Yes, this is what I do as well. I have quite a few rigs with only 4GB RAM and I just set the Virtual memory to 16 GB. I don't even think you need this much, but most of the excess capacity on a mining rig SSD goes to waste anyway, so why not.

I made a few more changes, namely I reduced the cvddc setting a bit more to 900 mv and decreased the core clocks from 1135 down to 1125 and was able to shave a bit more off my power consumption, which is closer to 102 watt/card now. I also tried Claymore 9.1 using -asm 1 and -dcri 8 and I can get 30.5 Mh/s per card solo mining with the above power draws. If I try dual mining with Decred my wattage shoots up past 120-125 watt/card.

If you use Claymore you can use -cclock, -mclock, -cvddc, and -mvddc parameters when you launch the miner. See the Readme file in the Claymore directory for usage information.

Since Wattool seems to not read values correctly on the newer cards, I been using the miner directly to control the clocks and voltages, GPU-z seems to confirm it is working as intended.

This is true, I got similar results as the OP, but my at-the-wall draw per card (6 rig system) was 108 watts after subtracting the PSU efficiency 92% and system idle draw of 50 watts. My GPU-z fluctuates around 80-85 watts, so I think the end results would be similar.

EDIT: To clarify I am getting 30 Mh/s per card (180 rig total) with core clocks set at 1135 and memory clocks set to 2275, core voltage set to 925 mv, memory voltage left at default. I did do the bios mod using the jaschaknack (sp?) PolarisBiosEditor by copying the 1750 memory strap (Samsung memory) from 1:1750 to 1:2000. I did try to copy the lower 1625 strap, but the system was unstable.

For full disclosure, I am also using the Crimson ReLive Edition 17.4.3 drivers on a Windows 10 64 bit system. Bonus, at least with the RX580 I did not have to apply the pixel patch after modding the BIOS.

My raw power usage data is:
Total rig draw: 760W (kill-a-watt reading)
PSU 92% efficient (8% inefficient):  760 - 60.8 = 699.2W (or just 760*0.92).
System idle draw (50 watts): 700 - 50 = 650W
Draw per card: 650 / 6 = ~108W

Yes, I can corroborate this as I have MSI RX580's with Samsung memory and had to use the 1:xxxx straps. I also seen a video with cards using Micron memory where the guy had to use to 2:xxxx, so I think the  1: is for Samsung and 2: is for Micron theory is correct.

I just got my MSI Armor RX580 8 GB cards in today and out of the box they do 25 Mhs with about 150 watt power draw from the wall. This is with the 17.4.3 ReLive drivers and Claymore 9.1 when mining Ethereum.

In Wattool they only show the memory clock as running 1000 MHz, but in the latest GPU-z they do show up correctly as 2000 Mhz. I am a bit hesitant to try to adjust things with Wattool if it isn't detecting the cards correctly, so I will probably just stick with experimenting with different settings using afterburner at first.

As far as heat is concerned, yes move them out of the house and to the garage. Trying to cool any amount of rigs inside your house with A/C will send you to the poor house in a hurry.

You will need to think a lot bigger than dryer vents. I have a lot of rigs in my garage and put in a screen door for the back service entrance, added a big window on the non-house side, and keep my main garage door open about a foot to keep airflow decent in the warmer months. I am thinking this year to add one or two of those bigger industrial type fans, Home Depot sells a 36" one for $99 to help move more air as I added a few more rigs over the winter. Even warm 90F degree air will keep your GPUs cool, but the key is good air flow to remove the excess heat from the rigs and garage.

For taxes, even if you keep it all in bitcoin, you might still want to claim enough in deductions to offset your electric bills and rig depreciation. In the US at least I believe there is a limit to how long you can claim deductions. So say you been mining Bitcoin, or Alts and converting it to BTC, for several years then try to sell them, the rules for depreciation may not let you go back that far. Same with electric bills, claim your excess electrical use each year.

I just ordered 5 of the MSI 8 GB RX580's that should be here by Friday. I have a rig already setup and ready to go with them and can post some results by then if no one else has. I didn't want to pay for next day shipping otherwise I would have had them sooner.

Just as an FYI, I just received in 5 more of these MSI Armors from NewEgg and the latest batch is all Hynix memory.

But of more interest is they went from a 8-pin PCIe connector down to a 6-pin connector. I do not know if this will translate into better performance power usage wise, but initial results show pretty much the same as the previous version. I will need to play around some more with the settings and BIOS to see. My other thought was that the original version probably didn't need the 8-pin's (150 watt) additional capacity so they simply downsized the connector with no other changes.

There is no part number differences that I can tell, so it still looks like it is flaovr of the week when ordering these from NewEgg as to which one you will get.

If you don't mind, could you expand upon the good power numbers portion?

I was thinking about picking one of these up for my gaming rig thinking I could mine ZEC during off-hours (when I am not gaming). However, I would like a little more information on the cost of operation. Since I wouldn't have the benefit of multiple GPU's to offset the system (mobo, cpu, etc.) wattage overhead, I need to work the numbers a bit to see if it would be a viable choice.

I have a mix of the MSI Armor cards and it's usually hit or miss if they will come with Hynix or Samsung memory (although my last few orders have been all Samsung). Anyway, what rjg says is correct about the Samsung doing better by using the 1750 straps, its the Hynix where you need to copy the 1500 strap.

As far as quality, I have several and have yet to encounter a bad card. I am fairly conservative with my overclocks however, and prefer to focus more on reducing power consumption so that my play into it.

