Search content
Sort by
Showing 20 of 61 results by alexis78
Night tests results:
As sp said, i used default intensity (28) and --gputhreads 2:
spmod skein: miner hashrate 5900-6200, pool hashrate - 4300-5300
Also i tried to launch Alexis miner with 2 threads per gpu -d 0,0,1,1 etc... and 3% less hashrate and without fees
sp, i know what your kernels speed always close to pool speed, but skein - need to fix it.
Aha-ha all sp "optimisations" of skein are just options setting.As sp said, i used default intensity (28) and --gputhreads 2:
spmod skein: miner hashrate 5900-6200, pool hashrate - 4300-5300
Also i tried to launch Alexis miner with 2 threads per gpu -d 0,0,1,1 etc... and 3% less hashrate and without fees
sp, i know what your kernels speed always close to pool speed, but skein - need to fix it.
Now I get 440 from 1060 (2200 from 5*1060) with Alexis78
Thank you
So, the real switch in Alexis miner to get these results is "-t 10" (for 5 cards) or "-d 0,0,1,1,2,2,3,3,4,4"?
Thanks.
Beware of pool side hashrate if you try it... It has failed on my end.
Mr. Tanguy also reported a failure on utilizing multiple host threads per single device on a chat we had, long time ago.
The failure we encountered is the same as above. Bigger hashrate miner side. Lower hashrate on pool.
Sp, how much faster is this?
I get around 330 with the slow alexis miner (80% tdp), so that's around +20%
The slow alexis miner is cuda7.5 for maxwell. I'm sure when cuda9 gets out of RC, everyone will replicate the increases easily
Here a preview.
evga 1060 6g 80% tdp and oc.
B/A/R
Blocks found / Accepted shares / Rejected shares
evga 1060 6g 80% tdp and oc.
B/A/R
Blocks found / Accepted shares / Rejected shares
You forgot to change something tiny... the values are still Solved/Accepted/Total
Also, looks like you removed the cpu validation since the diff of each share cant be 0.000 with stratum's diff at 0.062
Nice touch, but still gpled btw
2 weeks have past, and still nobody has managed to create a faster opensource signatum kernel based on your work.
And none will be.But I am not distrubuting my private kernels. Beta version are sendt out for free to my donators for testing purposes and feedback. I don't see where this break with the licensing..
And you are not one of my donators, so what are you bitching about?
2 weeks have past, and still nobody has managed to create a faster opensource signatum kernel based on your work. Stop trolling and show some skills.
I got lbryspmod6 from github. Can i see the source?And you are not one of my donators, so what are you bitching about?
2 weeks have past, and still nobody has managed to create a faster opensource signatum kernel based on your work. Stop trolling and show some skills.
Did you supply them with the source code? 
The ptx assembly is in the bin. You just need to extract. I left it wide open.
As for the leaked binary you have to ask the people who leaked my bins. They are violating the GPL. not me
Did you supply them with the source code? I was about to release signatum #6 with 35MHASH on the gtx 1070. But the release will be delayed.
will you release the obsolete (and leaked) signatum#1 to #5 though? from what i understand you would like to know how sp_ managed to get the miner as fast as he got.
So, if i release a faster of what he got, binary, to prove that i dont really care what he did. You think he will release the sources? I think not.The subject is the gpl violation, not what he did
The point is that I am not violating any GPL licence. Your Git has also been empty for almost a year.
I am still active in the ccminer development, and you are not.
You are because you dont provide sources when asked.I am still active in the ccminer development, and you are not.
Please show us your lbrymodsp6 and sigt
You are not active, nothing of the above points to active. Just copy pastes for a year
But I am not distrubuting my private kernels. Beta version are sendt out to my donators for testing purposes and feedback. I don't see where this break with the licensing..
A 3 year old "beta testing" And the only thing that saw the light on your open-source fork last 1 year are:
my fixes on skein
an empty lbry scanhash function
an updated fugue from klausT
an updated neoscrypt kernel from DJM
ok, keep up the good work
And why didn't you do it yourself?
I stopped open-sourcing because (...)I am not talking about opensourcing. I am talking about distributing of private binaries with new algos and more speed that you have made youself. If you do that, people might want to donate to you if they get a boost in their mining income.
Anyway, we have different perception of this so called community project...
But you still sell GPL licensed code, without providing sources
The best optimizer in the world playing dumb
I dont know, you tell, how can it?
I dont know, you tell, how can it?
My work is in all of your kernels, but the credit is not there. Epsylon copied my work from my fork, and then you fork him. Now I fork you.
My fork can mine Signatum, yours cannot.
Any tuning you've done on any kernels i used has the appropriate credits on the header, sometimes even when i had to clean your work to tune it better like the cuda_quark_skein hashfunction.
I dont see you forking my work on github, like others did and i have absolutely no problem with them.
I see you selling my fork
Private binaries of my fork, not yours.
Your fork cannot mine signatum right? So how can it be your fork?
The best optimizer in the world playing dumb
I dont know, you tell, how can it?
I don't sell anything. People donate for my work in the ccminer opensource project, and I give them access to private binaries.
Private binaries of my fork, not yours.
Your latest commit was in January about skein, and it had the fixes i've found regarding uint2 addition, and also the utilization of lop3.b32 ptx which i've introduced in keccak and used partially in skein and JH
And why didn't you do it yourself? Why do you bitch in my thread? It's almost been 1 year since your last commit to the ccminer opensource project.
When I run your ccminer with ccminer -skunk
It doesn't work does it?
No, because it is missing the scanhash function... (the simple thing which as we said can be applied with a copy/paste from another algo)When I run your ccminer with ccminer -skunk
It doesn't work does it?
I "bitch" in your thread because you sell GPL Licensed code without anyones of the actual developers consent.
I stopped open-sourcing because you sell everything i, and everyone else, contributes to the community. You do still sell my fork for the past 8-9months, right? I wonder what made you move from selling epsylon's work to selling mine, could it be the kernels?
You release nothing. You copy every algo that Epsy has released that you put in "your" miner.
I relased Skunk sp-mod #1 before there was an opensource ccminer implementation of skunk. My release is not even based on the same repo as Epsylons fork. My release was very profitable for the donators. The first buyers should have roi'd many times already.
Actually, since your "release" (private selling is not a release) is from my publicly available fork, it is based on the same repo as Epsylon's fork.
And for skunk, all that it was missing was a scanhash function, which anyone could reproduce by copy/pasting another simple algo, pretty straight forward procedure.
All the kernels are/were there.
Most of alexis work is based on my opensource optimized maxwell ccminer kernels. I have more than 1000 commits to github. Don't forget this... Anyway, the unreleased nist5 private is doing 55mhash. Ccminer 2.1 45 mhash and alexis 1.0.0 48mhash.
Gtx 1070 same clocks and launch config.
So... Why you sell my fork and not yours then? I've just 1 commit... I'm sure you can do yours 1001 commits and merge those "tiny" differences between our forksGtx 1070 same clocks and launch config.
do we have a deadline till sp#4 SIGT will be finished?
Since my kernel is stolen and resold, I have lost motivation to improve it further.The total hashrate is affected the first 2-5 minutes of mining because the gpu's doesn't start to work. If you open gpu-z you can see that only some of the gpu's are working. This isn't a problem if you mine at the same pool for a long period of time. (If all the 6 gpu's have started).
But it is a problem when you use it with a profit switcher software like the Nemosminer, minercontrol, nicehash(3rd party) or similar software.
But it is a problem when you use it with a profit switcher software like the Nemosminer, minercontrol, nicehash(3rd party) or similar software.
Kinda worked around it by starting 4 miner instances with 2 gpu per instance. I don't think that issue affected the total hashrate though, just an inconvenience.