HashFast announces specs for new ASIC: 400GH/s

[defcon23]

few minutes ago from hashafast twitter account:

""" Things are moving quickly! We finally have the first Golden Nonce chip, we began testing it —sending the first current through the 28nm chip to see how it performs. So far, things are looking very good. HashFast co-founder and CTO Simon Barber to give us the play-by-play of the first tests. """"

"things are going quickly !"  .... he said??? ....  when they are supposed to shipped  their first batch 2 month ago?!!....


what's a pitty...

(sorry for my bad english.. )

[cedivad]

BTC refunds... Denied BTC refunds everywhere.

From what i understand, they don't have the boards yet, so they are using some strange setup to test the chips, but we will have to wait for the boards for better results (and once that they have the boards everything will go smooth).

On how many things i'm wrong? :|

[minternj]

Nice hashfast, tell those Aussies they aren't building the hadron collider. These asics are pretty goddamn dumb. Slap em in and ship em out.

[Gyrsur]

when they are supposed to shipped  their first batch 2 month ago?!!....

one month ago to be correct but I'm afraid they manage to send out the devices before christmas.

[iCEBREAKER]

https://hashfast.com/hello-world/

Things are moving quickly! We finally have the first Golden Nonce chip, we began testing it —sending the first current through the 28nm chip to see how it performs. So far, things are looking very good. HashFast co-founder and CTO Simon Barber to give us the play-by-play of the first tests.

    We brought up IOVdd first. Up to 1.8V — draws less than the minimum 10mA the bench supply reads.
    Next core. Something odd was observed here. Initially, we saw high current draw as we brought the voltage up, reaching 0.7V where it was drawing 2.5A. As the voltage increased a little over 0.7v the current suddenly dropped to 82mA. It stayed there up to Vddcore=0.8v. This odd pattern of high current at low core voltage was repeatable. RST held low. MODE and BYPASS were held high. PLLVdd pulled to Vdd. All other inputs seemed to float low, with a 1Mohm impedance probe.
    Next CLK was applied — first 37Hz. 4mA increase in Vdd core current. Took chip out of reset. No change. Next increase clock speed, this increases current. In a couple of steps, we got it to 6Mhz, where the core Vdd was drawing 548mA.
    After a few minutes, we turned things off. On retrying, we didn’t find the same odd, high current during ramp of core Vdd. Perhaps this was due to CLK floating then, and not since?

Calling it a night, having trouble getting the serial port levels shifted correctly. Will continue in the morning.

— Simon

Over the next few days, the HashFast engineering team will be working intensely as they subject the GN to more tests. As we’re able to get them to break and give us an update, we’ll let you know how it’s going.

Simon will also be tweeting on the HashFast account — if you’re interested in seeing the results as the they come, follow us @HashFast on Twitter.

Hooray!

There is light at the end of the tunnel.  All you concern trolls, disgruntled obsessives, and misc haters can now stop worrying.

PCBs are fast and easy to make.  They have these newfangled things called "robots" that can populate a PCB with components in seconds.

[Gyrsur]

https://hashfast.com/hello-world/

Things are moving quickly! We finally have the first Golden Nonce chip, we began testing it —sending the first current through the 28nm chip to see how it performs. So far, things are looking very good. HashFast co-founder and CTO Simon Barber to give us the play-by-play of the first tests.

    We brought up IOVdd first. Up to 1.8V — draws less than the minimum 10mA the bench supply reads.
    Next core. Something odd was observed here. Initially, we saw high current draw as we brought the voltage up, reaching 0.7V where it was drawing 2.5A. As the voltage increased a little over 0.7v the current suddenly dropped to 82mA. It stayed there up to Vddcore=0.8v. This odd pattern of high current at low core voltage was repeatable. RST held low. MODE and BYPASS were held high. PLLVdd pulled to Vdd. All other inputs seemed to float low, with a 1Mohm impedance probe.
    Next CLK was applied — first 37Hz. 4mA increase in Vdd core current. Took chip out of reset. No change. Next increase clock speed, this increases current. In a couple of steps, we got it to 6Mhz, where the core Vdd was drawing 548mA.
    After a few minutes, we turned things off. On retrying, we didn’t find the same odd, high current during ramp of core Vdd. Perhaps this was due to CLK floating then, and not since?

Calling it a night, having trouble getting the serial port levels shifted correctly. Will continue in the morning.

— Simon

Over the next few days, the HashFast engineering team will be working intensely as they subject the GN to more tests. As we’re able to get them to break and give us an update, we’ll let you know how it’s going.

