Better performance, more energy efficient, it must be a completely different project. It's all two months before shipment. What will be next? What if BFL increase performance their Single "SC" to 120Ghash? Avalon beats them with 240Ghash? and BFL beats Avalon with 480Ghash? In a few months they probably will offer several Thash/s chips but in this way they not finish any project still starting new ones
I hope not
Unless BFL released disinformation, their SC rig with 8 chips can only go up to (upper limits) 120GH/s. @ 1GHz @ ~120watts
This is probably unsafe for the BFL hardware over the long term. They have so far hinted their (safe) overhead is somewhere around 75% of maximum...about 750MHz or somewhere near there with (I suppose) stock cooling. If they keep raising the temps the hardware is more likely to fail which increases returns and warranty claims etc.
--------------
Right now the two companies are seemingly playing a game of chicken. The first to veer off is going to lose alot of customers orders and pre-orders.
Right now some BFL customers are
clamoring for BFL to reflash the Single SC to 65GH/s.
Supposedly this is because of a dollars difference. I think it is just that they saw the new Avalon Specs and don't want to be stuck holding an inferior product.
All this costs BFL money, time and adds to delays in shipping out the hardware to their customers. Imagine if you are standing in BFL headquarters and BFL CEO asks you to reflash a thousand boxes from 40Gh/s to 60Gh/s. Then 15 days later when you are almost done, he asks you to do it again from 60Gh/s to 65Gh/s. Hopefully people will see the point I am making.
They (BFL) could fix this rat race issue by either telling their customers "NO, we won't reflash it again!" OR give the customers themselves the option to flash it themselves (distributing the workload to the customers). But this itself may cause secondary issues and firmware hacking as well as giving away a potential post-sale profit margin.
Consider this:
--If users can flash their box themselves, well then they can hack the firmware and run the boxes at any speed they want.
--If the boxes fail, how does BFL know what the clockspeed was when it comes in for warranty repair? (That costs money)
--If they give away a copy of the firmware, what prevents the competition from looking into the code and gleaning details about the way the first gen BFL Asic was made?
--------------------------------
There is a serious danger of BFL running themselves into the ground and many people losing their money if things don't play out just right.
Avalon can cause BFL harm (buisness wise) simply by incrementally releasing slightly better specs at regular intervals. The BFL customers will keep demanding more. It costs Avalon virtually nothing to keep this up, but it costs BFL quite a bit as they push their hardware closer and closer to the edge. (Thermally,componentary as well as warranty wise)
Consider this:
What if BFL just ordered 100,000 capacitors that will work up to 60 watts but not up to 85 watts? Do they just throw away several thousands of dollars worth of hardware because they need to appease customers demands??
(The reason behind the introduction of Little SC at 30Gh/s??) Every PCB design has an envelope with specific limits and design headroom. Each time they "up" something, they risk having to buy better components and analyzing whether the current design is capable of handling the extra performance or electrical requirements.
It's a game of chicken. BFL doesn't ultimately know what the intended final design specs are for the Avalon mining device.
It could be 60....could be 70....maybe 80 or 90?
All I do know is they (BFL) just jumped from 40Gh/s to 60Gh/s. They had clearly aimed for 40Gh/s. Now people want them to go higher up to 65Gh/s. As a company you gamble in all sorts of ways every time they increase the specs this late in the game.
