Topic: Butterfly Labs (Read 9783 times)

Thanks, I wasn't aware that I could do that. I thought it was a mod only feature. Here is an article proving one of my points btw http://www.edn.com/design/integrated-circuit-design/4313790/Making-ASIC-power-estimates-before-the-design.

Since you're the one who started the thread, you can easily close it yourself. Look for the lock option in the lower left hand corner.

This is why I'm asking for the thread to be closed. I don't believe anything more productive can be done here. I've stated my opinion and asked my questions. BFL's ceo has come here and given a direct response. You really can't ask for more than that. I wish for the best from BFL and I hope they can deliver what they have promised.

Am I mistaken or is somebody playing the semantics card with a person whose relatively straightforward and kind in regard to his recent posts, but is being called out as a tool?

I'm not getting this business practice at all. Somebody link me a memo showing that this is the new norm so I can start calling my custom base and suppliers cocksuckers, for currently I'm not on track to beat last year's sales, down about 66% to date.

This is a direct claim that your company has sold the 1500 GH/s rigs under the preorder model. You can claim it is a lie. It looks to me that you guys were pretty certain in the whole 5 GH/s 5 watt consumption estimate until you learned the hard way that ASIC design doesn't always work the way you'd like.

Even if the pre-order claim isn't true, why are you advertising on your site the minirig? https://products.butterflylabs.com/homepage/1500gh-bitcoin-miner.html. We both know it physically can't exist with the power required for the ASICs you are shipping at the end of the month. Thus you somehow have a dramatically more power efficient design in your lab and you're somehow going to release this in a reasonable timeframe? Don't you see how people can get a little POed with these claims?

I'm asking you to be a good CEO and do two things. First, white an open letter to your customers whom you've put through hell with persistent delays apologizing and promising to do better. It sucks, but its something you have to do as the leader. You guys dropped the ball. Second, pull the minirig from the market and refund any preorders. You have a legitimate product on the market that holds well against the competition. You'll get to the minirig, but we both know it isn't possible with what you have right now.

You guys have done some good work building a real product. The conversation we'd be having if you would have been conservative with release dates and the wattage claims would be nothing but praise for a beautiful and functional product for mainstream miners. You still have an opportunity to get it back on track. Don't waste it.   

According to SEC and FC rules pre-order customers (people paid 100% upfront, no product) are not customers - they are investors. Therefore they do have legal rights to get this kind of information.

Josh, why are you being an arse again? You are missing to opportunity to polish up Butterfly Labs corrupted image.


Old posts from BFL's Facebook page .....
20th September 2012


27th November 2012


13th December 2012


18th January 2013


22nd January 2013


22nd January 2013


26th February 2013


26th February 2013


Made me chuckle.... got to the 26th February and BFL are like "Screw it.... "

Really, can you point to where BFL is currently selling a 1500 GH/s rig?

You have a 1500 GH/s rig listed in your product catalog and jml has claimed you accepted preorders josh.

Can you point out where BFL is claiming that? 

Again, just another example of why the entire premise for this thread is complete nonsense.  You have done exactly zero due diligence and your claims are either invalid, irrelevant or already resolved.

If the top tier product was a 50 GH/s rig, then yes no one would really care. Let's assume the 5 GH/s unit consumed 100 watts. Then when you scale up to 50 (assuming something linear) you can target 1 KW which is on par with a quad-crossfire 7970 rig. You've beat the competition watt for watt dramatically. Everyone is happy and the world moves on.

But when you talk about a terahash. If your 5 Gh/s unit runs at 5 watts, then scaling makes sense. It's still very hot and very hard to cool, but it at least makes sense. In the 40-50 watt range, it doesn't to me and several other members here. More power = more heat.

Also when they were claiming the 5 watt range that meant you could run the unit off of a usb only. Put the unit in a small, very portable case, use a small fan if any at all, and finally long term persistent use is less likely to destroy the unit. Less heat, longer lifespan.

