Reclaim3r (OP)
|
|
July 15, 2014, 09:13:05 PM |
|
Just an intriguing idea I was thinking about. Atmel has a chip known as the ATSHA204 that appears to be used for simple password encryption and hardware security purposes. It DOES however, posses the ability to perform SHA256 hashes and I was wondering if it could be put to use in the world of SHA256 mining. Datasheet is here: http://www.atmel.com/Images/Atmel-8885-CryptoAuth-ATSHA204A-Datasheet.pdfAlong with a cute picture ( it comes in two formats, single wire and I 2C )[/sup] ( Single Wire ) ( I 2C)
|
|
|
|
dogie
Legendary
Offline
Activity: 1666
Merit: 1185
dogiecoin.com
|
|
July 15, 2014, 09:27:23 PM |
|
super tldr, no.
|
|
|
|
Swimmer63
Legendary
Offline
Activity: 1593
Merit: 1004
|
|
July 15, 2014, 09:58:17 PM |
|
super tldr, no.
Yes. How much does it cost?
|
|
|
|
dropt
Legendary
Offline
Activity: 1512
Merit: 1000
|
|
July 15, 2014, 10:08:09 PM |
|
super tldr, no.
Yes. How much does it cost? $1.32/ea in low Qty. Buy >1000 and it drops to $0.57/unit.
|
|
|
|
RoadStress
Legendary
Offline
Activity: 1904
Merit: 1007
|
|
July 15, 2014, 10:12:33 PM |
|
super tldr, no.
Yes. How much does it cost? $1.32/ea in low Qty. Buy >1000 and it drops to $0.57/unit. What's the hashrate that it could achieve? I'm asking just to see if this chip could've smashed the FPGAs.
|
|
|
|
Reclaim3r (OP)
|
|
July 15, 2014, 10:27:05 PM |
|
super tldr, no.
Super tldr? I'm sorry, I'm still somewhat of a newbie
|
|
|
|
|
Reclaim3r (OP)
|
|
July 15, 2014, 10:37:51 PM |
|
Thanks for clearing that up. Guess their isn't a definitive reply to this yet
|
|
|
|
-ck
Legendary
Offline
Activity: 4284
Merit: 1645
Ruu \o/
|
|
July 15, 2014, 10:43:31 PM |
|
The response was definitive enough. Take it from the cgminer maintainer if it wasn't definitive enough... It DOES however, posses the ability to perform SHA256 hashes and I was wondering if it could be put to use in the world of SHA256 mining.
Not remotely in any useful amount.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
Reclaim3r (OP)
|
|
July 15, 2014, 10:45:44 PM |
|
The response was definitive enough. Take it from the cgminer maintainer if it wasn't definitive enough... It DOES however, posses the ability to perform SHA256 hashes and I was wondering if it could be put to use in the world of SHA256 mining.
Not remotely in any useful amount. could we at least get a hashrate? Pretty please?
|
|
|
|
dropt
Legendary
Offline
Activity: 1512
Merit: 1000
|
|
July 15, 2014, 10:45:57 PM |
|
What's the hashrate that it could achieve? I'm asking just to see if this chip could've smashed the FPGAs.
Just by a quick skim of the datasheet, low as in the order of less than 40 Hashes per second. So to answer your question: Not a chance.
|
|
|
|
MrTeal
Legendary
Offline
Activity: 1274
Merit: 1004
|
|
July 15, 2014, 10:47:55 PM |
|
super tldr, no.
Yes. How much does it cost? $1.32/ea in low Qty. Buy >1000 and it drops to $0.57/unit. What's the hashrate that it could achieve? I'm asking just to see if this chip could've smashed the FPGAs. According to the datasheet, without any communications overhead it takes typically 11ms to do one hash of SHA256 (max 22ms), so for the BTC double SHA-256 you'd be looking at about 45 hashes per second, again not taking into account the overhead of loading the data into the chip. I'd guess more like 40 H/s or so.
|
|
|
|
Reclaim3r (OP)
|
|
July 15, 2014, 10:48:24 PM |
|
What's the hashrate that it could achieve? I'm asking just to see if this chip could've smashed the FPGAs.
