Bitcoin Forum

MCM is Monero (XMR) CPU miner with the best performance. Originally based on xmrig (net and config) with completely rewritten algo from scratch on ASM.
On SSE4 (non AES-NI) cpu performance boost is up to 50% (depends on cpu architecture). More description on Github: https://github.com/alex-tk/mcm (https://github.com/alex-tk/mcm)

Releases: https://github.com/alex-tk/mcm/releases (https://github.com/alex-tk/mcm/releases)

v0.9.6
 - added support for new algorithm cryptonight/r, short alias cn/r (also known as CryptoNightR or CryptoNight variant 4), for upcoming Monero fork

v0.9.5
 - added new algorithm cn-pico/trtl (aliases cryptonight-turtle, cn-trtl) for TurtleCoin (TRTL) fork.
 - added new algorithm cn/half for Masari and Stellite forks.

v0.9.4
+20% for cn/2 (CryptoNight variant 2 for Monero fork) --av=3 (Software AES implementation).

v0.9.3
- performance boost for CryptoNight variant 2 for Monero fork --av=3 (Software AES implementation)

v0.9.2
- added new algorithm CryptoNight variant 2 for Monero fork
- added CryptoNight-Lite and CryptoNight-Heavy support
- added SSL/TLS support for secure connections to pools
- based on xmrig 2.8.1 (https://github.com/xmrig/xmrig/blob/master/README.md)

v0.9.1
+2% performance boost for --av=0 (Software AES implementation)

Core2 Quad Q9300 2.5 GHz 6M, 1 & 4 threads, 64-bits, Software AES implementation

algo	cn/r	cn/2	cn-lite/1	cn-trtl
mcm 0.9.6	60	84	234	760
xmrig 2.13.1	54	50	158	760

Please post you hashrates.

Linux build added

How much is the devfee?

SSE4, so this can potentially make (E7+) Core2Duo's hashrate competent enough to earn a profit?
I'd love to try this with my scrap machines.

<Edit>
Waiting for a legit test response or a code review.

From the info in GitHub the dev fee is 1%.
I would love some feedback - if it's indeed faster than others and if it doesn't contain other.. surprises (no offense, dev, but being a little suspicious never hurts)

E7+ with 3M of L2 can be run in 2 threads and with expected hashrate about 60H/s. The biggest boost will be for older architectures like Yorkfield/Wolfdale about  40-50%, for newer it will be less (ex Haswell about 15-20%). For AES-NI it is about 1% for now.

I want to use a processor for mining XMR which processor is more efficient? Advise please.

Yes it is big problem that antiviruses detect it as dangerous software (Rule: XMRIG_Monero_Miner Ruleset: CryptoCoin Miner). People afraid to use such software. So I have to rewrite all from scratch. It will take some time.

Most crypto miners are detected by antiviruses. That's why I didn't even bother to check with VirusTotal.

But a miner you actually allow to run on your computer can do many more, even searching the HDD for wallets or downloading a payload that does that. I mean, it's just some code and the possibilities are limitless.
That's why I'll wait for trusted 3rd party to review the code.

If you plan to upgrade your pc new cpu on which you want get some xmr then AMD Ryzen is great choice - they have a lot of L3 cache. So even Ryzen 3 1200 will perform about 250H/s. To get more details you can use http://monerobenchmarks.info/singleCPUAMDINTEL.php (http://monerobenchmarks.info/singleCPUAMDINTEL.php). The main rule - 1.5-2M of cache per thread.

new release v0.9.1: 2% performance boost for --av=0 (Software AES implementation)

I've done today a test, since I have at hand an older computer where it doesn't matter if the miner is safe or not.
CPU is i7-920, CPU L2/L3 1/8
The hashrate reported by the pool is 87 H/s = the average for the last 3 hours.


If I remember right, the old miners (way more than one year ago) were showing 75-80 H/s, and they were using the CPU 100%, while this one uses 66%.
So the numbers look pretty good. All this is on default command line and config, do I have to change anything?

