Topic: [ANN]: cpuminer-opt v3.8.8.1, open source optimized multi-algo CPU miner - page 138. (Read 444067 times)

That's not even released yet. If it uses a new version of gcc there might be some incompatilibities. Other users have
reported problems with gcc 5 or 6.

I use Mint 17, based on Ubuntu LTS, not sure which version with gcc 4.8.4. Post the compile errors and I'll have a look.

joblo which Linux Distro have you used to compile it?
I tried to compile it but the AES-NI module u have built jst wont compile the TEMPLATE FORMATION issue is given.

So ..grr none of the miners i wanted to test built under Ubuntu Yakkety :-(

That's an odd combination of features. What kind of CPU is it?

I'm not going to start building one-off binaries for unusual CPU architectures, not for free anyway.
I suggest you try compiling yourself. There are instructions in README.md in the source package.
If you have problems you can always ask for help, which is always free.

That address works for pool payouts. I don't recall setting a payment ID at the pool but my wallet is displaying a
payment id with that address. maybe you do need one...

a531ddb280b51d491e8a12717f4b76f0bdbe44826b5e3be3f5c693a6f6f26bc1

You did but it still gives this error:

Failed to decode address (#0): Invalid encoded length

If it is not difficult to make the assembly of a -MSSE2 -MAVX2 -MAVX please

This is a reminder to those asking for help to post their question in this forum. Please do not use PM.
An open forum allows all users to learn as well as offer help.

It also helps if you post details such as miner version, CPU, error messages etc. If I have to ask for
further information it will delay a solution.

And make sure you're actually using cpuminer-opt before posting here.

Happy mining.

I thought I posted it but maybe I didn't.  It shouldn't need a payment id. Thanks for your consideration.

XMR: 497qTtunbvPgUT6auQnLLacLPKNnxW9fADEYy1ugYYLXd5FuyCR3RwoT44hanwKYnyXXq8QtLTYAr1Q AKe3jCWPfN8oEK3W

Joblo,

I am still waiting xmr address that works. If on exchange need that payment id too.

Confirmed benchmark is broken in v3.4.6. I didn't think I did anything to affect it but will have to take a closer look.

Edit: found  the problem, simple change in cpu-miner.c line 1334. Simply comment out the zrero initialization and
uncomment the original code.

Change:


to:


I'm too lazy to recompile all the binaries. Normal mining works.

cpuminer-opt v3.4.6 is released

source: https://drive.google.com/file/d/0B0lVSGQYLJIZekg3QUNCcE1ZU2c/view?usp=sharing

Windows bins: https://drive.google.com/file/d/0B0lVSGQYLJIZYVI1Rm40d2xEXzA/view?usp=sharing

New in v3.4.6

For users:

Cryptolight algo is now supported with AES otimizations. The code was always present but untested until now.

The dsplay format for share submissions is changed:
    - colour keyed "Accepted" or "Rejected" status.
    - reject count and rate displayed when share is rejected.

For developers:

Base code for managing new work in the miner thread loop was rewritten due to confusing and incorrect implementation.
There was redundant code and closely related code that spanned different functions, and It was also not fully compliant with
the spec https://en.bitcoin.it/wiki/Stratum_mining_protocol with respect to the "clean" parameter.

    - removed call to stratum_gen_work from miner_thread (redundant).
    - eliminated gen_work_now gate function (merged with get_new_work).
    - renamed gate function init_nonce to get_new_work.
    - moved most gate targets from algo-gate.c to cpu-miner.c removing
      most mining related code from algo-gate-api.c.

Some restructuring was also done to firm up algo_gate. All references to scratchbuf, used by scrypt and drop algos, is moved
from cpu-miner.c to the respective algos' source files. This allows for the removal of the scratchbuf argument of the scanhash
function.

hey joblo i've tested your miner at my pc(OS is windows 10 pro Redstone 1 ),the pc is asus x552cl with i5 3337u: https://www.asus.com/Notebooks/X552CL/specifications/
miner works but it doesn't work properly with other mean it works too fast:http://prntscr.com/cewmxn
i hope you will solve it,keep up the good work !

edit:many of the algos same with this build including m7m,cryptonight,quark,qubit,keccak and i think much more

And now for something completely different, with apologies to Monty Python.

I just saw a rerun of the premiere episode of Halt and Catch Fire, a TV series set in the early days of PCs.
At the start they had a description of the meaning of the title, but it was one I had never heard before.
Their definition was supported by this Wiki page.

https://en.wikipedia.org/wiki/Halt_and_Catch_Fire

But my recollection is different. Halt is a legitimate CPU instruction which is intended to stop all processing
and require an external reset signal to recover. It is also supposed to put the CPU in a low power state.

But early implementations tried to save silicon by mapping the halt instruction to a jump to self instruction
with all interrupts masked. This was thought to emulate a halted state.

But, not only did this implementation not put the CPU in a low power state it focussed all power to a very small
area of the CPU eventually burning it out. It didn't actually catch fire but it did go up in smoke.

Anyone else have this understanding of the meaning?

Edit: there was mention of my version later in the arcticle.

Just to ensure I include some on-topic content, my brand new 1070 essentially halted and caught fire the other day.
I actually saw a flame shoot out of the case. Fortunately it didn't take anything else with it, got a replacenent on
the spot from a dealer and was back up and running.

Thanks for testing. There isn't much I can do if it's already using AES.

I tested the on a win7 PC and yes the AES build is faster than SSE2 with the cryptolight algo, on linux I only tested one build.

yes you can take out the warning

I guess that means it is working and I can take out the warning about it being untested? If your tests show the AES build is faster than
SSE2 I'll change the display to indicate the algo has AES.

I haven't look at the Arux-BTT fork yet, let me know how it performs compared with cpuminer-opt.

Thanks joblo will try see if there is a dif..
btw I just find this https://github.com/Arux-BTT/CPUMiner-Multi-cryptonight-light/releases/tag/04082015 [with AES-NI] but I wanna keep using your version of cpuminer. I can use it easily on linux VPS (I'm kind of a noob with linux cli..lol)

Cryptolight is not a very popular algo and until recently I didn't even know there was a coin using it.
Therefore it is very low on my priority list. I haven't analysed the code so I don't know if the
cryptonight optimizations can be ported for use on cryptolight. If they can it might be possible for
me to implement it. Otherwise you will need to pay someone like Wolf to build one for you.

I am pretty busy keeping up with new algos, while trying to opitimize the more popular
existing algos. It is unlikely I will get to cryptolight any time soon.

Edit: I took a quick look at cryptolight and it appears to already have AES optimizartions implemented.
You can confirm this by testing on a CPU with AES_NI using the SSE2 build of cpuminer. If it is slower
than the AES version for your CPU then you are already using AES_NI.

Thanks for testing. cpuminer-opt only supports stratum mining.

I am curious about your performance results. Can you provide more info such as whether you compiled you own, Linux or Windows,
or used precompiled binaries. Thanks.