Pages:
Author

Topic: March 30th BFL ASIC update - page 2. (Read 7754 times)

legendary
Activity: 1988
Merit: 1012
Beyond Imagination
March 29, 2013, 02:52:57 PM
#84
5000X8=40TH, that's all added hash power if all of the first set of 6 wafers delivered
newbie
Activity: 56
Merit: 0
March 29, 2013, 02:46:11 PM
#83
Let me take one more shot at this correct me if I'm wrong

Quote
The first 5k chips will be arriving at the assembly house on Thursday
...
Quote
hat's the current timeline, and as usual, subject to change.
So in theory they could have been bumped and packaged, but not 100%. I assume they get shipped from Asia, so if he said that two days until arrival they were probably already shipped; so your right. He either lied about the wafers arrival date (earlier than expected) or they hadn't shipped yet and he said they did. Either way iffy. Thnks for the breakdown

No, for the last 2 months, josh has been talking as if all 5 wafers out of the first group were processed as one batch, which took 6 weeks to go from wafer to chips. He then suggested the second group of 6 wafers would ship 1-1.5 weeks after that.

The chat log last night now shows that 1 wafer was processed in that 6 weeks, and they are now finishing the remaining 4 wafers out of batch 1, with the 6 wafers from batch 2 nowhere in sight, and possibly delayed by a needed chip revision.

What this suggest to me is that they didn't actually order their chips when they said they did, and that what they recieved from the packaging facility was actually their engineering samples.

To compound things, josh has constantly shit on avalon for only shipping 300 units, saying BFL can ship 300 per day, today we learn BFL only has at most 1000 finished chips which can't even produce 300 units, unless 200 of them are jalapenos which due to the power usage of the current chips, will only be able to hash at 1-2 mh/s
legendary
Activity: 1876
Merit: 1000
March 29, 2013, 02:42:10 PM
#82


If the power/heat issues still exist with the BitForce Single SC(60GH unit), then BFL will option to ship two  Little Single SC(30GH units) in its place until the problem is resolved.


They will ship two full SC's that have been downclocked to total 60Gh.  all chips will be on each board.
sr. member
Activity: 279
Merit: 250
March 29, 2013, 02:29:39 PM
#81
First batch is 5000 chips, first order # for august maybe 3800, starting from 1600. Of course it won't be enough to cover june-23 till july-23 orders, no way an august one will be part of it.
If you ordered a big single they might do 2x little singles since the power limit is per board, not per chip. The chips run at full spec without a problem, the problem seems to be with the boards or power regulators. Same for the mini-rig's setup, if they don't find the culprit/a fix soon, they will send 2 mini-rigs, same total number of chips, but double amount of boards, cases, psu.

We've learned today that the first batch is 1000 chips. They don't even have the third wafer yet, and the first one was destroyed for nothing.

They are rolling 4 wafers off the line "shortly." They may even be ready now [assuming you take his word for it].

http://www.reddit.com/r/Bitcoin/comments/1b8hvt/bfl_josh_updates_on_asic_status_full_transcript/

Yeah, and now all that's needed is shipping, dicing, shipping, bumping, shipping, packaging, shipping, mounting, shipping, and testing. it only took 3 months last time!

Quote from: Josh on march 13
Soon. The first 5k chips will be arriving at the assembly house on Thursday, the second set about a week or week and a half after that, and the then rest of the chips another week or week and a half after that. That's the current timeline, and as usual, subject to change.

https://forums.butterflylabs.com/bfl-forum-miscellaneous/1300-when-bulk-first-75-wafers-will-delivered-bfl.html#post18270

Caught in a lie. At this point, had they only processed 1 wafer, josh would have known about it and wouldn't have said 5k chips.

Just playing devil's advocate here.... but that post you link to does kind of line up with what he said in the chat last night (save for ~2 week delay, which he does say is possibility in your link).


No it doesn't, not at all. His post suggests 5 wafers were bumped and packaged. The second set of 6 wafers would take another week, and the balance of 75 wafers the week after that. In the chat, we found out 1 wafer was bumped and packaged, they are still waiting on the other 4 wafers out of the first batch, and who knows what's going on with the second batch of 6 wafers. He lied, either then when all along he suggested 5 wafers were being processed, or now when he is saying wafers 3-6 are currently being finished.

