ASICMINER: Entering the Future of ASIC Mining by Inventing It

[bitfair]

Agree. The unoptimized test results were good news. I don't know what these people who bailed were expecting? Nice opportunity to acquire. Hopefully more reasonably good news will drive it down further in the short term. 

A power consumption 100% above expected is "good news"?

You can argue that it's not entirely unexpected, you can argue that it still beats many competitors, and you can argue that it may improve.

But calling it "actually good news" sounds somewhat disingenuous, don't you think?

[necro_nemesis]

Agree. The unoptimized test results were good news. I don't know what these people who bailed were expecting? Nice opportunity to acquire. Hopefully more reasonably good news will drive it down further in the short term. 

A power consumption 100% above expected is "good news"?

You can argue that it's not entirely unexpected, you can argue that it still beats many competitors, and you can argue that it may improve.

But calling it "actually good news" sounds somewhat disingenuous, don't you think?

Put in perspective of being in the gen3 game or not, yes. I gather people underestimate the complexity of the task and just accept the ASIC will materialize as specified.

[teek]

One thing to note is rockxie seems to have initially tested the chips at .72v.  I'm definitely no EE,  but FC did mention that the chips were versatile with several different modes.  The 0.2J/G he mentioned originally was stated to be at .55v,  which probably means a lower clock as well.  Rockxies boards appear to have only one crystal.. i think given this setup he hasn't really been able to see what low power mode does (or overclocking for that matter).   Will low power mode dump the consumption down ~50% ?  MY limited experience tells me probably not,  but .72 to .55 is quite a difference in and of itself..

teek

[rudi]

A power consumption 100% above expected is "good news"?

You can argue that it's not entirely unexpected, you can argue that it still beats many competitors, and you can argue that it may improve.

But calling it "actually good news" sounds somewhat disingenuous, don't you think?

58% more (0.5539/0.35) than expected, not 100%. 

[mc_lovin]

One thing to note is rockxie seems to have initially tested the chips at .72v.  I'm definitely no EE,  but FC did mention that the chips were versatile with several different modes.  The 0.2J/G he mentioned originally was stated to be at .55v,  which probably means a lower clock as well.  Rockxies boards appear to have only one crystal.. i think given this setup he hasn't really been able to see what low power mode does (or overclocking for that matter).   Will low power mode dump the consumption down ~50% ?  MY limited experience tells me probably not,  but .72 to .55 is quite a difference in and of itself..

teek

+1

[minerpumpkin]

One thing to note is rockxie seems to have initially tested the chips at .72v.  I'm definitely no EE,  but FC did mention that the chips were versatile with several different modes.  The 0.2J/G he mentioned originally was stated to be at .55v,  which probably means a lower clock as well.  Rockxies boards appear to have only one crystal.. i think given this setup he hasn't really been able to see what low power mode does (or overclocking for that matter).   Will low power mode dump the consumption down ~50% ?  MY limited experience tells me probably not,  but .72 to .55 is quite a difference in and of itself..

teek

Well, the rated voltage is 0.72 and is supposed to consume 0,35 J/GH/s and does consume 0,5539 J/GH/s in rockxie's test. We're not looking at low power modes at the moment. Right now we are at 160% of the advertised consumption. Yet, I don't see any reason why this can't be optimized further. Look at the improvements KNC managed to obtain through software, and we're even talking about hardware optimizations here. This is an experimental board!

[juicejoyce]

One thing to note is rockxie seems to have initially tested the chips at .72v.  I'm definitely no EE,  but FC did mention that the chips were versatile with several different modes.  The 0.2J/G he mentioned originally was stated to be at .55v,  which probably means a lower clock as well.  Rockxies boards appear to have only one crystal.. i think given this setup he hasn't really been able to see what low power mode does (or overclocking for that matter).   Will low power mode dump the consumption down ~50% ?  MY limited experience tells me probably not,  but .72 to .55 is quite a difference in and of itself..

teek

also wonder

[chriswilmer]

More selling on Havelock... *sigh* why do I feel like someone knows something I don't... based on what is publicly known, it would seem like a terrible time to sell right now.

[minerpumpkin]

More selling on Havelock... *sigh* why do I feel like someone knows something I don't... based on what is publicly known, it would seem like a terrible time to sell right now.

I doubt it's inside knowledge. People are afraid due to the increased consumption news rumours, while friedcat remains silent. There may be a problem, or everything is totally fine.

[neilol-real]

More selling on Havelock... *sigh* why do I feel like someone knows something I don't... based on what is publicly known, it would seem like a terrible time to sell right now.

On very low volume

[der_meister]

terrible time to sell

=good time to buy

Seriously, I think the final consumption will be a bit lower and it's good enough. In this case are more important the time and the size of batches.

[hdbuck]

preparing myself for a last scooping up of AM shares on Havelock

[Mabsark]

More selling on Havelock... *sigh* why do I feel like someone knows something I don't... based on what is publicly known, it would seem like a terrible time to sell right now.

I doubt it's inside knowledge. People are afraid due to the increased consumption news rumours, while friedcat remains silent. There may be a problem, or everything is totally fine.

The actual power consumption was never going to be as low as the simulated results, it never is. It's still one hell of an efficient chip though, especially considering that it's 40nm and it's beating its 28nm competition. The chip's working fine.

[chriswilmer]

terrible time to sell

=good time to buy

Seriously, I think the final consumption will be a bit lower and it's good enough. In this case are more important the time and the size of batches.

I think the winning strategy at this point is to get the largest batch of chips made that is humanly possible and then move sky and earth to get the supply chain logistics worked out for all of the parts needed to build miners. It seems like every company out there is 100% selling out on PREORDERS, let alone actual miners.

[bitcoin.newsfeed]

I think the winning strategy at this point is to get the largest batch of chips made that is humanly possible and then move sky and earth to get the supply chain logistics worked out for all of the parts needed to build miners. It seems like every company out there is 100% selling out on PREORDERS, let alone actual miners.

Yes, but this is Friedcat, and he's acting very slow in every aspect    ...meanwhile the competition is very aggressive

[necro_nemesis]

The story has been the dividend and to rebuild the revenue stream to pay that dividend takes time. The way to benefit from that dividend is in retaining stock and do so for a lengthy period of time. All ASIC makers face the same challenges as AM did so looking-glass numbers are irrelevant to actual data. Efficiency is not limited to a measurement of power it's cost/GHash of the whole system end to end for the lifespan of the product and from that perspective I remain optimistic that this product remains competitive. One would also believe AM to have a leg up on the mining piece in having existing data centers that can be reutilized. What the stock price does in the short term is somewhat irrelevant if you're in for the dividend to find the value in holding it.

[bitsalame]

I would like to ask him what's his end game.
Is he gonna attempt to accomplish 20%+ of the hashing power as once promised or is he switching the objectives, abandoning self-mining altogether and becoming purely a manufacturer?

[minerpumpkin]

preparing myself for a last scooping up of AM shares on Havelock

Was that you at 15:25? 

[hdbuck]

preparing myself for a last scooping up of AM shares on Havelock

Was that you at 15:25?  

hehe ^^

c'mon FC! its now or never!

[mc_lovin]

I think the winning strategy at this point is to get the largest batch of chips made that is humanly possible and then move sky and earth to get the supply chain logistics worked out for all of the parts needed to build miners. It seems like every company out there is 100% selling out on PREORDERS, let alone actual miners.

Yes, but this is Friedcat, and he's acting very slow in every aspect    ...meanwhile the competition is very aggressive

I was really excited in November with a tapeout date of January 20th...  Hopefully this week he gives us some good news to chew on..