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

[ranzo]

Hello,
I know it's not officially supported but I am trying to compile on freebsd and I thought maybe someone has an idea how to fix this.

CFLAGS="-O3 -march=native -Wall" ./configure --with-curl=/usr/local
make

Code:

algo/yescrypt/yescrypt.c: In function 'scanhash_yescrypt':
algo/yescrypt/yescrypt.c:388:25: error: expected declaration specifiers or '...' before numeric constant
         uint32_t _ALIGN(64) vhash[8];
                         ^
algo/yescrypt/yescrypt.c:388:18: error: expected ')' before '+' token
         uint32_t _ALIGN(64) vhash[8];
                  ^
algo/yescrypt/yescrypt.c:388:18: error: expected ')' before '&' token
         uint32_t _ALIGN(64) vhash[8];
                  ^
algo/yescrypt/yescrypt.c:389:25: error: expected declaration specifiers or '...' before numeric constant
         uint32_t _ALIGN(64) endiandata[20];
                         ^
algo/yescrypt/yescrypt.c:389:18: error: expected ')' before '+' token
         uint32_t _ALIGN(64) endiandata[20];
                  ^
algo/yescrypt/yescrypt.c:389:18: error: expected ')' before '&' token
         uint32_t _ALIGN(64) endiandata[20];
                  ^
algo/yescrypt/yescrypt.c:398:26: error: 'endiandata' undeclared (first use in this function)
                 be32enc(&endiandata[k], pdata[k]);
                          ^~~~~~~~~~
algo/yescrypt/yescrypt.c:398:26: note: each undeclared identifier is reported only once for each function it appears in
algo/yescrypt/yescrypt.c:402:59: error: 'vhash' undeclared (first use in this function)
                 yescrypt_hash((char*) endiandata, (char*) vhash, 80);

this is the part of yescrypt.c where it throws the error

Code:

int scanhash_yescrypt( int thr_id, struct work *work, uint32_t max_nonce,
                       uint64_t *hashes_done )
{
        uint32_t _ALIGN(64) vhash[8];
        uint32_t _ALIGN(64) endiandata[20];
        uint32_t *pdata = work->data;
        uint32_t *ptarget = work->target;

        const uint32_t Htarg = ptarget[7];
        const uint32_t first_nonce = pdata[19];
        uint32_t n = first_nonce;

        for (int k = 0; k < 19; k++)
                be32enc(&endiandata[k], pdata[k]);

        do {
                be32enc(&endiandata[19], n);
                yescrypt_hash((char*) endiandata, (char*) vhash, 80);
                if (vhash[7] < Htarg && fulltest(vhash, ptarget)) {
                        work_set_target_ratio( work, vhash );
                        *hashes_done = n - first_nonce + 1;
                        pdata[19] = n;
                        return true;
                }
                n++;
        } while (n < max_nonce && !work_restart[thr_id].restart);

        *hashes_done = n - first_nonce + 1;
        pdata[19] = n;

        return 0;
}

[1714120821]

1714120821

[1714120821]

1714120821

[1714120821]

1714120821

[1714120821]

1714120821

[1714120821]

1714120821

[1714120821]

1714120821

[joblo]

I know it's not officially supported but I am trying to compile on freebsd and I thought maybe someone has an idea how to fix this.

Don't expect  free ride, you'll have to do most of the work.

I fixed a FreeBSD compile error in v3.6.6 and it was reported to compile and run after that.
There have been no significant changes to yescrypt code since then.

It looks like an issue with the ALIGN macro. It's defined in compat.h. There were likely earlier errors,
try to find the first one.

[joblo]

cpuminer-opt-3.8.6.1

Faster argon2d* AVX2.
Untested AVX-512 for argon2d*, YMMV.

https://github.com/JayDDee/cpuminer-opt/releases/tag/v3.8.6.1

Additional notes:

AVX512 may compile on Linux but fails on mingw due to this bug:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65782.

Linux users with AVX512F capable CPUs please report your results.

Thanks to a user with a i9-7820x AVX-512 is confirmed to work with argon2d.
Please note that there is no reporting of AVX-512 features in this release,
it will be added in the next release.

More info: https://github.com/JayDDee/cpuminer-opt/issues/128

Thanks to Beastpool for the great work with AVX512 on argon2d. I hope to add
some of my own contributions once Cannonlake is released, if it gets released.
The latest rumour I saw is the next mainstream CPU will be a Coffeelake refresh.
I hope it's wrong.

[StelioKontos]

Any plans on adding monero7/cryptonight7 algos? Xmr Stak supports monero7 now: https://github.com/fireice-uk/xmr-stak/releases

[mylooderso]

hi,

i am getting an error after it worked with an older version for 8 hours straight. downloaded the newer version and set it up following the instructions, still the same issue. does anyone have an idea?

Code:

cpuminer-avx2.exe -a xevan -o stratum+tcp://yiimp.eu:3739 -u iPytqLnJukKjfor31gZXT5DQkcijqxQAJL -t 3

         **********  cpuminer-opt 3.8.6.1  ***********
     A CPU miner with multi algo support and optimized for CPUs
     with AES_NI and AVX2 and SHA extensions.
     BTC donation address: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT


***\Mining\Bitsend\cpuminer-opt-3.8.6.1-windows>timeout /t 60

waiting for 52 Seconds. Continue with random button...      <--- translated this part from german

running on an i5-2500k

thanks in advance!

edit: after the timer runs out it just starts over again. tried it withou the /t-command, doesn't work

[andnet]

running on an i5-2500k

thanks in advance!

edit: after the timer runs out it just starts over again. tried it withou the /t-command, doesn't work

i5-2500 have not avx2 . use avx or sse42 file

[mylooderso]

running on an i5-2500k

thanks in advance!

edit: after the timer runs out it just starts over again. tried it withou the /t-command, doesn't work

i5-2500 have not avx2 . use avx or sse42 file

sse42 works, thank's! it said use avx2 for sandy & ivy bridge in the readme. that's why i tried only that...

