiency than the S9.
The fans must eat 200w, trying to cool 3400w in that box.

Overclocked b2 with higher CFM fans?

[minefarmbuy]

Does no one have a problem with the miners phoning home?

[...]

Concerning, very. Would this allow them control of device as well? 10min "downtime" to mine for themselves?

[NotFuzzyWarm]

Concerning, very. Would this allow them control of device as well? 10min "downtime" to mine for themselves?

Until someone has one of these miners and goes through the code, unknown. That possible re-direction is what folks were worried about with the s7's.

[sidehack]

Yeah no, that's a pretty extreme fire hazard. If it requires 4 fans to keep cool, it should be in a box twice the size. Ask any mining datacenter in the world, they'll all tell you "keeping a miner from bursting into flames" matters quite a bit more than "being tiny".

Why are people so addicted to the S7 size/cooling concept that they'd try and shoehorn 2-3.5KW into what's already not a great plan for 1.5KW?

[philipma1957]

Yeah no, that's a pretty extreme fire hazard. If it requires 4 fans to keep cool, it should be in a box twice the size. Ask any mining datacenter in the world, they'll all tell you "keeping a miner from bursting into flames" matters quite a bit more than "being tiny".

Why are people so addicted to the S7 size/cooling concept that they'd try and shoehorn 2-3.5KW into what's already not a great plan for 1.5KW?

yeah  if I were in charge of this gear  it would have  under 1000 watts

say 800 watts and 10,000 gh   and slower fans that pull less power.

I would sell with home miners as my priority  and market it as such.

"why run this  when you can be paid for running  this"

[minefarmbuy]

Yeah no, that's a pretty extreme fire hazard. If it requires 4 fans to keep cool, it should be in a box twice the size. Ask any mining datacenter in the world, they'll all tell you "keeping a miner from bursting into flames" matters quite a bit more than "being tiny".

Why are people so addicted to the S7 size/cooling concept that they'd try and shoehorn 2-3.5KW into what's already not a great plan for 1.5KW?

4 fans? Not just wider 120mm? At least the s9 is longer than the s7. still though I can't think of augment to support GMO thoughts on this.


@philipma I had been bringing an s7 and down clocking to heat the house and I would love to see more efforts in regard to efficiencies. 

[sidehack]

It looked like two fans back to back on each end, but it could be just one extra-super-duper fan. For the kind of cooling that would be infinitely easier to achieve from, say, an S4-sized housing. And also quite a bit less susceptible to dust. A lot of people don't consider just how much fan power draw, and chip-level temperatures, can hurt machine-level efficiency.

A longer housing for the same power would help with per-chip power dissipation, which in general means lower chip temperatures so easier cooling and longer expected lifetime. However, don't forget that the last chips in the line are being cooled by air that's already gone over everything ahead. This is a lot of why Spondoolies' 3.5KW 2U rack miners were impossible to run at rated speeds, and let's not forget the AM Prisma debacle. All things being equal, a short wide miner will always cool more effectively than a long narrow one.

MFB, if you're using a 135-chip S7, you can also undervolt for improved efficiency.

[minefarmbuy]

It looked like two fans back to back on each end, but it could be just one extra-super-duper fan. For the kind of cooling that would be infinitely easier to achieve from, say, an S4-sized housing. And also quite a bit less susceptible to dust. A lot of people don't consider just how much fan power draw, and chip-level temperatures, can hurt machine-level efficiency.

A longer housing for the same power would help with per-chip power dissipation, which in general means lower chip temperatures so easier cooling and longer expected lifetime. However, don't forget that the last chips in the line are being cooled by air that's already gone over everything ahead. This is a lot of why Spondoolies' 3.5KW 2U rack miners were impossible to run at rated speeds, and let's not forget the AM Prisma debacle. All things being equal, a short wide miner will always cool more effectively than a long narrow one.

MFB, if you're using a 135-chip S7, you can also undervolt for improved efficiency.

You're correct got a closer look, two fans attached. Seem incredibly counter intuitive to blow into another fan to cool.

