Pages:
Author

Topic: HashFast launches sales of the Baby Jet - page 50. (Read 119626 times)

donator
Activity: 1218
Merit: 1079
Gerald Davis
September 02, 2013, 05:13:45 PM
Looking forward to seeing how quickly the GN chips go through spins.  To give everyone else here a sense of time, Intel's 45 nm Penryn 5500s did their first tape-out and wafer generation in the beginning of January 2007.  The initial chips were able to boot Windows (almost unheard of luck for a first tapeout), but the Q9550 chips had many spins and revisions and were not released until November 2007.  However, TSMC has been making solid progress on 28 nm yield for the past while, and I'd expect it'd be through spins in less than a year since it's no longer a bleeding edge process.

This is like comparing a rowboat to a super tanker in terms of complexity.  Bitcoin ASICs are little more than parallel SHA-2 calculators.  There is essentially no logic in the chips, all the smarts are in the general purpose microprocessor running the mining software.
donator
Activity: 1218
Merit: 1079
Gerald Davis
September 02, 2013, 05:11:21 PM
At 324 mm^2, HashFast will have a huge problem if they don't meet thermal specs at 250 W as they'll be next to impossible to cool or will not meet specified hash rates.  This was the same problem AMD hit early on with their Bulldozer processors, and the Bulldozer series of processors were slightly smaller than this.  I would say that anyone who is claiming a 324 mm^2 chip with a target TDP of 250 W can be easily overclocked 25% should be considered suspicious.

Not really.  It is all relative.  Say for the sake of the argument they can't exceed 250W (not sure what is true with water cooling and large die size).  Lets also say 400 GH/s is over 250W.  There is some clock speed and voltage combination which produces X GH/s stable at 250W TDP.  Now lets take that a step further.  What makes you think 250W is the target?  Maybe they are targeting 450 GH/s @ 200W and the rest is just to hedge their bets in the final silicon comes in hotter or slower than expected. 

Worse case scenario they end up needing to use multiple chips and/or offer customers a partial refund (which may be cheaper than providing multiple units if they miss by less than a certain percentage).  With the markup on BOM being >500% that simply means less profit, not a loss of profit.


legendary
Activity: 2156
Merit: 1072
Crypto is the separation of Power and State.
September 02, 2013, 12:13:31 AM
I would then assume you are just as skeptical of the other 28nm projects out there as well (KnC, Cointerra, etc.)?

I think comparing a general purpose CPU and a bag of logic that does a bunch of right rotates, AND, XORs, and appends is a bit disingenuous.
http://en.wikipedia.org/wiki/SHA-2#Pseudocode

The fact that Bitfury was able to produce a 55nm that hashes must be a freakin' miracle.


I'm not really skeptical, as almost 1/3 of what TSMC ships out right now is on 28 nm.  Which is, I believe, about the same quantity as on 55/65 nm at the current time.  But I would be surprised if the ease and cost of developing at 28 nm were the same as for 55 nm.  You should also keep in mind that Penryn was a die shrink of the Core 2 microarchitecture, so there was really not a lot new going on there (probably why the first tape-out was so dramatically successful).



What may really kill these chips in terms of development time is that they've selected pretty big die areas for them.  At 324 mm^2, HashFast will have a huge problem if they don't meet thermal specs at 250 W as they'll be next to impossible to cool or will not meet specified hash rates.  This was the same problem AMD hit early on with their Bulldozer processors, and the Bulldozer series of processors were slightly smaller than this.  I would say that anyone who is claiming a 324 mm^2 chip with a target TDP of 250 W can be easily overclocked 25% should be considered suspicious.

hi TT, always nice to see you around!

That chart is from 2009, back in the dark ages Before Satoshi.  28nm silicon is cheap and easy now.

Inherently vastly redundant much like a GPU, the GN is much less complex (logically and physically) than modern CPUs like Bulldozer and Core2.   Such redundancy vastly increases margin for error as faulty registers/hashing units are simply disabled.  Sync issues are also trivial due to the perfectly parallel nature of hashing and low clock speeds.

One advanced feature confirmed for GN is on-die thermal management.  While not exactly Powertune (sorry that IP isn't for sale Tongue) the function is equivalent - chip gets hot, chip slows down until it cools off.

IDK exactly how huge/hungry my 7970s are, but they definitely run well over 125% with stock cooling and Powertune +20.