[andnet]

running on an i5-2500k

thanks in advance!

edit: after the timer runs out it just starts over again. tried it withou the /t-command, doesn't work

i5-2500 have not avx2 . use avx or sse42 file

sse42 works, thank's! it said use avx2 for sandy & ivy bridge in the readme. that's why i tried only that...

In readme ..cpuminer-aes-avx.exe   "-march=corei7-avx"         Sandybridge, Ivybridge

[Digital_Seytan]

Hello CPUminers, I have a question among the users who have experience with this farmbatchbot download link: https://github.com/Undertrey/FarmWatchBot/releases
have used these with CPU-miner 3.8.6.1   ---->> please share.

[Bannedseller]

Will it be possible to use cpuminer-opt to mine Japanese Anonymous Cryptocurrency Koto: https://bitcointalk.org/index.php?topic=2728195.0
using RX480

[joblo]

Will it be possible to use cpuminer-opt to mine Japanese Anonymous Cryptocurrency Koto: https://bitcointalk.org/index.php?topic=2728195.0
using RX480

Yes and no. cpuminer-opt is a CPU miner, RX480 is a GPU.  I'm not aware of any GPU miners that support yescrypt algo.

[Larvitar]

Any plans on adding monero7/cryptonight7 algos? Xmr Stak supports monero7 now: https://github.com/fireice-uk/xmr-stak/releases

I don't think so. CPUminer-opt doesn't supports huge pages, and main focus are other algos. We have some good CN miners, but much less, for example, Yescrypt miners, Lyra2z miners, etc. This miner supports and optimizes CPU mining for it.

[joblo]

Any plans on adding monero7/cryptonight7 algos? Xmr Stak supports monero7 now: https://github.com/fireice-uk/xmr-stak/releases

I don't think so. CPUminer-opt doesn't supports huge pages, and main focus are other algos. We have some good CN miners, but much less, for example, Yescrypt miners, Lyra2z miners, etc. This miner supports and optimizes CPU mining for it.

Well said.

I'm trying to get AVX2 4 way finished before the AVX512 rush starts. There are still a couple of very challenging
algos that offer large potential. They are self contained and can't take advantage of any previous work but
will be worth it if I can get it done. If only people would stop creating new algos.

As far as cryptonight is concerned it's far down my list. Monero mining development is very competitive and
any work would be just to keep up. That would be a waste of time and effort when I can be productive elsewhere.

[joblo]

cpuminer-opt-3.8.7

https://github.com/JayDDee/cpuminer-opt/releases/tag/v3.8.7

Added argon2d4096 (alias argon2d-uis) for Unitus (UIS).
argon2d-crds and argon2d-dyn renamed to argon2d250 and argon2d500 respectively.
  The old names are recognized as aliases.
AVX512 is now supported for argon2d algos, Linux only.
AVX is no longer a reported feature and an AVX Windows binary is no longer
  provided. Use AES-SSE42 build instead.

Edit: it has been reported that the aliases for the argon2d algos aren't working.
Please use the folowing:

-a argon2d250  to mine Credits (argon2d-crds)
-a argon2d500 to mine Dynamic (argon2d-dyn)
-a argon2d4096 to mine Unitus (argon2d-uis)

[StelioKontos]

Any plans on adding monero7/cryptonight7 algos? Xmr Stak supports monero7 now: https://github.com/fireice-uk/xmr-stak/releases

I don't think so. CPUminer-opt doesn't supports huge pages, and main focus are other algos. We have some good CN miners, but much less, for example, Yescrypt miners, Lyra2z miners, etc. This miner supports and optimizes CPU mining for it.

Thanks for the answer.

[Enth]

about 3.8.7:

more hashrate per 3kh/s on argon2d
more stable for me on avx2

thanks!

[ZenFr]

I'm very surprised that a man like Joblo has only 2 personns who gave him merit !!!

I am the third : +5 merit for you Joblo and I think it is not so much for all you kindly job.
I invite each person who use your CPUMiner-opt to give you some merit : it is a minimum.
Thank you Joblo/JayDDee for your job ;-).

[C3Sector]

its sad to see avx support is dropped .. looks like its hard to maintain them all together .. . time to upgrade cpu !!
and you making great job there joblo, keep it upp !!

[StelioKontos]

I'm very surprised that a man like Joblo has only 2 personns who gave him merit !!!

I am the third : +5 merit for you Joblo and I think it is not so much for all you kindly job.
I invite each person who use your CPUMiner-opt to give you some merit : it is a minimum.
Thank you Joblo/JayDDee for your job ;-).

You're right, the problem is for me Merit System is crap so I tend to not support or use it because it basically froze my rank, so I'm holding a bit - to use an euphemenism - of a grudge against it.

Also he's already Legendary so I don't think he needs more merits anyway.

By the way I gave him one

[joblo]

its sad to see avx support is dropped .. looks like its hard to maintain them all together .. . time to upgrade cpu !!
and you making great job there joblo, keep it upp !!

Please explain. There should be no difference using the AES-SSE42 build on an AVX CPU.