Topic: [Archive] BFL trolling museum - page 7. (Read 69394 times)

Josh stated somewhere on the BFL forums that he wasn't allowed to make any pictures in the facility.
He only made some pictures of the outside. But he also can't post those cause there are details about the faccility on the pictures he doesn't want to leak out to prevent people harrassing the facility.
He could blur out the details he said. But then you just only have pictures of a big building.

Boxes... Cool...
Why the hell cant you just be at the chip fabrication, and snap a quick shot of the In Construction board?
It would shut everyone up. "chip fab plant isnt nearby" Well yeah, But you DID visit it once before, And I still havent seen photos

Oh no! You'll encourage that guy who keeps talking about 12th century Catholics

The thread title flat out says that the boxes are empty, and quite obviously exists only for people who like pictures of boxes (generally people who buy Apple products), and not as "evidence" for ASICs.

The quality of the trolling in this thread is just getting lower and lower :/

I suppose I can agree there, though I recall seeing a post about "0.8 to 1.2".

+1 LOL

This and many many many more post dont account for the price of BTC in the future...  If BTC worth 1000$ or 1$ in four years... All those nice post are worthless.. Even the seller of the mining equipement is'nt in the race at 1000$/BTC...

All useless post IMO.. All depends of what value BTC will exchange for when it will !

Ok, fair point.  I would argue that those 1U cases and the components within are specifically designed to withdraw that heat in a particular way, with hundreds or thousands of engineering hours dedicated to exactly that, but it does show it is possible.

water cooling maybe?

Probably. The avalon system apparently is going to run far cooler then expected, likely due to large fans and utilizing a form factor almost as big as the BFL shipping box. However, it would seem from posts that some of the space is occupied by a "radiator" and some sort of unrevealed secret

No, it's not. Lots of servers draw 400w in a 1U case. 4U cases can handle well over 1000w, easily.

I bet it's both... they use a lot of chips for each miner because of the 110nm process, which probably results in a largish PCB, and 400w is a lot of heat to dissipate out of a closed box.

It's cute the way BFL designs their products & packages to make them all pretty. I think that they think they are releasing the new iPad or iPod.

News flash: No one cares what this stuff looks like, just as long as it works. The Avalon is uglier than shit, but at least they're shipping.

Doesn't even begin to address the question. Everybody knows smaller die size = generally lower power consumption and better profitability for a chip maker. The problem is that BFL spits out this 1w/Gh number and people quote it like it's gospel. It's not even a power range as would have been prudent for BFL to release as they clearly do not know what the final power consumption spec will be. A jalepeno is very unlikely to have the same power consumption profile as a minirig SC, but they're the same according to BFL and people eat it up.

Josh,

What do you think of the mammoth case Avalon just released pics of?  Do you think the PCB takes up the entire footprint?  Did they do this to spread the ASIC chips apart because of heat?  Just curious.

Look! We can put boxes into boxes. More evidence that our ASICS are real:








We can even prove the existence of God!

Inaba runs out of money https://bitcointalksearch.org/topic/mtgox-is-apparently-out-of-money-137923

65nm designs almost always outperform 90nm and especially 110nm designs by a wide margin. Their ASIC designers would have to be unusually stupid in order to fail to produce chips with higher GH/J than all of the current competitors. FPGA power estimates are arguably easier to screw up, especially if you're working with an unusual FPGA.

To be a success after the difficulty rises, it doesn't have to be a "firm number". It just has to win.

So, the largest computer processor--Bitcoin--is about to get twelve times larger. And to think we don't need to build a moat around it.

I will definitely use http://blockorigin.pfoe.be/ next time...it does seem to be a bit more accurate, which will surely have an effect on overall stats.

As for workers, it's a pain in the ass to figure out.
For some pools, like deepbit or 50btc where they have a list of all the pools teams (containing all workers and hashing rates), it's pretty easy. I copy all of the data on workers/teams I can find into Excel. If it copies correctly, I can get a full table of all users, hash rates, etc...then I can clear the formatting and use the normal Excel functions to grab sums, averages, and the like. Pulling 7,000+ workers of data out of deepbit's tables, for example, isn't so hard after that.

For other pools, like BTCGuild, where they barely report anything other than "leaderboard" based hash rates, it's super hard to figure out accurate worker counts.

Overall, I try to use the accurate data displayed by the reliable pools to figure out what percentage of workers contribute the most to the hashrate...and then apply an average percentage to pools that don't give out precise information. Really it all comes down to an educated guess at best, but it's still better than assuming. Though nothing compared to your phenomenal work Organofcorti


Thanks for the kind words guys, I appreciate it. If anyone finds an easier way to pull data from some of the pools, let me know! Always happy to update my paper again