Let me take one more shot at this correct me if I'm wrong

Quote
The first 5k chips will be arriving at the assembly house on Thursday
...
Quote
hat's the current timeline, and as usual, subject to change.
So in theory they could have been bumped and packaged, but not 100%. I assume they get shipped from Asia, so if he said that two days until arrival they were probably already shipped; so your right. He either lied about the wafers arrival date (earlier than expected) or they hadn't shipped yet and he said they did. Either way iffy. Thnks for the breakdown
newbie
Activity: 56
Merit: 0
March 29, 2013, 02:21:05 PM
#80
First batch is 5000 chips, first order # for august maybe 3800, starting from 1600. Of course it won't be enough to cover june-23 till july-23 orders, no way an august one will be part of it.
If you ordered a big single they might do 2x little singles since the power limit is per board, not per chip. The chips run at full spec without a problem, the problem seems to be with the boards or power regulators. Same for the mini-rig's setup, if they don't find the culprit/a fix soon, they will send 2 mini-rigs, same total number of chips, but double amount of boards, cases, psu.

We've learned today that the first batch is 1000 chips. They don't even have the third wafer yet, and the first one was destroyed for nothing.

They are rolling 4 wafers off the line "shortly." They may even be ready now [assuming you take his word for it].

http://www.reddit.com/r/Bitcoin/comments/1b8hvt/bfl_josh_updates_on_asic_status_full_transcript/

Yeah, and now all that's needed is shipping, dicing, shipping, bumping, shipping, packaging, shipping, mounting, shipping, and testing. it only took 3 months last time!

Quote from: Josh on march 13
Soon. The first 5k chips will be arriving at the assembly house on Thursday, the second set about a week or week and a half after that, and the then rest of the chips another week or week and a half after that. That's the current timeline, and as usual, subject to change.

https://forums.butterflylabs.com/bfl-forum-miscellaneous/1300-when-bulk-first-75-wafers-will-delivered-bfl.html#post18270

Caught in a lie. At this point, had they only processed 1 wafer, josh would have known about it and wouldn't have said 5k chips.

Just playing devil's advocate here.... but that post you link to does kind of line up with what he said in the chat last night (save for ~2 week delay, which he does say is possibility in your link).


No it doesn't, not at all. His post suggests 5 wafers were bumped and packaged. The second set of 6 wafers would take another week, and the balance of 75 wafers the week after that. In the chat, we found out 1 wafer was bumped and packaged, they are still waiting on the other 4 wafers out of the first batch, and who knows what's going on with the second batch of 6 wafers. He lied, either then when all along he suggested 5 wafers were being processed, or now when he is saying wafers 3-6 are currently being finished.
sr. member
Activity: 279
Merit: 250
March 29, 2013, 02:13:45 PM
#79
First batch is 5000 chips, first order # for august maybe 3800, starting from 1600. Of course it won't be enough to cover june-23 till july-23 orders, no way an august one will be part of it.
If you ordered a big single they might do 2x little singles since the power limit is per board, not per chip. The chips run at full spec without a problem, the problem seems to be with the boards or power regulators. Same for the mini-rig's setup, if they don't find the culprit/a fix soon, they will send 2 mini-rigs, same total number of chips, but double amount of boards, cases, psu.

We've learned today that the first batch is 1000 chips. They don't even have the third wafer yet, and the first one was destroyed for nothing.

They are rolling 4 wafers off the line "shortly." They may even be ready now [assuming you take his word for it].

http://www.reddit.com/r/Bitcoin/comments/1b8hvt/bfl_josh_updates_on_asic_status_full_transcript/

Yeah, and now all that's needed is shipping, dicing, shipping, bumping, shipping, packaging, shipping, mounting, shipping, and testing. it only took 3 months last time!

Quote from: Josh on march 13
Soon. The first 5k chips will be arriving at the assembly house on Thursday, the second set about a week or week and a half after that, and the then rest of the chips another week or week and a half after that. That's the current timeline, and as usual, subject to change.

https://forums.butterflylabs.com/bfl-forum-miscellaneous/1300-when-bulk-first-75-wafers-will-delivered-bfl.html#post18270

Caught in a lie. At this point, had they only processed 1 wafer, josh would have known about it and wouldn't have said 5k chips.

Just playing devil's advocate here.... but that post you link to does kind of line up with what he said in the chat last night (save for ~2 week delay, which he does say is possibility in your link).
full member
Activity: 146
Merit: 100
March 29, 2013, 02:12:21 PM
#78
April Fool's joke
newbie
Activity: 56
Merit: 0
March 29, 2013, 02:00:54 PM
#77
First batch is 5000 chips, first order # for august maybe 3800, starting from 1600. Of course it won't be enough to cover june-23 till july-23 orders, no way an august one will be part of it.
If you ordered a big single they might do 2x little singles since the power limit is per board, not per chip. The chips run at full spec without a problem, the problem seems to be with the boards or power regulators. Same for the mini-rig's setup, if they don't find the culprit/a fix soon, they will send 2 mini-rigs, same total number of chips, but double amount of boards, cases, psu.

