So what would be the recommended PSU wattage for two Sapphire Radeon HD 7970s? I'm guessing around 750w-860w?
it depends on the cpu but should be enough.
|
|
|
A new opportunity may have come. Is anyone working on an argon2 miner? It just showed up on hashpower at the top of their profitability.
If I could code cuda I'd be all over it.
I'm over it since yesterday. Currently I'm getting about 1 milliBTC per cpu/day. The usual cpu mining non-profit ;-) About the algo: most of the time is spent doing standard scrypt-jane, I don't think there is a lot to optimise.
|
|
|
LINUX OPEN SOURCE ALGO SWITCHER--
If somebody good enough at coding could create a Linux algo-switcher that works as well as NiceHash Miner, they would have a personal product of value.
NiceHash Miner uses tthe .Net environment that is Windows oriented. There Are .Net implementations in Linux now, but they require some skill to configure. If anyone could even come up with a .Net installation tutorial for Linux that is mining orienteed, they'd have something. The .Net 2.0 environment is required to run NiceHash Miner, I tred and failed to configure Wine to do the job. --scryptr
i'm using an on the fly coded bash script for that, it works but it's very messy and it will unlikely fit for someone without bash coding skills... if i manage to clean it up a little i'll let you know... Thank You Skunk-- I have tried with the version of your script that you posted earlier. I did get it to mine, but not in a stable, reliable fashion. Several of the pools that it was customized for shut down. I discovered while toying with your script that "BASH" and "SH" were not equivalent commands, by the way. I'll try to work with your original scrypt again. I tried before to trim out the inactive pools and damaged the script. A simplified version for NiceHash only would be a really attractive piece of code. --scryptr in the mean time i've added a lot of features to the script and some of them even revealed useless, therefore it became bloated and bug prone... i'll rewrite it from scratch beginning with nicehash switching and slowly re-adding profit switching pools (there's just hashpower atm), local wallets switching and so on... but please, don't hold your breath, i'm very busy and just mining with my desktop's gpu, therefore it will be a free spare time task. please put you effort on git or svn so we can pull request it ;-)
|
|
|
This has been discussed before.
Without rewrites you get gpl violations.
No one cares about GPL violations in the crypto world. As long as you never form a legitimate business that makes million you'll never be worth suing. No one is going to send the coppers after you. It happens allll the time in the crypto world as well. Generally speaking it happens a LOT in the real world and usually nothing is ever done because the people who make GPL code don't have the resources to sue someone. If this is your shiny suit of armor you're wearing things aren't going to go anywhere. If the 'business' ever gets big enough, you'll be able to get someone to rewrite everything that you haven't yet and pay them BTC. I'm not saying it's a great idea to break 'laws', but in this case it really doesn't matter and no one gives two flips about it. That argument was just a easy way out to avoid actually doing something meaningful here. for info GPL licence doesn't mean 'free', you can request payment for a software even if it is published within GPL licence (see the definition of the licence). The GPL licence only means that the source has to be available (as far as I have understood) hmmmm well... "available for free": if you sell a private miner to a single person, he can request the sources and he has the rights to re-distribute them for free. business around opensource works in a different way than that.
|
|
|
This has been discussed before.
Without rewrites you get gpl violations.
No one cares about GPL violations in the crypto world. As long as you never form a legitimate business that makes million you'll never be worth suing. No one is going to send the coppers after you. It happens allll the time in the crypto world as well. Generally speaking it happens a LOT in the real world and usually nothing is ever done because the people who make GPL code don't have the resources to sue someone. If this is your shiny suit of armor you're wearing things aren't going to go anywhere. If the 'business' ever gets big enough, you'll be able to get someone to rewrite everything that you haven't yet and pay them BTC. I'm not saying it's a great idea to break 'laws', but in this case it really doesn't matter and no one gives two flips about it. That argument was just a easy way out to avoid actually doing something meaningful here. I've been an opensource supporter for more than 20 years. I don't break the GPL, I use and support it.
|
|
|
I see the level of blah blah is reaching record highs! I don't understand why sp_ and djm34 are "competing"... there is so little to compete for :-) Especially if you consider every developer is saying "I have a full time job giving me 1000 times what I get for miner coding". Maybe it's just for fun... if so, then go on, that's the spirit ;-)
|
|
|
it is - but it would MUCH more difficult for the devs to supply a compiled version ( even as a static build ) to work will most linux distributions ...
im very specific with what i have int he farm - fedora 23 x64 cuda 7.5 with all the latest updates ...
so if a miner can be compiled by the devs - that would be great AND i would pay another 50% more for their trouble ...
