don’t know, it’s I7 3612QM.
with 6 threads only get’s a bit higher:
Benchmark done!
Total time : 25350
Total hashes: 100
Total solutions found 198
Speed: 3.94477 H/s
Speed: 7.81065 S/s
I thought it’s suppose to be 188 solutions
Number of solutions over 100 consecutive nonces depends on starting nonce. A starting nonce of 1000 gives 188 solutions, but apparently the above picked a different start.
Btw, it is most unfortunate that nicehash refers to equihash runs as “hashes”,
since this leads to exactly the confusion between h/s and S/s that I warned about in
tx for the explanation
Does anyone expect they will use the default miner/solver, or will everyone be using a custom miner/solver?
Default solver is very very slow
what do you plan to use on launch?
The fastest solver that is opensource
2 Likes
@Valarray: it crashes if cpu does not have AVX, but there’s new sse release now
I’m a dev, but new to cryptoccy (last couple of years).
Is my understanding right that it’s completely cool to use custom solvers/miners? I just wonder how we can be sure they are 100% compatible with the network/protocol?
I guess just keep an eye on the forums and see who’s is best around launch time?
Does absolutely no-one use the default bitcoin miner then?
They are just combining the faster solver into zcashd. If you want to run the slow miner no problem more for us 
most of people where mining with the original miner until yesterday. If I’m right, nheqminer is first complete miner released to public today
2 Likes
thanks for info…
sararth/tromp seems better for CPU mine (Which i’m doing)
yes right, I forgot Sarath’s work sorry. I should have said “first binaries released to the public”
1 Like
one last question, is there something special to do to get the miner to take advantage of multiple cores?
or will it ramp up all by default?
genproclimit=numberofcores in your config file
Thanks!
in the zcash.conf?
1 Like
Ahhh got it, it’s genproclimit not procgenlimit
Correct , sorry I always make that mistake
1 Like
thanks for the help! I’m excited to be a ZEC billionaire 
1 Like
I haven’t looked at xenoncat’s blake2b, but I know in general about the following high-speed implementations of blake2b:
-
the official blake2b library (main author Samuel Neves) contains optimized implementations for various CPU instructions like SSE and AVX: GitHub - BLAKE2/BLAKE2: BLAKE2 official implementations
-
Floodyberry’s implementation was at one point the speed leader: GitHub - floodyberry/blake2b-opt: Optimized, portable implementations of BLAKE2b I think (but don’t remember for sure) that the official library subsequently caught up to it.
-
Samuel Neves’s SIMD implementation of BLAKE2bp is the current speed leader: GitHub - sneves/blake2-avx2: BLAKE2 AVX2 implementations You can’t benefit from this implementation unless you want to do (at least) 4 BLAKE2b’s in parallel. BLAKE2bp has got 4 parallel BLAKE2b’s inside it.
2 Likes