It seems that auto detection failed and set treads count to 5 (that's why cpu usage 66% - 5/8), but for your cpu it would be better 4, try to set it manualty "-t 4" for command line or "threads: 5" in config file, it should increase hashrate

The command line didn't work, it didn't care of my setting. I've set to 4 in the config.
The hashrate displayed by the pool has big variations, between 33 and 200 H/s, having the 6h average 90 H/s and 3h average 77 H/s. So the setting change doesn't seem to help.

You can check threads count at startup, it will display: "* THREADS: [number]"

The pool and the miner calculate hashrate in different ways. The pool calculates it based on found shares. The miner calculates it based on generated hashes.What hashrate is in the miner?

I know. It's 4 now.


85.7 H/s avg: 86.4 H/s max: 93.6 H/s

The hashrate shown by pool should be based on what they receive from my miner in a period of time.
I rely on what the pool tells because long ago the Monero miners used to show nice hashrates, but they've got disconnected very often leading to a lot of lost work and much lower paid hashrate.


However, I think that I will stop the miner, since it would need half a year to get to minimum payout :)

New release for Monero fork

Releases: https://github.com/alex-tk/mcm/releases (https://github.com/alex-tk/mcm/releases)

v0.9.2
- added new algorithm CryptoNight variant 2 for Monero fork
- added CryptoNight-Lite and CryptoNight-Heavy support
- added SSL/TLS support for secure connections to pools
- based on xmrig 2.8.1

0.9.2 Linux build added

Nice to see some support for non AES CPUs
I'm currently using JCE on a Q6600 and I'm getting 66.79h/s
mcm gives me only 61.7h/s
Is there room for improvement here?

 * ABOUT        MCM/0.9.2 based on XMRig 2.8.1 gcc/5.4.0
 * LIBS         libuv/1.8.0 OpenSSL/1.0.2g microhttpd/0.9.44
 * CPU          Intel(R) Core(TM)2 Quad CPU           @ 2.40GHz (1) x64 -AES
 * CPU L2/L3    8.0 MB/0.0 MB
 * THREADS      4, cryptonight, av=0, donate=1%
 * ASSEMBLY     auto:none
 * POOL #1      pool.supportxmr.com:3333 variant auto

Yes, there is room for improvement for v8 (non aes-ni version). I am working on it.

new release with performance boost for CryptoNight variant 2 for Monero fork

Need much more to reach xmr-stak

http://prntscr.com/lk1vn9

http://prntscr.com/lk1wkl

http://prntscr.com/lk1xbv

http://prntscr.com/lk2gtk

http://prntscr.com/lk2hzh

Test on my server.

Try to use double hash mode (--av=2), by default it is av=1

https://prnt.sc/lka8zj (https://prnt.sc/lka8zj)

0.9.3 Linux build added

new release 0.9.4:
+20% for cn/2 (CryptoNight variant 2 for Monero fork) --av=3 (Software AES implementation).

on Core2 Quad Q9300 4 threads, 64-bits, Software AES implementation - 84H/s

new release v0.9.5 (https://github.com/alex-tk/mcm/releases):
 - added new algorithm cn-pico/trtl (aliases cryptonight-turtle, cn-trtl) for TurtleCoin (TRTL) fork.
 - added new algorithm cn/half for Masari and Stellite forks.

I can't run this app. It closed automatically. any one can help?

What cpu & os do you use? Did you edit config.json (url & user)?

You can create run.bat with content:

mcm.exe
pause

to see messages and errors if any

fixing cn2 algo in v0.9.5

v0.9.5 fixed

new release v0.9.6 (https://github.com/alex-tk/mcm/releases):
added support for new algorithm cryptonight/r, short alias cn/r (also known as CryptoNightR or CryptoNight variant 4), for upcoming Monero fork

i can not seem to grab the release? is there something wrong with the links?
i tried in several different ways and on a few different machines along with several different operating systems?
seems like the link is empty

At the bottom you can see it.

http://i68.tinypic.com/30c3gpi.png

moved changelog to previous release to be able see assets at the top of the page

Do you plan to support algo-switching?

If your pool support mining algorithm negotiation, don't need change variant option (auto switching in place)

Updated release with enabled api functionality