|
Title: Memory based algorithm coin ? Post by: tytanick on March 26, 2013, 11:44:03 AM Hi guys.
I think that it would be very interesting coin that performs hashing mainly based on how much ddr3 memory i have in my pc. More memory = faster hashing + it uses strong cpu or graphic cards ? That would provide protection from ASICs that couldnt have so much memory? Anybody know how to do such alghoritm ? Meaby something like lzma compression extreme which needs big amount of memory ? tell me your opinion ? Title: Re: Memory based algorithm coin ? Post by: Buffer Overflow on March 26, 2013, 11:46:12 AM Protection from ASICs? ???
Title: Re: Memory based algorithm coin ? Post by: ehoffman on March 26, 2013, 12:40:49 PM There's one algo for that, it's called scrypt. There are parameters you can tweak that adjust the memory requirement. It is used for Litecoins (LTC). Litecoin use ~128K memory so it can fit easily in the CPU L2 cache. However, memory requirement is small enough that GPU can compete. But one could easily create a coin requiring 1GB of memory, in which case a GPU probably would hardly have more advantage (if any) than CPU.
Title: Re: Memory based algorithm coin ? Post by: tytanick on March 26, 2013, 12:41:48 PM protection i mean fair hashing power per spent dolars for hardware.
and also avaible to everyone, not for small amount of people (asic :) Title: Re: Memory based algorithm coin ? Post by: hanzac on March 26, 2013, 02:48:18 PM You are too late. LTC has been here more than 1 year.
Title: Re: Memory based algorithm coin ? Post by: tytanick on March 26, 2013, 08:15:05 PM There's one algo for that, it's called scrypt. There are parameters you can tweak that adjust the memory requirement. It is used for Litecoins (LTC). Litecoin use ~128K memory so it can fit easily in the CPU L2 cache. However, memory requirement is small enough that GPU can compete. But one could easily create a coin requiring 1GB of memory, in which case a GPU probably would hardly have more advantage (if any) than CPU. Litecoin can be mined on GPUs, but is there possibility to mine litecoins in asics ? Title: Re: Memory based algorithm coin ? Post by: DarkHyudrA on March 26, 2013, 08:40:10 PM There's one algo for that, it's called scrypt. There are parameters you can tweak that adjust the memory requirement. It is used for Litecoins (LTC). Litecoin use ~128K memory so it can fit easily in the CPU L2 cache. However, memory requirement is small enough that GPU can compete. But one could easily create a coin requiring 1GB of memory, in which case a GPU probably would hardly have more advantage (if any) than CPU. Litecoin can be mined on GPUs, but is there possibility to mine litecoins in asics ? The Bitcoin's ASIC no, since those ASICs are made for SHA-256, not Scrypt. Title: Re: Memory based algorithm coin ? Post by: tacotime on March 27, 2013, 05:31:43 AM See this thread for why it's a bad idea to make scrypt use tons and tons of memory
this comes up like every week or so https://bitcointalk.org/index.php?topic=122256.0 Title: Re: Memory based algorithm coin ? Post by: Buffer Overflow on March 27, 2013, 07:46:39 AM If it's just CPU mining, what protects the network from botnets?
Title: Re: Memory based algorithm coin ? Post by: Buffer Overflow on March 27, 2013, 09:21:24 AM I have some secret ideas "for botnets" :) But firstly we need to design CPU-only mining and then we'll see how to defeat botnets. Also you should distinguish my plans (backed by aliens) and the plans of this topic's TS. I've started my own thread in Russian section. There will be english sister-thread later. Sounds dubious. Title: Re: Memory based algorithm coin ? Post by: tacotime on March 27, 2013, 03:25:08 PM @tacotime Thanks for the link ! Quote 2. SHA256 followed by BLAKE256 followed by keccak256 (SHA3-256) for the crypt algorithm, to enhance circuit size in ASICs without strongly affecting hash speed. I prefer to "enhance" like this :Code: 2. Grøstl256 followed by keccak256 (SHA3-256) for the crypt algorithm, to enhance circuit size in ASICs without strongly affecting hash speed. That approach was naive, although I figured out another solution to it. Title: Re: Memory based algorithm coin ? Post by: tacotime on March 27, 2013, 04:15:49 PM No, I couldn't think of an implementation for radix sort in a hashing algorithm that couldn't be done more quickly and with less memory than quicksort, and also any way to really effectively insert it into the hashing algorithm and be easily invertible.
I'll pm you later about the other implementation, it's still FPGA and ASIC hard. I wish I had the time to work on it more or the money to pay a coder to implement it. Title: Re: Memory based algorithm coin ? Post by: mr_random on March 27, 2013, 06:08:33 PM I'll pm you later about the other implementation, it's still FPGA and ASIC hard. I wish I had the time to work on it more or the money to pay a coder to implement it. What makes you so sure it's FPGA and ASIC hard? |