Simon will also be tweeting on the HashFast account — if you’re interested in seeing the results as the they come, follow us @HashFast on Twitter.

Hooray!

There is light at the end of the tunnel.  All you concern trolls, disgruntled obsessives, and misc haters can now stop worrying.

PCBs are fast and easy to make.  They have these newfangled things called "robots" that can populate a PCB with components in seconds.

https://i.imgur.com/x95kGwG.jpg

Sorry, I can't enjoy my champagne. first I have to know if the will also deliver the 4 extra modules along with the initial shipment to make their customers (investors) happy and in a good mood to come back and by more from HF.

[amer]

https://hashfast.com/hello-world/

Things are moving quickly! We finally have the first Golden Nonce chip, we began testing it —sending the first current through the 28nm chip to see how it performs. So far, things are looking very good. HashFast co-founder and CTO Simon Barber to give us the play-by-play of the first tests.

    We brought up IOVdd first. Up to 1.8V — draws less than the minimum 10mA the bench supply reads.
    Next core. Something odd was observed here. Initially, we saw high current draw as we brought the voltage up, reaching 0.7V where it was drawing 2.5A. As the voltage increased a little over 0.7v the current suddenly dropped to 82mA. It stayed there up to Vddcore=0.8v. This odd pattern of high current at low core voltage was repeatable. RST held low. MODE and BYPASS were held high. PLLVdd pulled to Vdd. All other inputs seemed to float low, with a 1Mohm impedance probe.
    Next CLK was applied — first 37Hz. 4mA increase in Vdd core current. Took chip out of reset. No change. Next increase clock speed, this increases current. In a couple of steps, we got it to 6Mhz, where the core Vdd was drawing 548mA.
    After a few minutes, we turned things off. On retrying, we didn’t find the same odd, high current during ramp of core Vdd. Perhaps this was due to CLK floating then, and not since?

Calling it a night, having trouble getting the serial port levels shifted correctly. Will continue in the morning.

— Simon

Over the next few days, the HashFast engineering team will be working intensely as they subject the GN to more tests. As we’re able to get them to break and give us an update, we’ll let you know how it’s going.

Simon will also be tweeting on the HashFast account — if you’re interested in seeing the results as the they come, follow us @HashFast on Twitter.

Hooray!

There is light at the end of the tunnel.  All you concern trolls, disgruntled obsessives, and misc haters can now stop worrying.

PCBs are fast and easy to make.  They have these newfangled things called "robots" that can populate a PCB with components in seconds.

I feel bad for HF because icebreaker does more harm than good their reputation and they can't really silence him, because he's "trying to help".

[DeathAndTaxes]

Sorry, I can't enjoy my champagne. first I have to know if the will also deliver the 4 extra modules along with the initial shipment to make their customers (investors) happy and in a good mood to come back and by more from HF.

They have already stated they will not, "anticipated delivery of MPP is after 31 January".  Key words are after anticipated and after.

[Micky25]

PCBs are fast and easy to make.  They have these newfangled things called "robots" that can populate a PCB with components in seconds.

awww... all these robots are busy at the moment,



so this project has a very high probability to be not finished before xmas. So no delivery this year.

But hey, no problem if they deliver a great product, I have no doubt they work hard to do this and probably are on the right way.

..and (I nearly forgot to mention) if I get 2 full sierras instead of my batch 1 Baby-obsolete-Jet by mid January 2014

[Gyrsur]

Sorry, I can't enjoy my champagne. first I have to know if the will also deliver the 4 extra modules along with the initial shipment to make their customers (investors) happy and in a good mood to come back and by more from HF.

They have already stated they will not, "anticipated delivery of MPP is after 31 January".  Key words are after anticipated and after.

what should i say?  

EDIT: my plan in August was not to order from HF but then I got a very promising Email from them where they announced their super innovative MPP but I missed to ask a lawyer about that and so unfortunately I ordered. call it a good marketing strategie.

[iCEBREAKER]

Sorry, I can't enjoy my champagne. first I have to know if the will also deliver the 4 extra modules along with the initial shipment to make their customers (investors) happy and in a good mood to come back and by more from HF.

Reconsider your selfish, unreasonable expectations and you'll be able to enjoy some bubbly.   

HashFast has already paid for your extra MPP chips and they are being made at TSMC.  So relax and raise a toast.

You'll get them after the other orders have finished shipping.

Do you really expect HashFast to prioritize your MPP boards over other customers' initial shipments?

