Author

Topic: Cryptographic Coprocessor (Read 4046 times)

full member
Activity: 163
Merit: 100
April 04, 2013, 03:24:25 PM
#9
I think I have heard mixed reviews, I myself have not used it. However I believe I have heard more unpleasant complaints rather than compliments, I may be mistaken though.

Thank you.
full member
Activity: 196
Merit: 100
April 01, 2013, 11:31:31 PM
#8
It will not work very quickly.....!!!!!!!!!

1. bitcoin is Sha256(Sha256(x)), where x is the data+nonce
2. even if you were to implement this card, you would have to get the data from the card and then send it back in again!!!!
Then when it came out again, you would need to compare the  ass end for a row of zeros

So even with the fastest fucking CPU on market.. it is STILL going to be slower than even a shitty FPGA, why?
Well because an FPGA can collapse down 'Sha256(Sha256(x)) top bits =0' into one fucking clock cycle.....

Where as your setup is going to require several syncronised clock cycles to do the above.......(certainly to get the data over PCI)
 AND your main nonce increment is going to be OUTSIDE the high-speed logic & inside your overpowered CPU.....


Go learn a little more about bitcoin and how it works.....
hero member
Activity: 700
Merit: 500
April 01, 2013, 05:05:19 AM
#7
Hate to be the guy that just keeps coming back but what about something like the nShield solo?

http://www.thales-esecurity.com/products-and-services/products-and-services/hardware-security-modules/general-purpose-hsms/nshield-solo

All of these cards are designed to speed up a variety of cryptographic operations instead of raw SHA-256 hashes like bitcoin demands. These products are all kind of like Swiss Army knives in that they do a lot of things, but don't do any of them especially well. Many of these tasks they will handle better than general purpose hardware, but usually when they get used it is because guidelines require them as opposed to when performance demands them.
full member
Activity: 163
Merit: 100
March 31, 2013, 09:28:43 PM
#6
Hate to be the guy that just keeps coming back but what about something like the nShield solo?

http://www.thales-esecurity.com/products-and-services/products-and-services/hardware-security-modules/general-purpose-hsms/nshield-solo
full member
Activity: 163
Merit: 100
March 27, 2013, 11:03:21 AM
#5
What about something that is designed to run on x86_64 hardware:
http://h18004.www1.hp.com/products/servers/security/axl600l/

The HP/Atalla AXL600L SSL Accelerator Card is a 33-MHz, 32-bit device that incorporates a dual voltage signal bus and meets or exceeds PCI 2.2 specifications.

Maybe not the best idea then. Just really wondering if there is some hardware already available which could speed up hashing + low power usage.
hero member
Activity: 682
Merit: 500
March 27, 2013, 10:02:32 AM
#4
What about something that is designed to run on x86_64 hardware:
http://h18004.www1.hp.com/products/servers/security/axl600l/

The HP/Atalla AXL600L SSL Accelerator Card is a 33-MHz, 32-bit device that incorporates a dual voltage signal bus and meets or exceeds PCI 2.2 specifications.
full member
Activity: 163
Merit: 100
March 27, 2013, 09:09:46 AM
#3
What about something that is designed to run on x86_64 hardware:
http://h18004.www1.hp.com/products/servers/security/axl600l/
hero member
Activity: 952
Merit: 1009
March 27, 2013, 09:03:58 AM
#2
Considering the card supports neither standard PC architecture nor common OSes I'd guess no.

The systems the card can be plugged into usually cost millions.
full member
Activity: 163
Merit: 100
March 27, 2013, 08:52:58 AM
#1
Anyone tried mining with this:
http://www-03.ibm.com/security/cryptocards/pciecc/overview.shtml

It's an IBM cryptographic co-processor. It should be able to do hashing if my understanding is correct.
Jump to: