Oh we could send AM 10 million so they can make us some miners...who wants to join in? 
They need to trust AM first...
Topic: [ANN] Sfards:SF100, the first 28nm Dual-Mode Miner is accepting pre-order now - page 71. (Read 129801 times)
April 02, 2015, 12:12:06 PM
They need to trust AM first...
April 02, 2015, 12:07:03 PM
this looks like a DOA item.
It will do poorly if it does not show up like April 15th. I wish it would but I have very little hope for this item.
It will do poorly if it does not show up like April 15th. I wish it would but I have very little hope for this item.
The main problem is they have given zero specs , even if you were planning a may launch you would have preliminary specs and a data sheet.
I say again this is probably vaporware for this qurter, maybe they will release something 4th q who knows.
I think this inital post was a PR stunt to make people stop buying other hardware (s5s) and hold out for something better/new.
s5's are sold out the only new gear in stock from asic builders is the avalon 4.1 and it is costly
Oh we could send AM 10 million so they can make us some miners...who wants to join in?
April 02, 2015, 12:03:50 PM
this looks like a DOA item.
It will do poorly if it does not show up like April 15th. I wish it would but I have very little hope for this item.
It will do poorly if it does not show up like April 15th. I wish it would but I have very little hope for this item.
The main problem is they have given zero specs , even if you were planning a may launch you would have preliminary specs and a data sheet.
I say again this is probably vaporware for this qurter, maybe they will release something 4th q who knows.
I think this inital post was a PR stunt to make people stop buying other hardware (s5s) and hold out for something better/new.
s5's are sold out the only new gear in stock from asic builders is the avalon 4.1 and it is costly
April 02, 2015, 11:58:08 AM
At the end of the day, all they have done is tell the forum members that they have taped out their chip. They are not asking for your money nor are they promising something that they may not be able to deliver, sounds like good engineering practice to me.
They have done absolutely nothing wrong or underhand except share information on their plans. Whether you like their plans or not, at least they have given an indication of where they are planning to go. Personally, I don't think what they are doing is particularly cost effective but I do hope that they make it and that it works.
What they have done is encourage some debate, which is never a bad thing. I'm pretty sure that if they want to sell you something then they'll give out all the information you need.
They have done absolutely nothing wrong or underhand except share information on their plans. Whether you like their plans or not, at least they have given an indication of where they are planning to go. Personally, I don't think what they are doing is particularly cost effective but I do hope that they make it and that it works.
What they have done is encourage some debate, which is never a bad thing. I'm pretty sure that if they want to sell you something then they'll give out all the information you need.
April 02, 2015, 10:55:46 AM
this looks like a DOA item.
It will do poorly if it does not show up like April 15th. I wish it would but I have very little hope for this item.
It will do poorly if it does not show up like April 15th. I wish it would but I have very little hope for this item.
The main problem is they have given zero specs , even if you were planning a may launch you would have preliminary specs and a data sheet.
I say again this is probably vaporware for this qurter, maybe they will release something 4th q who knows.
I think this inital post was a PR stunt to make people stop buying other hardware (s5s) and hold out for something better/new.
April 01, 2015, 01:38:25 PM
Updates just make people stay interested. Going quiet really just allows for lots of speculation.
Some people will speculate one way or the other. While I, too, would prefer more frequent updates from developers, I'm just not much of a squeaky wheel type. This wasn't really speculation, though, just a question that could have been self-answered with a few minutes of time. 
April 01, 2015, 01:13:55 PM
still nothing?
We are expecting to release our first miner in April 2015.
It's only been April for part of 1 day
Only thing is most of us I think expected to see updates or something. They started off great showing off offices making it look like a good operation. Then it went dark. Updates just make people stay interested. Going quiet really just allows for lots of speculation.
Guys, they only taped out a couple weeks ago. They likely don't have any chips back from the foundry/packager, give it a rest.
Birdman, what are you doing out of the KnC threads? Also, you should know better by now.
April 01, 2015, 12:35:04 PM
still nothing?
We are expecting to release our first miner in April 2015.
It's only been April for part of 1 day
Only thing is most of us I think expected to see updates or something. They started off great showing off offices making it look like a good operation. Then it went dark. Updates just make people stay interested. Going quiet really just allows for lots of speculation.