That's totally unreasonable.  Stop being so greedy and stop ignoring the plain text of the ToS ("after 31 Jan") to which you agreed.

Bulls can make money.  Bears can make money.  Pigs get slaughtered.

[gateway]

https://hashfast.com/hello-world/

Things are moving quickly! We finally have the first Golden Nonce chip, we began testing it —sending the first current through the 28nm chip to see how it performs. So far, things are looking very good. HashFast co-founder and CTO Simon Barber to give us the play-by-play of the first tests.

    We brought up IOVdd first. Up to 1.8V — draws less than the minimum 10mA the bench supply reads.
    Next core. Something odd was observed here. Initially, we saw high current draw as we brought the voltage up, reaching 0.7V where it was drawing 2.5A. As the voltage increased a little over 0.7v the current suddenly dropped to 82mA. It stayed there up to Vddcore=0.8v. This odd pattern of high current at low core voltage was repeatable. RST held low. MODE and BYPASS were held high. PLLVdd pulled to Vdd. All other inputs seemed to float low, with a 1Mohm impedance probe.
    Next CLK was applied — first 37Hz. 4mA increase in Vdd core current. Took chip out of reset. No change. Next increase clock speed, this increases current. In a couple of steps, we got it to 6Mhz, where the core Vdd was drawing 548mA.
    After a few minutes, we turned things off. On retrying, we didn’t find the same odd, high current during ramp of core Vdd. Perhaps this was due to CLK floating then, and not since?

Calling it a night, having trouble getting the serial port levels shifted correctly. Will continue in the morning.

— Simon

Over the next few days, the HashFast engineering team will be working intensely as they subject the GN to more tests. As we’re able to get them to break and give us an update, we’ll let you know how it’s going.

Simon will also be tweeting on the HashFast account — if you’re interested in seeing the results as the they come, follow us @HashFast on Twitter.

Hooray!

There is light at the end of the tunnel.  All you concern trolls, disgruntled obsessives, and misc haters can now stop worrying.

PCBs are fast and easy to make.  They have these newfangled things called "robots" that can populate a PCB with components in seconds.

+1 !!

[Gyrsur]

Sorry, I can't enjoy my champagne. first I have to know if the will also deliver the 4 extra modules along with the initial shipment to make their customers (investors) happy and in a good mood to come back and by more from HF.

That's totally unreasonable.  Stop being so greedy and stop ignoring the plain text of the ToS ("after 31 Jan") to which you agreed.

nothing to do with greedy. the convinced me to pay with BTC (what I never did before neither after) with a marketing trick which is not half promise as it's sound. paid nearly 50 BTC and I will make back 10 BTC with their initial shipment. and never make back 40 BTC with the 4 extra modules which they will ship in the middle of February. be fair and stay to facts, buddy!

[Gyrsur]

https://hashfast.com/hello-world/

+1 !!

btw what happened to their public order system on their website, webmaster?

[defcon23]

https://hashfast.com/hello-world/

Things are moving quickly! We finally have the first Golden Nonce chip, we began testing it —sending the first current through the 28nm chip to see how it performs. So far, things are looking very good. HashFast co-founder and CTO Simon Barber to give us the play-by-play of the first tests.

    We brought up IOVdd first. Up to 1.8V — draws less than the minimum 10mA the bench supply reads.
    Next core. Something odd was observed here. Initially, we saw high current draw as we brought the voltage up, reaching 0.7V where it was drawing 2.5A. As the voltage increased a little over 0.7v the current suddenly dropped to 82mA. It stayed there up to Vddcore=0.8v. This odd pattern of high current at low core voltage was repeatable. RST held low. MODE and BYPASS were held high. PLLVdd pulled to Vdd. All other inputs seemed to float low, with a 1Mohm impedance probe.
    Next CLK was applied — first 37Hz. 4mA increase in Vdd core current. Took chip out of reset. No change. Next increase clock speed, this increases current. In a couple of steps, we got it to 6Mhz, where the core Vdd was drawing 548mA.
    After a few minutes, we turned things off. On retrying, we didn’t find the same odd, high current during ramp of core Vdd. Perhaps this was due to CLK floating then, and not since?

Calling it a night, having trouble getting the serial port levels shifted correctly. Will continue in the morning.

— Simon

Over the next few days, the HashFast engineering team will be working intensely as they subject the GN to more tests. As we’re able to get them to break and give us an update, we’ll let you know how it’s going.

Simon will also be tweeting on the HashFast account — if you’re interested in seeing the results as the they come, follow us @HashFast on Twitter.