I actually agree with you about the 5 GH/s system which is why I ordered one in the first place. I could care less about the power consumption. My 7950 consumes a heck of  alot more. But it just bothers me deeply they claim to be able to deliver something like a 1500 GH/s rig within any reasonable timeframe. I can't understand how they can justify that claim. Look at Intel with prescott http://www.theinquirer.net/inquirer/news/1036882/is-intel-prescott-p4-hot-handle

They had units in a lab running at 8 GHz with liquid nitrogen cooling  http://www.geek.com/chips/is-intel-planning-to-clock-pentium-4s-to-8ghz-547912/. It eventually became clear that the roadmap didn't make any sense. They had to bite a very bitter bullet and let AMD trash them while they retooled.

I see it as a small company that hasn't learned throwing everything at the wall and seeing what sticks isn't a good business strategy. Be good to your consumers and treat them like family. Be honest and open. Admit the mistakes you've made and when you over-promise understand that it comes with a cost.

If it's under a hundred watts and I can actually buy one, and it's actually 5 GH/s, that would all be enough for me.

Who cares if its 5 watts or 50 watts?  It's a best guess while you're developing something.

The bullshit comes when you claim you can build a 1500 GH/s rig and you accept money for it when you are having difficulty delivering a 5 GH/s rig under double digit wattage. It would be like IBM saying that they can't get a blade under 5000 watts, but don't worry about the server with 300 blades. You'll be able to cool it. I'm skeptical. I have every right to be. I did not purchase a minirig so I don't have a dog in this fight and Josh gets to do a mega victory dance if he delivers.

And frankly if he does, then I will personally apologize and recommend BFL to everyone. They would have done something really amazing from an engineering perspective akin to when AMD released the FX-51.

...that sounds vastly different from bullshit excuses.

As a disclaimer, my only experience in VLSI is with some projects in the aerospace industry. IMHO, usually the last thing you make a comment on is power consumption. It can take months of fine tuning and hard work, which is why Nvidia and AMD for example almost never comment on power consumption for their new graphics cards until they start shipping review units. Reliability testing and turn around time from the fabricator also has a lot to do with the schedule you live on and your relationship with the fabricator. Generally speaking it depends on the volume you ordered and how much time you have allocated for quality control. Firmware can take a great deal of time as well.

ASIC development is hard and the people who do it well are few and far between. Which is why it is a terrible idea to announce any release date until you have a working prototype in your lab and you're doing things like quality and firmware. Even then small factors can result in delays for several months. For many larger companies, some projects simply die because the internal benchmarks are just simply unobtainable for example Intel's larrabee project: http://blogs.intel.com/technology/2010/05/an_update_on_our_graphics-rela/. Here we have probably the most advanced and respected chip designer in the market admitting defeat for a very public and well financed project. This should give you an idea how hard developing a product can be. I commend BFL for taking on the project and dealing with the delays.

Another example of problems: http://www.pcper.com/news/Graphics-Cards/UPDATE-NVIDIA-Fermi-continues-have-problems-according-multiple-sources

In your experience designing and testing electronics requiring new chips to be designed, what is the average time to market?  What kind of delays do you usually experience, and what reasons/excuses have you found to be valid?

yes I have spent money at BFL. I don't care if it is a lost investment or not as it is sub-500 dollars. I can live. My issue is that their response to any question has always been with personal attacks to both intelligence and character. This is not how you run a company and when bitcoins enter the mainstream, they will be crushed. No one wants to do business with a company that feels it can have long delays without any legitimate excuse while the customer watches other making money with a competitor's products. But again this format has lost its ability to accomplish anything. I request that the thread be closed. 

I think we should do an audit of whether or not you spent any money, or are just bellyachin' for free.

Answering any further questions is pointless. We can't have a conversation in this format. I will wait and see what happens.

You still haven't answered my question - how much money do you have invested in this?

Alright we are done here. Moderator could you close this thread. Nothing more can be accomplished in this format.