[ANN] cpuminer-opt v3.14.2, open source optimized multi-algo CPU miner - page 12.

Nighthog

jr. member

Activity: 133

Merit: 5

edit...

Fixed the issue, used the wrong port on the pool with the "correct parameters"...

Code:

cpuminer-zen3.exe -a yespower --param-key "CPUpower: The number of CPU working or available for proof-of-work mining" --param-n 2048 --param-r 32 -o stratum+tcp://eu.blockmasters.co:4250 -u -p c=

JayDDee

full member

Activity: 1424

Merit: 225

Quote from: Nighthog on January 31, 2021, 04:32:30 AM

Does this support 'cpupower' in it's current state? (CPUchain Core)

Code:

PoW Mining Algorithm: yespower / 'cpupower' with N = 2048, r = 32 and perslen = 73

I tried the 'yespower' algorithm and got rejects spammed for to low difficulty on the hash.
I suppose it's not correct. A variant with these variables needed?

Scroll to the end...
https://github.com/JayDDee/cpuminer-opt/wiki/Supported--Algorithms

Nighthog

jr. member

Activity: 133

Merit: 5

Does this support 'cpupower' in it's current state? (CPUchain Core)

Code:

PoW Mining Algorithm: yespower / 'cpupower' with N = 2048, r = 32 and perslen = 73

I tried the 'yespower' algorithm and got rejects spammed for to low difficulty on the hash.
I suppose it's not correct. A variant with these variables needed?

dottat

newbie

Activity: 48

Merit: 0

Quote from: SpceGhst on May 26, 2020, 09:40:10 PM

It looks like once you find 10-12 blocks in solo mining, the periodic hash rate starts to even out. It’s hard with Minotaur because blocks don’t come that often, unlike pool mining where shares are 3-4 per minute.

You still around lcc/ring?

JayDDee

full member

Activity: 1424

Merit: 225

Quote from: nina34 on January 21, 2021, 02:43:34 PM

can i compile it on 32 bit linux system ?

No, 64 bit only. Read the requirements in the first post.

nina34

newbie

Activity: 1

Merit: 0

can i compile it on 32 bit linux system ?

i've got this error messing during compiling
"simd-utils/simd-int.h:94:18: error: ‘__int128’ is not supported on this target"

JayDDee

full member

Activity: 1424

Merit: 225

Quote from: UspesenRudar on December 28, 2020, 06:07:59 AM

Can you add new BALLOON algo for Zentoshi.
From https://github.com/zentoshi/cpuminer-opt-balloon
have trouble creating win EXE.
THX

It's only for one coin that isn't listed on any exchanges, so I'm not interested.
Zentoshi should provide Windows EXEs for mining.

scryptr

legendary

Activity: 1797

Merit: 1028

HAPPY HOLIDAYS! --

Here's to BTC peaking at $50K+ in 2021! --scryptr

UspesenRudar

newbie

Activity: 33

Merit: 0

Can you add new BALLOON algo for Zentoshi.
From https://github.com/zentoshi/cpuminer-opt-balloon
have trouble creating win EXE.
THX

7d2mine

newbie

Activity: 80

Merit: 0

Just tried this out for the first time, wanted to say THANKS for doing this!! Was able to figure it out and get it running with very minimal BS in about 5-10 mins.

JayDDee

full member

Activity: 1424

Merit: 225

cpuminer-opt-3.15.5

Fix stratum jobs lost if 2 jobs received in less than one second.

JayDDee

full member

Activity: 1424

Merit: 225

Quote from: spilk76 on December 13, 2020, 06:18:47 AM

hello humans I can't compile cpuminer-multi on my arm device
help me

This thread is for cpuminer-opt, not cpuminer-multi. You might have better luck in the cpuminer-multi thread.