Just by a quick skim of the datasheet, low as in the order of less than 40 Hashes per second. So to answer your question: Not a chance. Would it be worth the cost and power?
|
|
|
|
-ck
Legendary
Offline
Activity: 4284
Merit: 1645
Ruu \o/
|
|
July 15, 2014, 10:53:31 PM |
|
Wow that's persistence for you. Current ASICs are about $1 per billion hashes per second with all their attached hardware and systems. You work it out now for your chip.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
dropt
Legendary
Offline
Activity: 1512
Merit: 1000
|
|
July 15, 2014, 10:54:32 PM |
|
What's the hashrate that it could achieve? I'm asking just to see if this chip could've smashed the FPGAs.
Just by a quick skim of the datasheet, low as in the order of less than 40 Hashes per second. So to answer your question: Not a chance. Would it be worth the cost and power? Assuming MrTeal and I are reasonably accurate with our rough estimation of ~40H/S... A Gen1 AM Asic did ~333MH/S, which is 333000000H/S. So you'd need 16,650,000 of them to do the work of one (1) Gen1 AM ASIC. That'd set you back like 9 Million dollars, require ~8,325A at 5.5V, or ~45.8kW. Soooo.. No. Edit: That assumes you use two chips for the double SHA256 round.
|
|
|
|
Reclaim3r (OP)
|
|
July 15, 2014, 11:00:24 PM |
|
What's the hashrate that it could achieve? I'm asking just to see if this chip could've smashed the FPGAs.
Just by a quick skim of the datasheet, low as in the order of less than 40 Hashes per second. So to answer your question: Not a chance. Would it be worth the cost and power? Assuming MrTeal and I are reasonably accurate with our rough estimation of ~40H/S... A Gen1 AM Asic did ~333MH/S, which is 333000000H/S. So you'd need 16,650,000 of them to do the work of one (1) Gen1 AM ASIC. That'd set you back like 9 Million dollars, require ~8,325A at 5.5V, or ~45.8kW. Soooo.. No. Edit: That assumes you use two chips for the double SHA256 round. Feels...okay. I was hoping I could waste my summer time buy buying in bulk and soldering them to a massive PCB. Maybe I'll find another SHA256 non-mining ASIC
|
|
|
|
Reclaim3r (OP)
|
|
July 15, 2014, 11:01:17 PM |
|
Wow that's persistence for you. Current ASICs are about $1 per billion hashes per second with all their attached hardware and systems. You work it out now for your chip.
Ayup I get especially persistent with Bitcoin or cryptocurrency related info and what not. Got my answer though :-D
|
|
|
|
Lorenzo
|
|
July 16, 2014, 09:36:27 AM |
|
Interesting find OP. Isn't this chip technically speaking, an ASIC? I don't think it's an FPGA and it's definitely not a CPU or GPU. And if is an ASIC, then why can it only reach 40 hashes per second? I thought ASICs were supposed to be much faster than this? Also I found another thread about this chip where someone came up with a figure of 710 hashes per second (still very low however): http://bitcointalk.org/index.php?topic=177561.0
|
|
|
|
MrTeal
Legendary
Offline
Activity: 1274
Merit: 1004
|
|
July 16, 2014, 02:46:37 PM |
|
Interesting find OP. Isn't this chip technically speaking, an ASIC? I don't think it's an FPGA and it's definitely not a CPU or GPU. And if is an ASIC, then why can it only reach 40 hashes per second? I thought ASICs were supposed to be much faster than this? Also I found another thread about this chip where someone came up with a figure of 710 hashes per second (still very low however): http://bitcointalk.org/index.php?topic=177561.0That guy was just looking at how long it takes to load the data into the chip. He was ignoring how long it actually takes to compute the hash. And yes, the chip is an ASIC. It's not a bitcoin ASIC however, it's a very small, very slow ASIC designed to store keys and hash values. Completely different purpose, so it's not useful for Bitcoin mining.
|
|
|
|
|