Added another 5% or so on all my rigs with older cards using the -i 2 switch.
On my Rigs with newer cards, such as the RX series, I actually seen a slight decrease versus version 2.0. I have tried all three intensity levels and even the -i 2 gives a bit less sols and the CPU usage is greater than in version 2.0. I am still playing around as it might have been my lowered clocks, but it seems this version was geared more toward the older cards.
Yes I did not have time to improve performance on Polaris yet. |
|
|
|
|
v2.1:
- added "-i" option for adjusting mining intensity and CPU usage.
- improved speed for high intensity mode ("-i 2"), about 60H/s on stock 390X, 50H/s on stock 280X.
- since nicehash cannot upgrade their ddos protection, now "-allpools 1" is required for this pool.
PS. OpenCL API uses CPU to wait response from GPU in high intensity mode so it's not easy to fix this issue completely, currently you can use "-i 0" option if you have low-end CPU.
|
|
|
|
dev fee not wokring,then it stops altogether
Claymore im sorry to my bad english but i hope you understand what i say. On nicehash btc pool when i use your miner for mining zcash i have many disconection. I ask nicehash support and they told me what you miner mining same time for your support on one ip adress. So you can use other pool for minig for you support or do option where user can disable it?. Because, because its problem appear when your miner try to mining for your support at one btc adress. Nicehash not allow it. Can you solve this problem? i belive we can find way mining for your support without conflict nicehash.
https://bitcointalk.org/index.php?topic=1670733.msg16787287;topicseen#msg16787287In a few hours I will release an update with reduced CPU usage and this version will require "-allpools 1" for nicehash. |
|
|
|
CPU over 85% hash rate like past and a little lower I think V1.0 Beta better than other versions .  |
|
|
|
I had to remove v1.0 and v1.1 from google drive and mega, because I was pointed that I used a small part of mrb's work.
It will take up to 24 hours to revise my code and create own implementation, then I will upload updated version.Details are: mrb> Claymore, you took and reused my blake.c code and didn't include the copyright notice as required by silentarmy's MIT license:
mrb> https://forum.z.cash/t/silentarmy-v3-now-a-full-miner-multi-gpu-stratum/3580/640
mrb> I request that you cease infringing by:
mrb> (1) either including the full "License" section of silentarmy's README.md along with your distribution of your zcash miner,
mrb> (2) or stop using my blake.c code.
mrb>
mrb> Of course this is only ~100 lines of code, and we are all lazy programmers and I understand why you did it :) but you ought to do the right thing.
mrb> Thanks,
mrb> -Marc Bevand
[/quote]
Oh, you are correct, I really missed this point. I was in a big hurry to release first version asap, so I concentrated on main Wagner's algo,
blake is not interesting for me at this point since it takes only 10% of total GPU time and it's trivial in fact, blake algo is not that I can call something new.
But to make some start point asap I included your blake.c into my sources and then forgot about it and did not create my implementation. That's what happens if you are in a hurry.
It is very unpleasant situation for me, I will do the following:
1. I remove v1.0 and v1.1 versions of my miner. In fact, I already removed current packages from google drive and MEGA, I will upload updated versions soon.
In next version I will remove your blake.c from my sources.
2. I apologize, it's only my fault that I used this part of your code. I was not going to do it in public version.
3. I want to send you some compensation from my ZEC dev fee if you don't mind, just send me your ZEC address.
4. I'm not going to hide my fault in any way, I will post this message in my miner thread in first post.
Thank you for doing the right thing. I fetched your new tarball Claymore's ZCash AMD GPU Miner v1.1 Beta - LINUX - Catalyst 15.12.tar.gz and can confirm you included the MIT license as mandated in your "Readme!!!.txt" file, as quoted below. You are in no way obligated to "pay" me. But if you feel this can help re-establish some of your image and reputation, feel free to publicly send to my donation address t1cVviFvgJinQ4w3C2m2CfRxgP5DnHYaoFC You are now fully compliant with the MIT license. I hope my SILENTARMY code was useful to you  ==========
This miner uses some code of SILENTARMY OpenCL GPU Zcash Equihash solver:
# License
The MIT License (MIT)
Copyright (c) 2016 Marc Bevand
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"),
to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense,
and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
-Marc Bevand I'm happy that we close this issue! I also donated 5 ZEC (about 2BTC at current prices) to your address t1cVviFvgJinQ4w3C2m2CfRxgP5DnHYaoFC |
|
|
|
Try silentarmy versus this - at least on Linux, it seems... the same. So I poked around and pulled the IL and ISA from Claymore's.... the shit IS basically silentarmy, from what I can see. I didn't check every kernel, but it looks like it.
Have you implemented your own kernel for ZEC PoW? If it has good speed, it will look like silentarmy because there is no way to create really different implementation of fast binary tree for Wagner's algo, especially on GPU. But it does not mean that it is silentarmy code. Another example is blake implementation - implement it and its machine code will look similar for all possible implementations because it requires a concrete set and order of operations. |
|
|
|
v2.0:
- improved speed by 5-15% (depends on card model).
About 55H/s on stock 390X, 52H/s on stock 390 and 48H/s on stock 280X.
Where is the Linux build? |
|
|
|
|
v2.0:
- improved speed by 5-15% (depends on card model).
About 55H/s on stock 390X, 52H/s on stock 390 and 48H/s on stock 280X.
|
|
|
|
|
FILES UPDATED TO INCLUDE MIT LICENSE DETAILS IN README.
People started asking me about miner package in PM and some people started to share alternate links for downloading, it's not a good thing because of possible viruses.