The error is in ARM ASM code, possibly a syntax error, but I know nothing about ARM ASM so it's just a guess.

spilk76

newbie

Activity: 4

Merit: 0

hello humans I can't compile cpuminer-multi on my arm device
help me

Code:

```
u0_a96@localhost:~/cpuminer-multi$ make
make all-recursive
make[1]: Entering directory '/data/data/com.termux/files/home/cpuminer-multi'

Making all in compat
make[2]: Entering directory '/data/data/com.termux/files/home/cpuminer-multi/compat'
make[3]: Entering directory '/data/data/com.termux/files/home/cpuminer-multi/compat'
make[3]: Nothing to be done for 'all-am'.
make[3]: Leaving directory '/data/data/com.termux/files/home/cpuminer-multi/compat'
make[2]: Leaving directory '/data/data/com.termux/files/home/cpuminer-multi/compat'
make[2]: Entering directory '/data/data/com.termux/files/home/cpuminer-multi'
gcc -DHAVE_CONFIG_H -I. -fno-strict-aliasing -I. -Wno-pointer-sign -Wno-pointer-to-int-cast -D__arm__ -g -O2 -MT algo/cpuminer-rainforest.o -MD -MP -MF algo/.deps/cpuminer-rainforest.Tpo -c -o algo/cpuminer-rainforest.o `test -f 'algo/rainforest.c' || echo './'`algo/rainforest.c
algo/rainforest.c:562:16: error: unknown token in expression
  asm volatile("stp %0, %1, [%2,#%3]\n\t"
   ^
:1:21: note: instantiated into assembly here
   stp x12, x11, [x9,##0]
   ^
algo/rainforest.c:562:16: error: unknown token in expression
  asm volatile("stp %0, %1, [%2,#%3]\n\t"
   ^
:1:21: note: instantiated into assembly here
   stp x11, x10, [x9,##16]
   ^
algo/rainforest.c:562:16: error: unknown token in expression
  asm volatile("stp %0, %1, [%2,#%3]\n\t"
   ^
:1:21: note: instantiated into assembly here
   stp x10, x12, [x9,##32]
   ^
algo/rainforest.c:562:16: error: unknown token in expression
  asm volatile("stp %0, %1, [%2,#%3]\n\t"
   ^
:1:21: note: instantiated into assembly here
   stp x12, x11, [x9,##48]
   ^
algo/rainforest.c:562:16: error: unknown token in expression
  asm volatile("stp %0, %1, [%2,#%3]\n\t"
   ^
:1:21: note: instantiated into assembly here
   stp x14, x10, [x9,##64]
   ^
algo/rainforest.c:562:16: error: unknown token in expression
  asm volatile("stp %0, %1, [%2,#%3]\n\t"
   ^
:1:21: note: instantiated into assembly here
   stp x13, x12, [x9,##80]
   ^
algo/rainforest.c:562:16: error: unknown token in expression
  asm volatile("stp %0, %1, [%2,#%3]\n\t"
   ^
:1:21: note: instantiated into assembly here
   stp x11, x14, [x9,##96]
   ^
algo/rainforest.c:562:16: error: unknown token in expression
  asm volatile("stp %0, %1, [%2,#%3]\n\t"
   ^
:1:21: note: instantiated into assembly here
   stp x10, x13, [x9,##112]
   ^
8 errors generated.
make[2]: *** [Makefile:2435: algo/cpuminer-rainforest.o] Error 1
make[2]: Leaving directory '/data/data/com.termux/files/home/cpuminer-multi'
make[1]: *** [Makefile:2920: all-recursive] Error 1
make[1]: Leaving directory '/data/data/com.termux/files/home/cpuminer-multi'
make: *** [Makefile:699: all] Error 2
```

JayDDee

full member

Activity: 1424

Merit: 225

cpuminer-opt-3.15.3

Yescrypt algos now use yespower v0.5, a little faster.
New implementation of sha256 using SHA CPU extension.
Replace Openssl with SPH for sha256 & sha512.
AVX512 optimization for sha256t & sha256q.
Faster sha256t, sha256q, x21s, x22i & x25x on CPUs with SHA without AVX512.
AVX512+SHA build for Intel Rocketlake added to Windows binary package.