#crysx
it's not if you staticize the binary with ermine, the trial version works good enough... ![Wink](https://bitcointalk.org/Smileys/default/wink.gif) Nice, but it makes linux more like windows ![Cheesy](https://bitcointalk.org/Smileys/default/cheesy.gif) (i.e. distributing the shared libraries with every program makes them... well... no longer shared)
|
|
|
Actually it is possible to compile a linux version there is no problem in that (I did it already for the first neoscrypt miner). The problem is that there is many differents version of linux, and you can't get away with only one linux exe... hence what might work for some linux brand won't work for other (example, fedora and ubuntu won't necessarily work with a same exe... however it comes at some point to the packages which are used... compiler gcc/cuda)
So basically it is a pain to create a linux exe and be sure it works for all linux OS
If you make it work on the latest ubuntu and fedora you are ok with most users. And if it doesn't work you can always refund. Or, you can ask a higher price and compile for that specific setup.
|
|
|
"We are short of the Bitcoin (BTC) currency. Please select another one for payments until we find more blocks."
I thought that was impossible with BTC, as they come from exchange, not mining...
|
|
|
So what is the best miner for x11 right now and what speeds could one expect?
Not strictly on topic is it? :-) And are you talking about cpu mining? In that case, forget x11.
|
|
|
I added aome more quark hash on compute 5.2 devices(@github) +200KHASH(g1 970)
Thanks! Maybe those jh and keccak improvements are good for X11 as well.
|
|
|
Djm34, does this version work fast on cuda 7.5? How does it compare to itself compiled on 6.5? Thanks.
|
|
|
Folding@EVGA http://forums.evga.com/tt.aspx?forumid=28 will get you more money quicker then any algo, $10. 00 for 2 days work for me with my hardware, takes 5 to 6 days with the same hardware to make $10 mining the top paying algo. ;-) I do folding at home for 2 days each month for the 10 EVGA Buck to go towards a new video card. Those EVGA buck bought me my new GTX960 GTX760 and 2x GTX660Ti and other video cards over the years. Sure you have to save them all year to really have anything since its a once a month thing and 10 buck max, but it still makes a nice Christmas gift for yourself. http://folding.extremeoverclocking.com/user_summary.php?s=&u=610174 Are you "merge mining" foldingcoin, curecoin etc.?
|
|
|
Does the Fury supports newer instructions than the 7970 or 290 seriers?
7970 = GCN 1.0 290 = GCN 1.1 FURY = GCN 1.2 I can tell you from my experience that the opencl compiler is much smarter on 1.1 than 1.0, currently almost the same as 1.2. That means it's much easier to optimise for hawaii than for tahiti. But that's just a small part of the equation.
|
|
|
Prices seem a little high, but the savings on power and cooling might make them worth the price?
Not worth it. You can buy a 7990 for $300 and undervolt and under clock it. You have higher efficiency and more mining hash. True but you will have bought an old technology, old card, plus, i'm not so sure about your statement though. His statement belies the fact that Southern Islands aka 7xxx and all their rebadges (R9 2xx/3xx) is still the best hash/watt/$$ you can get on the AMD side. The only new chips AMD has put out at a more efficient nm process are the Fury cards, and their mere cost alone removes them from hash/watt/$$ considerations. they are still 0.28 nm: AMD just added some CUs and changed memory interface.
|
|
|
HI all
I never followed the original thread, but i'm interested in how the mining side of things works...
How do you keep the mining fair between each different Algos? i mean is there a factor involved that adjust the difficulty between each different Algo? e.g. if i was mining using NIST5 would i find a block quicker than if i was using X15? as there are 3x more rounds of hashing involved...
Each algo has its own difficulty retarget. So hashrate between NIST5 and X15 doesn't matter, your earnings are based on how much your hashrate compared to every else's hashrate is (total network hashrate). ...compared to everyone else mining on the same algo. the chains are independent. i.e. if someone is mining on x11 at 1 Th/s, you can still get a lot of coins by mining x13 (if it has low diff).
|
|
|
As far as I understand (please correct me, djm34, if I'm wrong), the 50% gain is based on the latest opensource kernel by djm34, NOT the one on the sp_ fork, which includes about 10% improvement by sp_ and another 10% by myself. So the gain at the end should be about 30%.
don't know for sure actually, if you can tell me what is the hashrate of the opensource code (or which sp version to use, I can do a comparison, I think sp change on blake mixing was incorporated before I gave a quote) with SP_ ver .74 my 970's get 563 with his ver 7.7 I get 368 Kh/s on neoscrypt ![Grin](https://bitcointalk.org/Smileys/default/grin.gif) ok so it is even more than 50% (I don't have a 970, I only have 980's) Hmmmm no.... your version should do 650-700 while mine does 560 on the 970.
|
|
|
Thanks for the new thread. Please add a link to the bittrex exchange page.
|
|
|
As far as I understand (please correct me, djm34, if I'm wrong), the 50% gain is based on the latest opensource kernel by djm34, NOT the one on the sp_ fork, which includes about 10% improvement by sp_ and another 10% by myself. So the gain at the end should be about 30%.
|
|
|
guys the 7990 is just two 7970, so double the speed and almost double power usage. and much more than double cost to purchase.
|
|
|
|