On TEThashV1 , and ProgPoW
Finally I’ve made FNV changes in MIX parts.
Here’s my issue suggestion on ethereum issues
request to change Keccak algorithm to newer one
and change MIX algorithm in EThash.
I’ve analysed FNV , and it is depricated current usages on ethash.
So, I changed FNV following original Offset Based FNV1A.
On This time, PoW algorithm change to newer one, it needs crypto analysis and long time verification time. it’s not suitable for prevent ASIC , so late!
But, in case of TEthashV1 (Trust Ethash Version1) has most of things are based on ethash, which is verified several years, but MIX FNV part uses deprecated implementation , So, if this is changed then more strong PoW algorithm based on ETHASH.
Clearly, make obsolete the current ASIC miners too.
And, I also researching FPGA Mining, recent days of synthesizer (eg. Xilinx or Altera) it has tremendous optimization options and great synthesis abilities, So, It is not possible to prevent ASIC.
And Clearly agreed on @timolson 's Hash per Watts opinion.
So, Above reasons, it may be regular small parts of algorithm changing is more good for resist centralized mining and prevent ASICs.
In case of ASIC, at least 6 months needs, redesign and consumer products delivery.
Another My opinion on Ethash , is remove 64 rounds of Mix to decrease to 32 rounds.
decrease DAG Size or limit Max DAG Size, sustain current Memory accessing bw.
So, then cheaper or some years ago GPU also effective hash on PoW, and many guys can participates mining.
Now, Above TethashV1 is appied on
- TEO PreTest-Mainnet ,
- Miner and
- rTEO parity forked client
- go-wrapper TethashV1 for Pool
- TethashV1 cpp library (chfast/MariusVanDerWijden/ethash) it is insided in teominer/libethash
also patched to new TEthashV1 algorithms for go-ethash wrapper , cpp ethash library.
please check and comments for us.
Next 6 months after minor algorithm patch for make a time.