We've learned today that the first batch is 1000 chips. They don't even have the third wafer yet, and the first one was destroyed for nothing.

They are rolling 4 wafers off the line "shortly." They may even be ready now [assuming you take his word for it].

http://www.reddit.com/r/Bitcoin/comments/1b8hvt/bfl_josh_updates_on_asic_status_full_transcript/

Yeah, and now all that's needed is shipping, dicing, shipping, bumping, shipping, packaging, shipping, mounting, shipping, and testing. it only took 3 months last time!

Quote from: Josh on march 13
Soon. The first 5k chips will be arriving at the assembly house on Thursday, the second set about a week or week and a half after that, and the then rest of the chips another week or week and a half after that. That's the current timeline, and as usual, subject to change.
https://forums.butterflylabs.com/bfl-forum-miscellaneous/1300-when-bulk-first-75-wafers-will-delivered-bfl.html#post18270

Caught in a lie. At this point, had they only processed 1 wafer, josh would have known about it and wouldn't have said 5k chips.
sr. member
Activity: 279
Merit: 250
March 29, 2013, 01:52:52 PM
#76
First batch is 5000 chips, first order # for august maybe 3800, starting from 1600. Of course it won't be enough to cover june-23 till july-23 orders, no way an august one will be part of it.
If you ordered a big single they might do 2x little singles since the power limit is per board, not per chip. The chips run at full spec without a problem, the problem seems to be with the boards or power regulators. Same for the mini-rig's setup, if they don't find the culprit/a fix soon, they will send 2 mini-rigs, same total number of chips, but double amount of boards, cases, psu.

We've learned today that the first batch is 1000 chips. They don't even have the third wafer yet, and the first one was destroyed for nothing.

They are rolling 4 wafers off the line "shortly." They may even be ready now [assuming you take his word for it].

http://www.reddit.com/r/Bitcoin/comments/1b8hvt/bfl_josh_updates_on_asic_status_full_transcript/
newbie
Activity: 56
Merit: 0
March 29, 2013, 01:47:20 PM
#75
I was going to suggest they could just slap single heatsinks into the jalapeno, and ship out a big pile of jalapenos, but if a single chip is using 9-15 watts, and the jalapeno was designed for 4.5 watts, and is powered through USB through 2 USB plugs designed to pull at most 2.5 watts each.

Well the jalapeno won't work at all with the current chips.
sr. member
Activity: 280
Merit: 250
March 29, 2013, 01:36:47 PM
#74
First batch is 5000 chips, first order # for august maybe 3800, starting from 1600. Of course it won't be enough to cover june-23 till july-23 orders, no way an august one will be part of it.
If you ordered a big single they might do 2x little singles since the power limit is per board, not per chip. The chips run at full spec without a problem, the problem seems to be with the boards or power regulators. Same for the mini-rig's setup, if they don't find the culprit/a fix soon, they will send 2 mini-rigs, same total number of chips, but double amount of boards, cases, psu.

If you ordered an Audi A6 and got 2x Nissan Micras that wouldn't really cut it. Yea, you would have the same amount of horsepower and topspeed combined but still it's a lesser product.

The increased power comsumption makes all the difference when ASICs really flood the market. When difficulty skyrockets all the powersavings promised to you but not delivered could make the difference between profitable mining and not profitable mining. The lifetime earnings of BFL ASIC is also affected by higher power usage.

And the delivery dates... If they will ship 2x amount of chips to fill one order those delivery delays will be tremendous. All currently sold Avalons will be shipped before BFL has shipped all units purchased last year.
sr. member
Activity: 434
Merit: 250
March 29, 2013, 12:48:37 PM
#73
First batch is 5000 chips, first order # for august maybe 3800, starting from 1600. Of course it won't be enough to cover june-23 till july-23 orders, no way an august one will be part of it.
If you ordered a big single they might do 2x little singles since the power limit is per board, not per chip. The chips run at full spec without a problem, the problem seems to be with the boards or power regulators. Same for the mini-rig's setup, if they don't find the culprit/a fix soon, they will send 2 mini-rigs, same total number of chips, but double amount of boards, cases, psu.

We've learned today that the first batch is 1000 chips. They don't even have the third wafer yet, and the first one was destroyed for nothing.

Not for nothing. That decision could very well make Josh's gear more valuable.
newbie
Activity: 56
Merit: 0
March 29, 2013, 12:44:59 PM
#72
First batch is 5000 chips, first order # for august maybe 3800, starting from 1600. Of course it won't be enough to cover june-23 till july-23 orders, no way an august one will be part of it.
If you ordered a big single they might do 2x little singles since the power limit is per board, not per chip. The chips run at full spec without a problem, the problem seems to be with the boards or power regulators. Same for the mini-rig's setup, if they don't find the culprit/a fix soon, they will send 2 mini-rigs, same total number of chips, but double amount of boards, cases, psu.

We've learned today that the first batch is 1000 chips. They don't even have the third wafer yet, and the first one was destroyed for nothing.
hero member
Activity: 568
Merit: 500
March 29, 2013, 12:41:42 PM
#71
First batch is 5000 chips, first order # for august maybe 3800, starting from 1600. Of course it won't be enough to cover june-23 till july-23 orders, no way an august one will be part of it.
If you ordered a big single they might do 2x little singles since the power limit is per board, not per chip. The chips run at full spec without a problem, the problem seems to be with the boards or power regulators. Same for the mini-rig's setup, if they don't find the culprit/a fix soon, they will send 2 mini-rigs, same total number of chips, but double amount of boards, cases, psu.
sr. member
Activity: 279
Merit: 250
March 29, 2013, 12:37:28 PM
#70
If you ordered today just read Josh's policy:
Quote
If would would prefer to wait for a unit after we've made some changes to the systems that need a bit of tweaking, we will be happy to put your shipment on hold. However, if you'd rather have the units right now at an increased power usage, we will ship you as many units as required to get you to the hashrate your purchased, if we end up having to scale back any given class of unit to fit within the power envelope of the current board design.

If you DONT want the power hungry ones, contact BFL and let them know; they will hold your order until the new ones are ready. They don't have an specific estimate as to when a fix would be rolled out to production units, so if they fix quickly, yes you probably would get a newer one if you ordered today, but if it is difficult to solve, well, then, your SOL. We need more info tbch.

"X: josh do we have any idea of if we will reach sept or aus orders in this batch
BFL_Josh: In the first 5 wafers? No, I doubt it."

http://www.reddit.com/r/Bitcoin/comments/1b8hvt/bfl_josh_updates_on_asic_status_full_transcript/

It implies there are not enough chips and boards to fullfill even last year orders, so called 1st batch. Ordering now means getting into 3rd batch, or 4th?  Huh
Depends on how many cancelled orders too. Time will tell I guess.
full member
Activity: 148
Merit: 102
March 29, 2013, 12:22:36 PM
#69
Once the finished boards gets back from UL, they'll know exactly where the leak is. There is no way in hell that any of their units are going to leave their factory without UL approval, especially now that they're on record for stating there's a concern. Same true for FCC approval, of which Josh stated back in September he was waiting for the results from them then, expected within a couple weeks.

I'm pretty sure they will not have to deal with UL since they are not making the power supplies and will only be dealing with low voltage DC.
sr. member
Activity: 392
Merit: 250
March 29, 2013, 12:21:26 PM
#68
its not march 30 yet
full member
Activity: 141
Merit: 102
March 29, 2013, 12:19:38 PM
#67
well here is another way to look at it:

1. bought 1 single sc, got 2 single sc
2. stuck the biggest water cool heatsink you can find on each
3. overclock it as much as possible back to original spec (might need firmware hack)
4. Profit ?? Smiley


Lol, the power usage will skyrock and will probably damage the power circuit on the pcb.
newbie
Activity: 56
Merit: 0
March 29, 2013, 12:17:39 PM
#66
Quote
BFL_Josh: What's causing even more consternation is the fact that the wafer we burned for tests runs at far less power than a second wafer we mounted on the BGA package... so it may be a wafer by wafer thing, and sicne we only have two datapoints, it's hard to nail down the issue. We're rolling the other 4 wafers off the line shortly (they may be done already I think) - and that wil at least give us an idea of the power consumption between wafers with 6 data points.

They only have 1 wafer? so only 1000 chips to use so far. so josh gets his 2 minirigs which has turned into 4 minirigs because they've been underclocked to not catch fire, that's 800 chips, gone. that leaves 200 chips to make 25 singles, which now equals 12.5 singles because of the power/heat thing..

That is your 300 units a day, people, the equivalent of 12 singles shipped in april.
vip
Activity: 756
Merit: 503
March 29, 2013, 12:05:41 PM
#65
Ship it to me.  Cheesy
Pages:
Jump to: