[LOCKED] cpuminer-opt v3.12.3, open source optimized multi-algo CPU miner

[guytechie]

Sorry, been out of the loop for a while.  Any new algo optimizations between 3.8.3.3 and 3.8.4.1?

Looks like some bug fixes, an API feature, and a new algo (yescryptr32).

So far, I don't see anything I would benefit from so far mining on Windows using zergpool, ahashpool, blazepool, and MPH.

[Gigi75]

@joblo
I've already tried re-installing libcurl and libssl / libcrypto (different versions) and modify LD_LIBRARY_PATH & LD_RUN_PATH without any results.
I use a chrooted env. because it's easier for modyfing everything but basicaly I have exactly the same isssu on my bare Ubuntu 16.10
I know it's for sure a problem of bad shared library compatibility but i cannot figure out which one are the root cause of the issu.

What are the shared lib. version does cpuminer use?
Can you send me the result of 'ldd cpuminer-opt' in your working system?
It'll be a great help to overcome this issue!

Thank you for your help,
Guillaume

[joblo]

@joblo
I've already tried re-installing libcurl and libssl / libcrypto (different versions) and modify LD_LIBRARY_PATH & LD_RUN_PATH without any results.
I use a chrooted env. because it's easier for modyfing everything but basicaly I have exactly the same isssu on my bare Ubuntu 16.10
I know it's for sure a problem of bad shared library compatibility but i cannot figure out which one are the root cause of the issu.

What are the shared lib. version does cpuminer use?
Can you send me the result of 'ldd cpuminer-opt' in your working system?
It'll be a great help to overcome this issue!

Thank you for your help,
Guillaume

Providing my shared lib lineup won't help because it's a different OS release, you meed the default for
what you have. The best way to do that is build a virgin Ubuntu 16.10 or Debian VM and make sure it
works with all defaults. Then apply your mods one at a time until it breaks.

[felixbrucker]

one thing you could try: execute the build in docker, iirc there already is a docker file ready

then once it is built and everything works you can check the libs inside the docker container and compare them to your ones

[Bannedseller]

Hi Dev,
I will ask the same question as above:
Do you have any intentions to add sha224 that Bismuth algorithm is using: https://bitcointalk.org/index.php?topic=1896497.0
How hard will it be to implement sha224 ?
Thank you

[ZenFr]

Hi JayDDee,
Over 70 algo with CPUMiner-opt : very good, congratulation :-).
But one of the most popular is not in : EquiHash.
Is there any chance to have one day EquiHash in CPUMiner-opt ?

[4ward]

Hi JayDDee,
Over 70 algo with CPUMiner-opt : very good, congratulation :-).
But one of the most popular is not in : EquiHash.
Is there any chance to have one day EquiHash in CPUMiner-opt ?

Any chance to get the equihash algorithm added?  I'm still using nheqminer for my cpu.

Completely different design, too much work to integrate for no benefit.

[joblo]

Too many requests for algos, I don't respond anymore just to say no. If I'm interested you will know
soon enough.

Feel free to post of a new algo you believe will fit with cpuminer, the more info about the algo the better.
The best references are C source code of a working unoptimized example.

Typical reasons why I won't implement a given algo:

1. Requires a complete rewrite.
2. Existing miner is already optimized.
3. Lack of details/specs about the algo.
4. Lack of support by pools, exchanges, etc.

[Enth]

i'll be waiting for argon2d)

[joblo]

cpuminer-opt-3.8.5

Added argon2d-crds and argon2d-dyn algos.
sha256t 8 way AVX2 & 4 way SSE4.2 optimized.
CPUs with SSE4.2 get optimizations previously reserved for AVX.

https://github.com/JayDDee/cpuminer-opt/releases/tag/v3.8.5

[reeyon]

@joblo

Can you add new algo x16s?

[Enth]

cpuminer-opt-3.8.5

Added argon2d-crds and argon2d-dyn algos.
sha256t 8 way AVX2 & 4 way SSE4.2 optimized.
CPUs with SSE4.2 get optimizations previously reserved for AVX.

https://github.com/JayDDee/cpuminer-opt/releases/tag/v3.8.5

AMAZING!!! Thank you!

[4ward]

cpuminer-opt-3.8.5

Added argon2d-crds and argon2d-dyn algos.
sha256t 8 way AVX2 & 4 way SSE4.2 optimized.
CPUs with SSE4.2 get optimizations previously reserved for AVX.

https://github.com/JayDDee/cpuminer-opt/releases/tag/v3.8.5

Excellent )

significant improvement in SH256t (9mh -> 32mh)
argon2d algos seem somewhat slower than the reference on avx2 / windows

[Gigi75]

@joblo
@felixbrucker
Thank you guys, I'll give a try with Docker...

[4ward]

@joblo

Can you add new algo x16s?

a reference to the algo details helps

https://bitcointalk.org/index.php?topic=3208091.0

[Enth]

argon2d algos seem somewhat slower than the reference on avx2 / windows

yes, but more stable for me on 4 treads. Original gives me bsod...

[joblo]

argon2d algos seem somewhat slower than the reference on avx2 / windows

If by reference you mean the Beastpool fork, it was forked from cpuminer-opt and I just imported
all the argon2d code as-is. It should be identical.

How much slower is it? Are both algos similarly affected? Since you're using the binary my only guess
is a compile issue.

[4ward]

argon2d algos seem somewhat slower than the reference on avx2 / windows

If by reference you mean the Beastpool fork, it was forked from cpuminer-opt and I just imported
all the argon2d code as-is. It should be identical.

How much slower is it? Are both algos similarly affected? Since you're using the binary my only guess
is a compile issue.

yes, the Beastpool fork

crds 26k vs 32k
dyn 8k fvs 10k

I can try compiling both from source to verify, but that will probably only happen on the weekend

[joblo]

argon2d algos seem somewhat slower than the reference on avx2 / windows

If by reference you mean the Beastpool fork, it was forked from cpuminer-opt and I just imported
all the argon2d code as-is. It should be identical.

How much slower is it? Are both algos similarly affected? Since you're using the binary my only guess
is a compile issue.

yes, the Beastpool fork

crds 26k vs 32k
dyn 8k fvs 10k

I can try compiling both from source to verify, but that will probably only happen on the weekend

Compiling beastpool requires a few extra steps. First you need to reverse the pthreadGC2 change for Linux
and chmod +x autogen.sh and build.sh.

I get virtually the same speed for both on crds: 29.8k opt vs 29.6k beast.
What compiler was used for the Beaspool bin?

[4ward]

argon2d algos seem somewhat slower than the reference on avx2 / windows

If by reference you mean the Beastpool fork, it was forked from cpuminer-opt and I just imported
all the argon2d code as-is. It should be identical.

How much slower is it? Are both algos similarly affected? Since you're using the binary my only guess
is a compile issue.

yes, the Beastpool fork

crds 26k vs 32k
dyn 8k fvs 10k

I can try compiling both from source to verify, but that will probably only happen on the weekend

Compiling beastpool requires a few extra steps. First you need to reverse the pthreadGC2 change for Linux
and chmod +x autogen.sh and build.sh.

I get virtually the same speed for both on crds: 29.8k opt vs 29.6k beast.
What compiler was used for the Beaspool bin?

SW built on Mar 12 2018 with GCC 5.3.1.