So according MIT license I added notice to README and uploaded updated miner packages back to original links. You can also see a license notice in original post of this thread.
|
|
|
|
@Claymore your miner is having a real strange behavior, I was experiencing a huge amount of duplicate shares, so we started some debugging to investigate what is happening, and I see this happening on all miners:
....
The Rig is running your v1.0
any idea why is this happening ?
v1.0 has problems with nonce after devfee mining until new main job is received. In v1.0 every hour after devfee mining nonce was reset to zero and miner checked same nonce range again until new job is received. So it could lose about 1-2% of shares. Fixed in v1.1, see history.txt or OP: - fixed issue with rejected shares after devfee mining. |
|
|
|
I had to remove v1.0 and v1.1 from google drive and mega, because I was pointed that I used a small part of mrb's work.
It will take up to 24 hours to revise my code and create own implementation, then I will upload updated version.Details are: mrb> Claymore, you took and reused my blake.c code and didn't include the copyright notice as required by silentarmy's MIT license:
mrb> https://forum.z.cash/t/silentarmy-v3-now-a-full-miner-multi-gpu-stratum/3580/640
mrb> I request that you cease infringing by:
mrb> (1) either including the full "License" section of silentarmy's README.md along with your distribution of your zcash miner,
mrb> (2) or stop using my blake.c code.
Of course this is only ~100 lines of code, and we are all lazy programmers and I understand why you did it :) but you ought to do the right thing.
Thanks,
-Marc Bevand
[/quote]
Oh, you are correct, I really missed this point. I was in a big hurry to release first version asap, so I concentrated on main Wagner's algo,
blake is not interesting for me at this point since it takes only 10% of total GPU time and it's trivial in fact, blake algo is not that I can call something new.
But to make some start point asap I included your blake.c into my sources and then forgot about it and did not create my implementation. That's what happens if you are in a hurry.
It is very unpleasant situation for me, I will do the following:
1. I remove v1.0 and v1.1 versions of my miner. In fact, I already removed current packages from google drive and MEGA, I will upload updated versions soon.
In next version I will remove your blake.c from my sources.
2. I apologize, it's only my fault that I used this part of your code. I was not going to do it in public version.
3. I want to send you some compensation from my ZEC dev fee if you don't mind, just send me your ZEC address.
4. I'm not going to hide my fault in any way, I will post this message in my miner thread in first post.
|
|
|
|
@claymore
i try your miner with 2 instances, it work fine, same hashrate.
but this give me a idea :
can you make a flag start for choose a config file pls ?
by default config.txt and if i put "-conf config-coinmine.txt" or "-conf config-nicehash.txt" zecminer use my config file.
Read "CONFIGURATION FILE" section of readme or OP to see how it works. |
|
|
|
Part of DevFee problem solved - added -allpools 1 - miner changed DevFee and problem removed, Claymore - i think what error in your wallet on nicehash, miner gives error because not working on nicehash without workername after the wallet, maybe you forgot to write your workername in the source Everything is ok with the miner, I have checked equihash.usa.nicehash.com:3357 many times, it works for me and for most of people. But nicehash has paranoid ddos protection, I think it's the reason. I see same thing for my XMR miner when they block devfee connections sometimes. I asked them if they can help somehow, but they don't want to change anything. So if you have this problem just use "-allpools 1". |
|
|
|
i can confirm that its working faster and for now stable  Also id there will be option with dual mining ? I heard that we can add ethereum as second miner and it will be more profit overall. I want to improve speed of this miner firstly. Anyway, ZEC is not very good for dual mining. |
|
|
|
Since flypool is stabile enough i didn't set epools file for failover pools but i decided to do that just to guarantee mining nonstop. After that, i get this error in red text at the beginning of miner start : "no address specified - ignore". It seems it's not a problem since i'm mining with no problem whatsoever. I checked my epools.txt, it seems in order but maybe i'm missing something. I'm just curious what creates that error?
Copy default epools.txt and remove all "#" characters from it to enable sample pools in it, do not change anything else, we just want to test. Start miner with default "config.txt", it will write at the beginning "ZEC: 5 pools specified". Four in epools.txt and fifth (main) pool in config.txt. If you see red text about ignoring pool it means that you specified it incorrectly in epools.txt and you need to see sample pools in epools.txt as example and may be read Readme. |
|
|
|
The new 1.1 gives me 100% rejected shares on suprnova havent tested any other pools but since i dont need any of the features of 1.1 and i like suprnova.
May be wrong credentials in your command line or config.txt? For such issues I need to see the log file. The version 1.1 beta reports the total speed as H/s instead of Sol/s like version 1.0. However, actually it is Sol/s not H/s since when I check my speed on my nicehash page it says so. Then, there is a typo needs to be fixed. As everybody involved in Zcash knows that 1 H/s equals 2 Sol/s.
Both v1.0 and v1.1 report speed in H/s. Miner does not send any hashrate info to pool. Your pool may calculate speed in Sol/s, but it is not related to the miner. Again: all versions of the miner show speed in H/s. |
|
|
|
|
v1.1:
- added Linux version.
- fixed issue with rejected shares after devfee mining.
- remove "-allpools 1" option if you use it and mine on miningpoolhub, suprnova, coinmine or nicehash pools.
- added "XnSub" support for Nicehash.
- fixed bugs.
PS. Read OP or Readme for updated samples of using on different pools.
|
|
|
|
Thank you... Can we expect better speeds on the 470 and 480 cards in your next update? Also when I find shares it says (300ms and over) How can I make that faster? Thanks Next update is bug-fixes only. 300ms is a server response time, it depends on your internet connection, pool location and its quality. |
|
|
|
|
Show Posts