https://github.com/JayDDee/cpuminer-opt/releases

JayDDee

full member

Activity: 1424

Merit: 225

cpuminer-opt-3.15.2

Zen3 AVX2+VAES optimization for x16*, x17, sonoa, xevan, x21s, x22i, x25x, allium.
Zen3 build added to Windows binary package.

https://github.com/JayDDee/cpuminer-opt/releases

bobben2

full member

Activity: 279

Merit: 104

Quote from: adamvp on November 11, 2020, 03:09:30 PM

Is there any coin worth cpu mining right now?

That depends on a couple of things (at least):
1. CPU type. Generally, Ryzen cpus will be a better choice for mining these days, especially those with higher core counts and larger L3 cache.
2. Your strategy. You can mine a coin believing it will moon one day in the future. Then you might not mind it not returning more than a few cents a day now. You can let the miner do its work and forget about it until the day the coin moons and then rake in the profits selling.
Or, you can spend a lot of time researching the coin of the day and mine it short term to see if that will earn you a quick buck.
I think XMR has potential to go higher in price and am mining it.

miner29

full member

Activity: 1275

Merit: 141

XMR

adamvp

hero member

Activity: 1246

Merit: 708

Is there any coin worth cpu mining right now?

JayDDee

full member

Activity: 1424

Merit: 225

Quote from: nsummy on November 10, 2020, 03:10:28 PM

I have a feature/documentation request. I think it would be good to document which algos can take advantage of some of these newer CPU instruction sets. I've been mostly a GPU miner but CPU mining really intrigues me and would like to do it as a side project. I also think documenting which algos are no longer "supported" would be beneficial, or somehow segregating them from the rest. Just looking at the algo list its apparent that most of the included algos will never be seriously mined by a CPU. I definitely appreciate the work though, I have been using cpuminer-opt off and on for years now

The short answer is any algo derivative of x11 or containing parts of x11, otherwise known as sha3 based.
The irony is that the algos that can use AVX512 etc can't compete with GPUs.

The long answer requires searching through the code of each algo.

I'm not sure what you mean by no longer supported. I focus more on newer CPUs and algos that are still mineable
with a CPU (although difficult for some) but all algos are still supported.

JayDDee

full member

Activity: 1424

Merit: 225

Quote from: sech1 on November 10, 2020, 02:16:01 PM

Quote from: JayDDee on November 10, 2020, 12:10:47 PM

RandomX is a little different. It can benefit proportionaly more with VAES512. Some of the AES sequences
alternate AESENC with AESDEC so they can't be paired. VAES512 can still provide a near 2x improvement
in the AES performance, whlile the pure AESENC or AESDEC sequences get near 4x. I don't know how much
AES factors in the performance of RandomX as a whole.

RandomX is limited by AES instruction latency. Main AES loop has 8 128-bit AES instructions and runs in 4 clock cycles per iteration on Ryzen. With VAES it's 4 256-bit AES instructions but still 4 clock cycles per iteration. It can't be parallelized because each iteration depends on the previous one. AESENC/AESDEC interleaving can be worked around with some clever use of _mm256_permute2x128_si256().

The extra permutes would kill the advantage of 2 way parallel AES, but yes it can be done,
4 way parallel (avx512) might overcome the penalty.

I looked at RandomX VAES a while ago but couldn't figure out how to enable AVX512 to compile with cmake. I'm not good with c++ either.

I'm playing with AVX2+VAES on my Icelake laptop, it looks like x17 with get a 7% boost by using VAES for groestl, Shavite & Echo.
I assume similar for Zen3.

If things work out the next release may include a zen3 build with VAES in addition to AVX2 & SHA.

Topic: [ANN] cpuminer-opt v3.14.2, open source optimized multi-algo CPU miner - page 12. (Read 10289 times)