Time to refresh the discussion
![Grin](https://bitcointalk.org/Smileys/default/grin.gif)
range 2^64
BSGS HT=2^30 = 9GB
OP to scan whole range 2^64/2^30=
2^34 op BSGS with DP and endomorphisms
BSGS HT=2^30 * DP 2^8 = 2^38 = 9GB (We have 2^30 points but range computed is 2^38)
HT contain DP 2^29 pointG, 2^28 pointLambda1, 2^28 pointLambda2, totaly 2^30 points
jump will be equil to double HTsize = 2^38 * 2 = 2^39
OP to scan whole range 2^64/2^39=2^25 and each jump cost 2^8 op
We will use 2 endomorphism beta1 and beta2 so each point computed from pubkey give us 3 point(point,beta1 and beta2)
So we can 3 times faster(in average) calculate each DP and jump will cost 2^6.4 op
Totaly to scan whole range we need 2^25 jumps and each jump cost 2^6.4 =
2^31.4 opHT table content:
G,2G,3G...2^29G
(2^29G+G)*Lambda1,(2^29G+2G)*Lambda1,(2^29G+3G)*Lambda1...(2^29G+2^28G)*Lambda1
(2^29G+2^28G+G)*Lambda2,(2^29G+2^28G+2G)*Lambda2,(2^29G+2^28G+3G)*Lambda2...(2^29G+2^28G+2^28G)*Lambda2
Not Bad !!! I thnk this is a faster then kangaro or =. Kangaroo is not stable result. I may start kangaroo 2 times and get 2 result- not found and found. BSGS eting many many memory, but BSGS as I remeber work with big ranges for ex 128 bytes, I not remeber about 2 mptnhs for calculation in BSGS, but Kangaroo many times say me what wayt 2 monts....
There is the link to BSGS code with endomorthism ? Have you implement hash dump from memory to disk ?
To all many respected members. THEN YOU SHOW CODE FOR ETHEREUM ![Huh](https://bitcointalk.org/Smileys/default/huh.gif)
ETH is simle then BTC as I know(BTC pubkey is a SHA+SHA..., ETH ONLY 1 SHA !!!)
![Huh](https://bitcointalk.org/Smileys/default/huh.gif)
?
Br.