Hooray!

There is light at the end of the tunnel.  All you concern trolls, disgruntled obsessives, and misc haters can now stop worrying.

PCBs are fast and easy to make.  They have these newfangled things called "robots" that can populate a PCB with components in seconds.

https://i.imgur.com/x95kGwG.jpg

Sorry, I can't enjoy my champagne. first I have to know if the will also deliver the 4 extra modules along with the initial shipment to make their customers (investors) happy and in a good mood to come back and by more from HF.

+10 man !

[Visesrion]

Sorry, I can't enjoy my champagne. first I have to know if the will also deliver the 4 extra modules along with the initial shipment to make their customers (investors) happy and in a good mood to come back and by more from HF.

+10 man !

They wont do that. Is sugested that and the answer was a big no long time ago. The ugrade chips are coming after the 2nd batch and the MPP chips after 31st of January. They could have done the things well time ago by switching the BJs into the sierras to compensate, but as their schedule was to deliver by late December they are doing the things really well covering their back with legal tos.

By the way, did I miss the part of how much hashing power is the chip mining or they just wanted to post some technical update that says "we are still testing the chip" and we have no pcbs 

[iCEBREAKER]

https://hashfast.com/hello-world/

Things are moving quickly! We finally have the first Golden Nonce chip, we began testing it —sending the first current through the 28nm chip to see how it performs. So far, things are looking very good. HashFast co-founder and CTO Simon Barber to give us the play-by-play of the first tests.

Over the next few days, the HashFast engineering team will be working intensely as they subject the GN to more tests. As we’re able to get them to break and give us an update, we’ll let you know how it’s going.

Simon will also be tweeting on the HashFast account — if you’re interested in seeing the results as the they come, follow us @HashFast on Twitter.

Hooray!

There is light at the end of the tunnel.  All you concern trolls, disgruntled obsessives, and misc haters can now stop worrying.

PCBs are fast and easy to make.  They have these newfangled things called "robots" that can populate a PCB with components in seconds.

https://i.imgur.com/x95kGwG.jpg

Sorry, I can't enjoy my champagne. first I have to know if the will also deliver the 4 extra modules along with the initial shipment to make their customers (investors) happy and in a good mood to come back and by more from HF.

+10 man !

You both know the number of chips initially available is limited and all of them are spoken for.

And yet, you both persist in your delusion that HashFast can magically create 400% more chips than actually exist right now.

Such wishful thinking is childish.  Your pony will arrive soon.  You don't need to complain that anything less than 5 ponies is unacceptable.

   How do you expect HashFast to provide everyone with 400% more chips than actually exit?   

Neither of you have a good answer to that question, but that won't stop you from pouting like spoiled, entitled brats.   


Visesrion, it's obvious you don't know enough about EE to follow (IE comprehend) HashFast's technical updates.

Let me translate for you: The PCB is done, but the chip is still being brought up to full speed in the test bed.

[MrTeal]

Sorry, I can't enjoy my champagne. first I have to know if the will also deliver the 4 extra modules along with the initial shipment to make their customers (investors) happy and in a good mood to come back and by more from HF.

They have already stated they will not, "anticipated delivery of MPP is after 31 January".  Key words are after anticipated and after.

Actually, in that ambiguous and useless statement the word anticipated is actually a good thing. Although I'm sure they didn't mean it this way, it still leaves open the possibility of before Jan 31.
The more accurate wording they probably should have used would be the wonderfully open ended "delivery of MPP is after 31 January"

[Visesrion]

Sorry Icebreaker, but long time ago HF claimed that they have already ordered and paid the 2nd batch and MPP chips, so now 2 months after that I think they got plenty of time to get the chips.
Apart from that, why they refuse from day 1 to change (by paying the difference at that time) a BJ order into a Sierra one ?
Why the hell you offer a BJ range product if you are thinking on offering a product with x3 hashing power one month later at the same price ? Now with the delay that gap will be nothing

And about the chip or pcb, until I see the chip mining on a video I do not care what they write down. Sorry but if I were at the front of a company with so much money invested by the customers, I will take the mobile or any electronic gadget around and make lot of videos and pictures even it was a tiny update, specially if you are already 1 month late and making them lose money each day

[DPoS]

lulz

"After a few minutes, we turned things off. On retrying, we didn’t find the same odd, high current during ramp of core Vdd. Perhaps this was due to CLK floating then, and not since?"


are they REALLY going to ship this into your hands this month?


https://hashfast.com/hello-world/