Given their sweet performance-enthusiast-sourced aftermarket liquid-cooling, HashFasts will crush.  Cool
full member
Activity: 238
Merit: 100
September 01, 2013, 09:54:57 PM
I would then assume you are just as skeptical of the other 28nm projects out there as well (KnC, Cointerra, etc.)?

I think comparing a general purpose CPU and a bag of logic that does a bunch of right rotates, AND, XORs, and appends is a bit disingenuous.
http://en.wikipedia.org/wiki/SHA-2#Pseudocode

The fact that Bitfury was able to produce a 55nm that hashes must be a freakin' miracle.


A general purpose CPU can be quite simple. The 6502 (the one that powered the Apple-2, and the NES) had about 3,000 transistors  However, something like a modern X86 chip is a beast, with hundreds of millions of transistors that all need to work perfectly and in sync.
legendary
Activity: 1484
Merit: 1005
September 01, 2013, 09:40:56 PM
I would then assume you are just as skeptical of the other 28nm projects out there as well (KnC, Cointerra, etc.)?

I think comparing a general purpose CPU and a bag of logic that does a bunch of right rotates, AND, XORs, and appends is a bit disingenuous.
http://en.wikipedia.org/wiki/SHA-2#Pseudocode

The fact that Bitfury was able to produce a 55nm that hashes must be a freakin' miracle.


I'm not really skeptical, as almost 1/3 of what TSMC ships out right now is on 28 nm.  Which is, I believe, about the same quantity as on 55/65 nm at the current time.  But I would be surprised if the ease and cost of developing at 28 nm were the same as for 55 nm.  You should also keep in mind that Penryn was a die shrink of the Core 2 microarchitecture, so there was really not a lot new going on there (probably why the first tape-out was so dramatically successful).



What may really kill these chips in terms of development time is that they've selected pretty big die areas for them.  At 324 mm^2, HashFast will have a huge problem if they don't meet thermal specs at 250 W as they'll be next to impossible to cool or will not meet specified hash rates.  This was the same problem AMD hit early on with their Bulldozer processors, and the Bulldozer series of processors were slightly smaller than this.  I would say that anyone who is claiming a 324 mm^2 chip with a target TDP of 250 W can be easily overclocked 25% should be considered suspicious.
sr. member
Activity: 490
Merit: 255
September 01, 2013, 09:22:21 PM
I would then assume you are just as skeptical of the other 28nm projects out there as well (KnC, Cointerra, etc.)?

I think comparing a general purpose CPU and a bag of logic that does a bunch of right rotates, AND, XORs, and appends is a bit disingenuous.
http://en.wikipedia.org/wiki/SHA-2#Pseudocode

The fact that Bitfury was able to produce a 55nm that hashes must be a freakin' miracle.
legendary
Activity: 1484
Merit: 1005
September 01, 2013, 09:04:56 PM
Looking forward to seeing how quickly the GN chips go through spins.  To give everyone else here a sense of time, Intel's 45 nm Penryn 5500s did their first tape-out and wafer generation in the beginning of January 2007.  The initial chips were able to boot Windows (almost unheard of luck for a first tapeout), but the Q9550 chips had many spins and revisions and were not released until November 2007.  However, TSMC has been making solid progress on 28 nm yield for the past while, and I'd expect it'd be through spins in less than a year since it's no longer a bleeding edge process.
sr. member
Activity: 490
Merit: 255
September 01, 2013, 05:42:09 PM
What are the die size, package type, core voltage, transistor count, operating frequency, and projected TDP of the chip?  

Hey Bogie -

Looks like Amy from HF answered a couple of these questions:
https://bitcointalksearch.org/topic/m.2975145

Die Size: 18x18mm
TDP: 250W

I don't think the other pieces have been revealed but I imagine they will be required for the independent board program, which also doesn't have a firm date. 