I imagine manufacturers don't want to disrupt spacing for hardware on mid to large scale operations. Like GMO who sells hosting or hash power (?) makes it easier for them specifically to insert new machine in place of old. Where I, at home wouldn't matter as my set up is much more malleable. I suppose a fan hanging off a shelf is much better than part of the housing. Just feels off as a whole and hard to take GMO seriously without efficiencies and design function. My first view was a literal wft moment.   

I got word we're going to list, as B2 did well for us but efficiencies were better. Nov is forever time like philipma said so we'll see.

Sidehack, I am brushing up on a couple of your threads on the s7. I plan solo mine unless market rises, or I have a friend who may host at their work for free power.

[Raymond_B]

I find this hashrate and electrical consumption asterisk curious.

* It could operate as the same conditions as GMO miner B2.

And then on the B3 page there's this.

Operating Environment   
Temperature: 0°C to 30°C
Humidity: 30% to 80% RH, prevent condensation
Altitude: Below 2,000m

[minefarmbuy]

I find this hashrate and electrical consumption asterisk curious.

* It could operate as the same conditions as GMO miner B2.

And then on the B3 page there's this.

Operating Environment   
Temperature: 0°C to 30°C
Humidity: 30% to 80% RH, prevent condensation
Altitude: Below 2,000m

I wonder if the phrasing is a translation error? "Could operate as the same conditions" referring to environment? Or maybe that it can be down clocked to hit b2 hash/efficiencies? Still not instilling confidence in this unit.

We've tried to get testing units form them for the b2 and now their b3, zero service outside automated reply's. I guess everyone waits till November. . ..

[HagssFIN]

3.4 kW is way too much in such small steel tube case.

[r01k]

As was the case with the B2, Antminers are a better deal now.

GMO B3:

• 103W per TH
• 0.0165 TH per Dollar

Antminer S9i 13.5TH w/ PSU:

• 97W per TH
• 0.0193 TH per Dollar

For now it's wise to keep your money and see what Bitmain next SHA256 is going to bring eventually.

[Thetaj]

Yeah no, that's a pretty extreme fire hazard. If it requires 4 fans to keep cool, it should be in a box twice the size. Ask any mining datacenter in the world, they'll all tell you "keeping a miner from bursting into flames" matters quite a bit more than "being tiny".

Why are people so addicted to the S7 size/cooling concept that they'd try and shoehorn 2-3.5KW into what's already not a great plan for 1.5KW?

You pretty much hit the nail in the head there. Especially in Asia, having this miner is probably not the brightest idea, especially if you don't wanna see your 2m$ mine go up in flames because one of the (4! wtf?) fans died.

[Razon]

Heya,

Same conclusion I came to when I saw the first announcement of GMO;s 7nm Miners. Whoever designed this board, really did a piss poor job in regulating power consumption.

But, in all honesty it could be coming down to the limitation of the first generation of 7nm chips just not being up to spec. (Yes 7nm is still in first generation stages, these could also be just out of proof of concept stage)

Bitmain will most likely make a better machine since they wont be using Samsung, and will be sticking with TSMC who generally makes much better semiconductor technology. You wont see better systems until  last quarter of 2018 to first quarter of 2019 as most 7nm fabs are only being completed around this time right now.

[philipma1957]

Heya,

Same conclusion I came to when I saw the first announcement of GMO;s 7nm Miners. Whoever designed this board, really did a piss poor job in regulating power consumption.

But, in all honesty it could be coming down to the limitation of the first generation of 7nm chips just not being up to spec. (Yes 7nm is still in first generation stages, these could also be just out of proof of concept stage)

Bitmain will most likely make a better machine since they wont be using Samsung, and will be sticking with TSMC who generally makes much better semiconductor technology. You wont see better systems until  last quarter of 2018 to first quarter of 2019 as most 7nm fabs are only being completed around this time right now.

who ever designed these did a huge fail.

but the boss of the company most likely does not give a shit and has his own agenda .