April 01, 2015, 12:30:45 PM
still nothing?
We are expecting to release our first miner in April 2015.
It's only been April for part of 1 day
April 01, 2015, 12:08:25 PM
still nothing? Nothing on their other channels either... no pics of the elusive test product.
SVARDS...
If you had nothing to actually show us, why bother posting the announcement?
SVARDS...
If you had nothing to actually show us, why bother posting the announcement?
April 01, 2015, 10:43:54 AM
this looks like a DOA item.
It will do poorly if it does not show up like April 15th. I wish it would but I have very little hope for this item.
It will do poorly if it does not show up like April 15th. I wish it would but I have very little hope for this item.
April 01, 2015, 05:37:31 AM
tl;drt
No, <2.0, no and no.( Based on the staggering 7 pages of thread, site, twitter, weibo and other public sources. But mostly the staggering 7 pages of thread. )
April 01, 2015, 01:47:42 AM
Does anyone know the total scrypt MHs speed of these, and also the wattage draw. Also what is the total sha-256 speed? Any pictures anywhere?
Vegas
Vegas
March 30, 2015, 04:22:21 PM
How can you determine that without knowing the specifications of scypt mode only, sha256 mode only and combined?
Both are fixed algorithms. If you are familiar with silicon design you can accurately determine how many times each gate/transistor in an engine stage switches in one arithmetic cycle. Knowing roughly how much power that consumes based on the characteristics of the process node - and looks from the initial power figures quoted that it's still a good 'ol fashioned 28nm straight gate based design - then it's very easy to get a good approximation of the power consumed. It's also very easy to calculate the power of the embedded DRAM blocks. Try it and see.
Good info there ..... certainly above my paygrade and have no intention of pretending otherwise, unlike the usual suspects, but interesting nevertheless.
You don't give yourself enough credit, the best technologists are the ones that realise they don't have a piece of knowledge instead of bullshitting that they are an expert. Then they go and learn about it. Silicon design and power estimation are defined by fairly simple mathematics and physics, and for anyone with an inquiring mind, a desire to learn and a bit of patience, it's easy to learn the basics.
Particle physics? That's a whole different ballgame!
March 30, 2015, 09:36:23 AM
How can you determine that without knowing the specifications of scypt mode only, sha256 mode only and combined?
Both are fixed algorithms. If you are familiar with silicon design you can accurately determine how many times each gate/transistor in an engine stage switches in one arithmetic cycle. Knowing roughly how much power that consumes based on the characteristics of the process node - and looks from the initial power figures quoted that it's still a good 'ol fashioned 28nm straight gate based design - then it's very easy to get a good approximation of the power consumed. It's also very easy to calculate the power of the embedded DRAM blocks. Try it and see.
Good info there ..... certainly above my paygrade and have no intention of pretending otherwise, unlike the usual suspects, but interesting nevertheless.
March 30, 2015, 08:49:18 AM
Dual mining technologies on a single chip is not a good idea, there are almost no common parts between the two hashing engines, the silicon for the scrypt engine is a waste of space unless LTC price goes up considerably. Also, the scrypt engine will be relatively power hungry.
...As for SFARDS I wish them good luck with their chip, especially with routing it.
i agree. There's very little real benefit to this compared to simply having seperate sha256 and scrypt mining devices. pretty sure whatever gridseeds are still in use are 100% dedicated to BTC at this point....As for SFARDS I wish them good luck with their chip, especially with routing it.
I believe the sentiment this time around is that it allows one set of hardware, one set of investments and one setup to swap algorithms as profitability shifts. With a flatter difficulty, each generation *should* be installed longer. Last time it didn't really work because almost immediately the sha256 side wasn't competitive / wasn't worth the cooling hassle compared to the profitability of the scrypt side.
Its hard for us to comment this time around without knowing the exact specs of the 3 modes and the arrangement of the hardware.
Us? Do you think you're Queen Victoria?
Seriously, do the maths and you'll see exactly why this configuration is a very bad idea. If you can make any such arrangement work, then there's something seriously wrong in your calculations.
If you need any help, just post here.
How can you determine that without knowing the specifications of scypt mode only, sha256 mode only and combined?
Both are fixed algorithms. If you are familiar with silicon design you can accurately determine how many times each gate/transistor in an engine stage switches in one arithmetic cycle. Knowing roughly how much power that consumes based on the characteristics of the process node - and looks from the initial power figures quoted that it's still a good 'ol fashioned 28nm straight gate based design - then it's very easy to get a good approximation of the power consumed. It's also very easy to calculate the power of the embedded DRAM blocks. Try it and see.
March 30, 2015, 02:04:30 AM
Dual mining technologies on a single chip is not a good idea, there are almost no common parts between the two hashing engines, the silicon for the scrypt engine is a waste of space unless LTC price goes up considerably. Also, the scrypt engine will be relatively power hungry.
...As for SFARDS I wish them good luck with their chip, especially with routing it.
i agree. There's very little real benefit to this compared to simply having seperate sha256 and scrypt mining devices. pretty sure whatever gridseeds are still in use are 100% dedicated to BTC at this point....As for SFARDS I wish them good luck with their chip, especially with routing it.
I believe the sentiment this time around is that it allows one set of hardware, one set of investments and one setup to swap algorithms as profitability shifts. With a flatter difficulty, each generation *should* be installed longer. Last time it didn't really work because almost immediately the sha256 side wasn't competitive / wasn't worth the cooling hassle compared to the profitability of the scrypt side.
Its hard for us to comment this time around without knowing the exact specs of the 3 modes and the arrangement of the hardware.
Us? Do you think you're Queen Victoria?
Seriously, do the maths and you'll see exactly why this configuration is a very bad idea. If you can make any such arrangement work, then there's something seriously wrong in your calculations.
If you need any help, just post here.
How can you determine that without knowing the specifications of scypt mode only, sha256 mode only and combined?
March 30, 2015, 01:39:19 AM
Dual mining technologies on a single chip is not a good idea, there are almost no common parts between the two hashing engines, the silicon for the scrypt engine is a waste of space unless LTC price goes up considerably. Also, the scrypt engine will be relatively power hungry.
...As for SFARDS I wish them good luck with their chip, especially with routing it.
i agree. There's very little real benefit to this compared to simply having seperate sha256 and scrypt mining devices. pretty sure whatever gridseeds are still in use are 100% dedicated to BTC at this point....As for SFARDS I wish them good luck with their chip, especially with routing it.
I believe the sentiment this time around is that it allows one set of hardware, one set of investments and one setup to swap algorithms as profitability shifts. With a flatter difficulty, each generation *should* be installed longer. Last time it didn't really work because almost immediately the sha256 side wasn't competitive / wasn't worth the cooling hassle compared to the profitability of the scrypt side.
Its hard for us to comment this time around without knowing the exact specs of the 3 modes and the arrangement of the hardware.
Us? Do you think you're Queen Victoria?
Seriously, do the maths and you'll see exactly why this configuration is a very bad idea. If you can make any such arrangement work, then there's something seriously wrong in your calculations.
If you need any help, just post here.
March 29, 2015, 10:14:01 PM
Question: Who at Sfards should I submit my invoice to for the hours I put in trying to discredit this outfit to no avail? 
Consider this post an endorsement from Bruno Kucinskas.
Later, guys. I'm done here, but PM me if anything changes.
Consider this post an endorsement from Bruno Kucinskas.
Later, guys. I'm done here, but PM me if anything changes.
March 29, 2015, 04:05:50 PM
Dual mining technologies on a single chip is not a good idea, there are almost no common parts between the two hashing engines, the silicon for the scrypt engine is a waste of space unless LTC price goes up considerably. Also, the scrypt engine will be relatively power hungry.
...As for SFARDS I wish them good luck with their chip, especially with routing it.
i agree. There's very little real benefit to this compared to simply having seperate sha256 and scrypt mining devices. pretty sure whatever gridseeds are still in use are 100% dedicated to BTC at this point....As for SFARDS I wish them good luck with their chip, especially with routing it.
I believe the sentiment this time around is that it allows one set of hardware, one set of investments and one setup to swap algorithms as profitability shifts. With a flatter difficulty, each generation *should* be installed longer. Last time it didn't really work because almost immediately the sha256 side wasn't competitive / wasn't worth the cooling hassle compared to the profitability of the scrypt side.
Its hard for us to comment this time around without knowing the exact specs of the 3 modes and the arrangement of the hardware.
Jump to: