|
Title: Bitcoin and ECC memory Post by: AliceWonderMiscreations on February 04, 2016, 02:47:46 PM Is there benefit to running ECC memory with a computer that generates a lot of payment addresses?
I guess what I am asking, is there risk of a faulty memory module with a stuck bit producing an incorrect public key or sha256 hash or ripemd160 hash? I suspect if the ripemd160 hash is correct, that a stuck bit impacting the sha256 hashes would usually result in a bad checksum, but what about generating the ECDSA public key or the ripemd160 hash of the public key? Thanks for any thoughts on this, especially from people who know. Title: Re: Bitcoin and ECC memory Post by: 2112 on February 04, 2016, 05:18:38 PM Is there benefit to running ECC memory with a computer that generates a lot of payment addresses? Depends on how you generate those addresses. If you have your own code and can add the function to re-check, then no. If you are just running Bitcoin Core then ECC is mandatory for reliable operation. Bitcoin Core lacks error checking in many of its operations.I guess what I am asking, is there risk of a faulty memory module with a stuck bit producing an incorrect public key or sha256 hash or ripemd160 hash? I suspect if the ripemd160 hash is correct, that a stuck bit impacting the sha256 hashes would usually result in a bad checksum, but what about generating the ECDSA public key or the ripemd160 hash of the public key? Thanks for any thoughts on this, especially from people who know. For details check the posts by etotheipi about his adventures with bit-flipping RAM with his non-ECC Intel i7 desktop machine used for development. Title: Re: Bitcoin and ECC memory Post by: AliceWonderMiscreations on February 05, 2016, 09:28:04 AM Thank you.
And the meek shall inherit the earth ;) |