I mined for a very tiny 'profit' for a few months when BTC was hovering in the $1-3 range. I however kept mining and hoarded the coins until I sold at $12 USD in December of last year. I now reget it but it was still more of a profit than had I sold the coins the moment I mined. I'm now at the point where I'll be mining with GPUs until I'm mining for a slight loss and then I'll switch to LTC. I'll be holding these bitcoins as I've recovered all my original investment and then some. Hopefully my current 'loss' will yield a big profit in the longterm.

I doubt it. However, it would cost more per die and require QA time and that isn't worth it to AMD.

The communication between the chips and the controller is minimal. I am also of the opinion that the more compact design would be better but for compute density reasons. BFL excites me for the possibility of watercooling their chips and ending up with an insanely dense mining cluster. I hope in a few generations BTC ASICs are something that we could order like FPGAs now and assemble custom PCBs to suit our own desires.

Careful with that wording.

I do agree that right now is a bad time to get into GPU mining. I'm still making ~$600 USD/month at the current conversion with my GPUs but I fully expect that in ~1 month I'll be losing money running them. At the moment I'd just wait until BFL and Avalon have their backlogs trimmed to reasonable lengths.

Excellent point. I have minimal knowledge of using the the ground pad and vias to transmit heat through the PCB but it makes logical sense that Avalon would use a larger than electrically necessary PCB to give each chip some buffer from neighboring chips.

Avalon went with a very simple chip on a large(110nm) process. I imagine this strategy gave them very high yields out of the wafers because of the small chip size. This came at the cost of a large PCB to route to all the chips. BFL appears to be going the other way with a few very large chips based on a smaller(65nm) process. This should give them better power efficiency and more hashing power per chip at the risk of random manufacturing defects giving them fewer of these chips per wafer. Of course one hopes for 100% yields but that isn't always the case. We'll see how this all plays out very soon.

Gave up on cgminer. Compiled reaper, plugged in the following just to get running..

worksize 256
vectors 1
aggression 20
threads_per_gpu 1
sharethreads 32
lookup_gap 2
gpu_thread_concurrency 8192

and I'm up and running at ~600 KH/s reported from reaper and the pool. cgminer was a mess trying to find a combination of thread-concurrency, lookup-gap and aggression that produced low hardware errors and high utility. Highest I achieved with cgminer was ~440 KH/s but lots of hardware errors. Best utility was 26 shares/minute compared to ~31 shares/minute with reaper. Not sure why I couldn't run a thread_concurrency higher than lookup_gap * 1024 but I'm happy now.

Ok, I'll dump the card into another rig with more ram and try increasing thread-concurrency.

Any idea what temps the modules and heatsinks are hitting? Curious if those heatsinks were overkill or necessary.

Bump.

Not necessarily true. I'd purchase a unit from the 3rd batch if the price were cut in half* or so. Once their fixed costs are paid off it should cost very little to product additional units as they're just paying wafer/packaging costs and chassis/pcb/psu/shipping costs. Let the first two batch customers pay off their gear and then slash the price.


*- fraction pulled out of thin air. I'd have to do the maths but if it paid back in 6-9 months I'd be happy with lower efficiency.

Is it normal to have so many reported hardware errors?

Here are my cgminer settings: cgminer --scrypt -o http://LHPool.eu:8080 -u Gomeler.1 -p x -w 256 -I 18 -g 1 --lookup-gap 2 --thread-concurrency 1120

Drivers: 13.1 x64

OS: Ubuntu 12.04 x64

and it running a bit. This is a HD 6870x2 with 1120 shaders. Setting --thread-concurrency to higher than 1120 causes it to throw some bizarre openCL error about memory being greater than available. Top shows ~1.1GB free.

cgminer version 2.10.4 - Started: [2013-02-01 01:22:24]
--------------------------------------------------------------------------------
 (5s):487.8K (avg):456.8Kh/s | Q:10  A:19  R:2  HW:0  E:190%  U:5.0/m
 ST: 0  SS: 0  DW: 1  NB: 4  LW: 24  GF: 0  RF: 0
 Connected to LHPool.eu diff 15 with LP as user Gomeler.1
 Block: 33dae6ae1d602569...  Diff:1.33M  Started: [01:24:40]  Best share: 275
--------------------------------------------------------------------------------
 [P]ool management [G]PU management [ S ]ettings [D]isplay options [Q]uit
 GPU 0: 243.1K/241.9Kh/s | A:10 R:2 HW:34 U:2.64/m I:14
 GPU 1: 243.1K/214.9Kh/s | A: 9 R:0 HW:41 U:2.37/m I:14
--------------------------------------------------------------------------------

 [2013-02-01 01:21:56] Started cgminer 2.10.4
 [2013-02-01 01:21:57] Probing for an alive pool
 [2013-02-01 01:22:24] Long-polling activated for http://LHPool.eu:8080/LP
 [2013-02-01 01:22:44] Rejected 0cb20944 Diff 20/15 GPU 0  (prevhash-stale)
 [2013-02-01 01:22:44] LONGPOLL from pool 0 detected new block
 [2013-02-01 01:22:48] Rejected 072b736f Diff 35/15 GPU 0  (prevhash-stale)
 [2013-02-01 01:22:56] Accepted 0e7e1c78 Diff 17/15 GPU 1
 [2013-02-01 01:22:56] LONGPOLL from pool 0 detected new block
 [2013-02-01 01:23:01] Accepted 095d0243 Diff 27/15 GPU 1
 [2013-02-01 01:23:16] Accepted 03b91ba9 Diff 68/15 GPU 1
 [2013-02-01 01:23:20] Accepted 0588028e Diff 46/15 GPU 1
 [2013-02-01 01:23:21] Accepted 06f464e0 Diff 36/15 GPU 0
 [2013-02-01 01:23:22] Accepted 0d398263 Diff 19/15 GPU 0
 [2013-02-01 01:23:38] Accepted 00ee0f22 Diff 275/15 GPU 0
 [2013-02-01 01:24:06] Accepted 085aae8c Diff 30/15 GPU 0
 [2013-02-01 01:24:27] Accepted 06abc42b Diff 38/15 GPU 1
 [2013-02-01 01:24:40] Accepted 0e21986c Diff 18/15 GPU 0
 [2013-02-01 01:24:40] LONGPOLL from pool 0 detected new block
 [2013-02-01 01:24:56] Accepted 0a849586 Diff 24/15 GPU 1
 [2013-02-01 01:24:56] Accepted 0488b40e Diff 56/15 GPU 0
 [2013-02-01 01:25:03] Accepted 03d595f3 Diff 66/15 GPU 0
 [2013-02-01 01:25:03] Accepted 097c66d4 Diff 26/15 GPU 1
 [2013-02-01 01:25:17] Accepted 0372c6fd Diff 74/15 GPU 1
 [2013-02-01 01:25:18] Accepted 0e1c6ccb Diff 18/15 GPU 0
 [2013-02-01 01:25:25] Accepted 0b376340 Diff 22/15 GPU 0
 [2013-02-01 01:25:38] Accepted 049bbd06 Diff 55/15 GPU 1
 [2013-02-01 01:25:43] Accepted 0f036fea Diff 17/15 GPU 0

Stopped reading after that. I'll just check back in a few weeks to see what ever happened to bASIC.

There is a valid fear of a savvy entity that sees BTC as a threat. If said entity decided it wanted to try and destroy BTC it could invest a few million dollars to protect hundreds of millions of dollars of future revenue. At the moment you could attack BTC legally by funding lobbyists or by attacking the network itself. Attacking the network, given a few days of reading about BTC, would logically happen via an ASIC of comparable efficiency to what is being released to consumers now.

Assuming Avalon made $0 USD on their first batch, I believe it was 300 units @ $1300/each. That's a paltry $390000 USD for 18 TH/s. In reality the fixed costs of the masks are likely a fraction of this and the chips are in the $1-10 USD/chip range. Pulling numbers out of my ass, throw $1-5 million USD in renting rackspace and producing 300+ TH/s to counter the expected production of BFL and Avalon for the next 6 months and you can effectively control the network. If you are wise you'll wait until you have an astronomical amount of hashing power and then flip the switch at one moment and double-triple the network hash rate. Guaranteed within hours BTC miners/developers will notice "oh shit, someone owns the network". Crash user trust in the security of the network and crash the value of BTC. Before the developers can change the protocol and get enough users over to the new network BTC will be trashed. Sure, other more complicated cryptocurrencies may crop up(LTC for example) but the damage is done.

THIS is what I fear and it'll be an issue until the cost to attack the network exceeds what I'd consider to be 'pocket change' for larger entities. At this moment if my net worth was sufficient that I could throw $2-3 million at a private ASIC endeavor as 'fun' I'd do it just to fuck with people. It would be supreme trolling, not casual forum trolling.

Some motherboards may require you to bridge pins on the PCIe slot to indicate the link-width. I had this problem when using x1 to x16 adapters. It's simple work with a soldering iron or a little piece of wire.

Doesn't make financial sense. Icarus boards likely are worthless to them but they are running a similar program to BFL to ensure repeat customer business. Perhaps value an Icarus at $100 trade-in against a module though and take in another $200-250 in cash to cover costs + profit.

I think he's making a statement on the fact that previously 60 GH/s would have consumed 10-30 kW of power compared to 400 W. All sorts of implications with the power consumption figures but they're best discussed in a separate thread.

http://garzikrants.blogspot.se/2013/01/avalon-its-alive.html

At this point I'm just mashing ctrl + r to see what else he posts 

Hopefully someone will disassemble the unit after they report on the power consumption and hashrate.

That's a massive single heatsink, I guess there are multiple modules that connect to the visible PCB 'backplane' via ribbon cables. Then 3 backplanes at ~22 GH/s @ ~6 watts per gigahash yields ~132w which could be easily powered by a pair of 6-pin PCIe connectors. Top of the unit looks like a bundle of just yellow and black so.. PCIe connectors. Can't wait for someone to tear this thing apart.

Just a +1 to coretechs being an honest trader. I'd pick these up but I'm in the same nervous boat re: Avalon.