Still not sure how to feel about the independent board program.  It means that they are planning on selling raw chips at some point after they have produced their own miners in unknown quantities.  It starts to remind me of my Avalon experience which isn't one I want to be reminded of or repeated.  Hopefully the MPP whether legal or not will be a benefit to the HF purchasers.
hero member
Activity: 574
Merit: 501
September 01, 2013, 05:24:31 PM
And while you're at it, how about telling us the underwear and shoe sizes of each of the corporate officers...
legendary
Activity: 966
Merit: 1000
September 01, 2013, 04:08:16 PM
What are the die size, package type, core voltage, transistor count, operating frequency, and projected TDP of the chip?  
member
Activity: 74
Merit: 10
September 01, 2013, 03:52:24 PM
It was clearly copy-pasted from the link, which was posted on the 29th  Undecided
hero member
Activity: 546
Merit: 500
Owner, Minersource.net
September 01, 2013, 12:43:38 PM
HashFast’s Golden Nonce GN ASIC successfully taped-out yesterday, Wednesday the 28th, and has been released for 28nm fabrication to a well-known, leading-edge foundry. https://hashfast.com/tapeout/

More details will follow in next week’s joint press release.

-The HashFast Team

Yesterday was Saturday Aug 31st.
Yeah... These are why people have doubts. If you cant even get the post date on a forum correct...
legendary
Activity: 2492
Merit: 1473
LEALANA Bitcoin Grim Reaper
September 01, 2013, 12:41:28 PM
HashFast’s Golden Nonce GN ASIC successfully taped-out yesterday, Wednesday the 28th, and has been released for 28nm fabrication to a well-known, leading-edge foundry. https://hashfast.com/tapeout/

More details will follow in next week’s joint press release.

-The HashFast Team

Yesterday was Saturday Aug 31st.
full member
Activity: 154
Merit: 100
September 01, 2013, 12:34:27 PM
HashFast’s Golden Nonce GN ASIC successfully taped-out yesterday, Wednesday the 28th, and has been released for 28nm fabrication to a well-known, leading-edge foundry. https://hashfast.com/tapeout/

More details will follow in next week’s joint press release.

-The HashFast Team
full member
Activity: 210
Merit: 100
August 21, 2013, 06:25:34 AM
"The results of our mock tapeout are due back any minute now and so far it all looks good." ?

So why not wait an hour, get the results and publish them?

If you reread the quote, you'll learn that the "results of [the] mock tapeout are due back any minute..."  Mock tapeouts are noteworthy if you're planning to build mock ASICs.  Unlike Mock Turtle Soup, mock ASICs are neither cheap nor real Sad
sr. member
Activity: 1176
Merit: 265
August 21, 2013, 03:36:21 AM
"The results of our mock tapeout are due back any minute now and so far it all looks good." ?

So why not wait an hour, get the results and publish them?
legendary
Activity: 966
Merit: 1004
CryptoTalk.Org - Get Paid for every Post!
August 20, 2013, 10:23:11 PM
My guess is that they won't commit to an OC speed until they have the actually chip in hand and characterized.  I bet their simulations show that 500GH/s is very doable, but simulations and reality are not the same... Just ask bitfury.  He was expecting his chip to run 5GH with 10GH overclock.  Instead he got 2GH with 3GH overclock.


Sims show 540!

https://bitcointalksearch.org/topic/m.2976695


& this post


https://bitcointalksearch.org/topic/m.2976622



sr. member
Activity: 490
Merit: 255
August 20, 2013, 09:41:11 PM
My guess is that they won't commit to an OC speed until they have the actually chip in hand and characterized.  I bet their simulations show that 500GH/s is very doable, but simulations and reality are not the same... Just ask bitfury.  He was expecting his chip to run 5GH with 10GH overclock.  Instead he got 2GH with 3GH overclock.
newbie
Activity: 56
Merit: 0
August 20, 2013, 09:37:17 PM
Hashfast's website says that the Babyjet is "designed" to be overclocked up to 500 gh/s, but they cannot guarantee that it could reach 500 gh/s could they? I simply cannot buy one unless it is for sure that it can reach that speed. If it does not, then it doesn't have a good ROI for me.

Now, I emailed them, and they told me they cannot release information about the potential overclock speed. I call that a crock because it is 5.7k usa cash and I want to know what I am buying if I buy it. They may say they do not need my business, but I represent all these other buyers when I say I want to know the overclock speed.

Now, I also find 100+ gh/s more overclock speed as pretty high amount. If it were 25-50, then it might be more believable to me.

Again, I represent others when I say I want to know the overclock speed
full member
Activity: 210
Merit: 100
August 20, 2013, 08:59:26 PM
We've posted a new update on our blog:
https://hashfast.com/countdown-to-tapeout/

-John

That's way awesome, John! Now go post in another thread that you just posted here!1!1!!

Pages:
Jump to: