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

[4ward]

Hello, here are some results from myr-gr with 3.8.1.1 and 3.8.0.1, same Ryzen 1800x with Win 8.1 64-bit, cpuminer-avx2-sha:

But I have reproduced that neoscrypt is broken on Windows, it works on Linux I will investigate.

Neoscrypt is broken since the cross-compile, I believe you said that you will have to rewrite it, but since its not very useful on cpu, its low on the priority list )

[1714943529]

1714943529

[1714943529]

1714943529

[1714943529]

1714943529

[joblo]

Hello, here are some results from myr-gr with 3.8.1.1 and 3.8.0.1, same Ryzen 1800x with Win 8.1 64-bit, cpuminer-avx2-sha:

But I have reproduced that neoscrypt is broken on Windows, it works on Linux I will investigate.

Neoscrypt is broken since the cross-compile, I believe you said that you will have to rewrite it, but since its not very useful on cpu, its low on the priority list )

It looks like yet another alignment issue on Windows. I'll be testing a potential fix later today.
About the rewrite, That has been delayed and will more likely be done incrementally instead
of a complete rewrite.

[gingerbutt]

Hey there. I sent you an email yesterday about adding an algorithm. Did you get a chance to see it? I'd gladly tip you for your help.
I stayed up pretty late last night troubleshooting, and I've already added the algo, but I can't get CPU temps or accepted/rejected hashes to show up when actually connected to a testnet pool.
https://github.com/TheDankstoneMemery/cpuminer-opt

[joblo]

Hey there. I sent you an email yesterday about adding an algorithm. Did you get a chance to see it? I'd gladly tip you for your help.
I stayed up pretty late last night troubleshooting, and I've already added the algo, but I can't get CPU temps or accepted/rejected hashes to show up when actually connected to a testnet pool.
https://github.com/TheDankstoneMemery/cpuminer-opt

There was no link in the email, not that I had already read it.

Try using LYRA2REV2 instead of LYRA2RE. I'm not sure why it makes a difference but it does.
You'll need to add another argument for the matrix, take a look at lyra2rev2.c for how it's done.

You left a lot of defaults for gate functions. are you sure that's correct?

I can add the algo, when it's working, and it will be much faster (with AVX2) compared to the your
fork. CPU temps are only avaiable on Linux.

[gingerbutt]

Hey there. I sent you an email yesterday about adding an algorithm. Did you get a chance to see it? I'd gladly tip you for your help.
I stayed up pretty late last night troubleshooting, and I've already added the algo, but I can't get CPU temps or accepted/rejected hashes to show up when actually connected to a testnet pool.
https://github.com/TheDankstoneMemery/cpuminer-opt

There was no link in the email, not that I had already read it.

Try using LYRA2REV2 instead of LYRA2RE. I'm not sure why it makes a difference but it does.
You'll need to add another argument for the matrix, take a look at lyra2rev2.c for how it's done.

You left a lot of defaults for gate functions. are you sure that's correct?

I can add the algo, when it's working, and it will be much faster (with AVX2) compared to the your
fork. CPU temps are only avaiable on Linux.

Currently working in Linux Mint 18. Will try your suggestion and give it a shot. Thank you! The source I used for Allium came from github.com/GarlicoinOrg/cpuminer-multi. After reading more about the algo itself, I did switch it over to LYRA2REV2 like you suggested, but I haven't taken a look at the extra args yet. Probably why I'm getting "too few arguments to function" when building.

As for the gate functions, I really don't know. I haven't dived into C for a couple years, and I just wanted to add the algo to the miner I prefer.

[gingerbutt]

I added null and 1 in the matrix, and now it builds, but it gives me a segmentation fault with either option. I think I'm missing something. Is there a chance I could chat with you through discord or something like that? Do you have time?

[joblo]

I added null and 1 in the matrix, and now it builds, but it gives me a segmentation fault with either option. I think I'm missing something. Is there a chance I could chat with you through discord or something like that? Do you have time?

You need to allocate the matrix when you register the algo, take a look at lyra2rev2-gate.c.
You're also missing some non-default gate functions. Take a look at the allium hooks in cpu-miner.c

Or do you just want me to do it?

[gingerbutt]

If you can do it, I'll gladly tip in GRLC if you'll accept. I'll still try my hand at hacking it together with my fork. I want to learn. Will try figuring this out on my end. Thanks!

[theboymaker]

Hi there,

I'm mining on x17 and I have this:

Code:

localhost:9286 detected new block, diff 15.186, target 0.000

is the fact of having target at 0 a problem?
what does it mean?

Thanks for your answers!

[joblo]

cpuminer-opt-3.8.2

https://github.com/JayDDee/cpuminer-opt/releases/tag/v3.8.2

Fixed and faster myr-gr.
Added x12 algo (Galaxie Cash), allium algo (Garlicoin).
Faster lyra2rev2, lbry, skein.
Large reduction in compiler warnings.

The Windows binaries crash mining neoscrypt, use the legacy version instead.
There is no estimate for resolution.

4 way has been implemented for sha256 and will be used instead of the SHA extension
on ryzen CPUs, as well as all CPUs with AVX2. Performance of sha256-4way is expected
to be better than SHA. Please report if otherwise. Affected algos are skein, lbry & myr-gr.

Garlicoin is  about to do a hard fork to switch to the allium algo. Please ensure the fork has
occurred before using allium.

There are occasional rejects mining allium (testnet) reporting low difficulty share. Based
on when they occur they appear to be stale instead of low diff. Low diff rejects are usually
a SW bug and affect all shares. Stale shares can occur and can sometimes occur in groups
until a new block is issued by the pool.

I messed up build.sh, changed to link to my local libraries and forgot to change it back.
It needs to be editted before using.

[gingerbutt]

Got a Garlicoin wallet I can donate to?

[wildzer0]

Can't mine any Garlicoin, shares are always rejected (tried 3 different pools):

Code:

[2018-02-15 23:16:32] Share submitted.
[2018-02-15 23:16:32] CPU #1: 107.11 kH, 113.17 kH/s
[2018-02-15 23:16:32] Rejected 14/14 (100.0%), diff 0.0294, 883.56 kH/s
[2018-02-15 23:16:32] reject reason: low difficulty share of 0.000040906673679398166

[gingerbutt]

Can't mine any Garlicoin, shares are always rejected (tried 3 different pools):

Code:

[2018-02-15 23:16:32] Share submitted.
[2018-02-15 23:16:32] CPU #1: 107.11 kH, 113.17 kH/s
[2018-02-15 23:16:32] Rejected 14/14 (100.0%), diff 0.0294, 883.56 kH/s
[2018-02-15 23:16:32] reject reason: low difficulty share of 0.000040906673679398166

You can't mine Garlicoin with Allium until the fork. You have to use -a scrypt:2048 until then.
You can try mining on a testnet pool like stratum+tcp://testnet.garlic-factory.fun:3333

[wildzer0]

Okay, thanks for the info. I was going by this list: http://pools.garlicoin.fun/ and choosing pools where it said "Allium ready: yes".

[gingerbutt]

Okay, thanks for the info. I was going by this list: http://pools.garlicoin.fun/ and choosing pools where it said "Allium ready: yes".

Yeah, that just means they're ready for the switchover this Friday. Should be seamless. Until then, you have to use scrypt:2048.

[zorachus99]

I've read about 20 pages and the readme on the download, but I can't seem to find some decent Windows command line examples for AES-NI and non AES-NI CPU's.

Basic configs to start would be nice.  What algo's are most efficient for this miner?

Thanks!

[ruplikminer]

I want to try this cpu జ్ఞ‌ా mining

[Enth]

At now for my 4690k @4700GHz on lyra2z algo the best is 3.8.1 - 522.17 kH/s, 3.8.2 gives me only 505 kH/s

joblo, what do you think, intel hardware prefetch must be turn on for max performance?

[theboymaker]

Hi there,

I'm mining on x17 and I have this:

Code:

localhost:9286 detected new block, diff 15.186, target 0.000

is the fact of having target at 0 a problem?
what does it mean?

Thanks for your answers!

Any answer please?

[host21]

Hi is this normal?

[2018-02-16 16:46:12] Share submitted.
[2018-02-16 16:46:12] CPU #0: 2 H, 3.04 H/s
[2018-02-16 16:46:13] Accepted 39/39 (100%), 2 H, 3.04 H/s, 100C


Intel xeon 1.7ghz for cryptonight . Only 3.04 H/s.

Can I optimize something?