Doubtful it is the miners, as most who are going to sell, sell small amounts daily and not accumulate 40-60 BCT worth of ZEC to just to place a sell wall as you describe a few pennies higher than the market rate. Much more likely the action of speculators and/or whales trying to manipulate the price.

Yes, as Tmdz mentioned the newer drivers check for the BIOS signatures. There is a workaround to this, but the older drivers work just as well. Since there is not a significant difference in hashrate, even when you get the newer drivers to work, its probably not worth the effort.

Also, you technically did not "brick" your card as that would mean you were unable to restore the BIOS at all, which would then basically render your GPU unusable or about as useful as an expensive brick, hence the term.

You simply attempted to modify the BIOS and then restored the original. I know, not as click-baity of a title, but then what do I know.

I would recommend looking into some sort of under-volting/underclocking if going this route. I have been testing dual mining on a 5xRX470 rig and power draw went from ~600 watts Eth only to ~800 watts ETH+DCR. I am using a modified vbios with -100 mv undervolt and not pushing the clocks too hard (1100/1850), so I would expect default settings to be a lot more power hungry.

What you ask really make no sense to me. Do you own these computers, if so why wouldn't you want to mine with them overnight instead when staff is not trying performing their essential job functions? Running a mining process in the background can degrade system performance, especially on minimally speced workstations that you would probably find in a setting such as a call center.

To answer your question, if you have administrative access to the workstations, you can do as felixbrucker suggested and utilize the task scheduler. Yes you can run for all users and run the process as hidden, a simple google search should give you all the help you need.

Hardwired LAN is always the best since it is running on fairly short lengths of wire designed to minimize interference and cross-talk and you will enjoy sub 1 ms pings (latency) inside of your network. This helps keep additional contributions to latency negligible when you communicate with a remote server. Hardwired is also the most reliable as network drops are going to be either the router, your rig, or ISP causing issues (rarer), not just random wifi drops (more common).

With that said, Wireless LAN can be useful in limited situations. If you want to spread a few rigs around the house to minimize heating one area too much, or for noise, wireless allows you to do this easier without having to run wires all over. However, there will be a slightly higher latency penalty, which depending on your setup and how much interference your wifi is encountering, could add from 5 to 20 ms on top of your normal router to server latency. As others have correctly pointed out, this can sometimes cause an slight increase in stales at the pool.

The big problem is when you start to increase your rig count. Each rig will increase the wireless or RF "noise floor" in your house and depending on a multitude of other factors, can quickly get to the point where wireless become unusable. You will also notice this with other wireless technologies such as remote controls for garage door openers no longer working properly. This combined with the open air nature of most dedicated mining rigs and lack of proper grounding/shielding to dissipate the RF being generated, will quickly saturate your house with RF noise.

There is no hard and fast rule for this, as it will depend on your current noise floor, how many other wireless devices you have in your home (not just wifi), proximity to neighbors with wifi and wireless devices, power lines, etc., but once you hit anywhere from 5-10 rigs on wireless, you will probably start to notice the effects.

TLDR: Use hardwired LAN if you can, but for a small number of rigs on wifi the small hit on profits probably won't be noticeable.

I use dummy plugs mainly to get a usable screen resolution when using remote management software such as VNC. Since my main workstation is a 4k screen, the default 640x480 or whatever is way to small and even some windows such as the Wattool do not display fully.

I use RX470/480's almost exclusively and all my rigs will boot without the dummy plugs, indeed I have only about half the number of plugs as I do rigs so I often have to swap them around.

This is incorrect, see Geenstijl's post above.

They only raised about 895 1,000 BTC during the crowdsale, or in other words sold 15,217,819 DataBits. This is still worth over $1 million USD, so they are going ahead with the project and feel they can make this amount work. They are also reducing the funds they are keeping for development down from 30 mil to only 6,521,922 DataBits. This new number keeps the development fund at the original goal of 30% of all DataBits outstanding, so the net effect is a new reduced max supply of 21,739,741 DataBits.

The remaining 78 million or so DataBits will supposedly be burned at some point, this is yet to be confirmed (if they have been burned) though.

They also received some investment capital from their appearance on Shark Tank South Africa according to their website, so they may be working with more than 1 million USD, but the tokens should remain as explained above. So thinking of this only from a cryptocurrency perspective, the new max coin supply is pretty much in-line with Bitcoins, so valuing this at only 5,000 satoshis, or 5 cents each seems rather low. I guess it depends on how big of an impact their game will make on the market. Myself, I see a lot of other alts based upon far less and they have much larger market valuations.


--EDIT-- I said 1,000 BTC above originally, but according to their twitter they actually raised 895 BTC. I am assuming they sold while BTC was $1,118 or higher, as they claim they still reached their $1 million USD goal. Just wanted to have all of the correct facts out there.



--EDIT 2-- Reading the quote more closely, it appears Vinny Lingham only invested 500,000 ZAR, or roughly 59 bitcoins (~$60,000), for 20% of the business. Relevant quote from the article below:


So it remains unclear if Vinny gets his cut now, or essentially takes $200,000 from the ICO, or if he is in it to the end. Either way, I am sure they needed the inital seed money to get this far, as clearly there is more development going on than just the DataBits portion. I am still excited about the project and its incorporation of a blockchain token.

It's all relative. The ICO price ranged from 6.67K to 10k each, so a first day trading price of roughly 2x the ICO price isn't unheard of. I am sure the price will fluctuate a lot in the coming days and even go down more, indeed it already has been as low as 13.6k on Bittrex.

But short-term price is nothing compared to the long-term potential. If they can deliver the game and it gains traction, even 22k will be considered low.

Sounds interesting, I actually first seen it on Bittrex this morning and came here to look into it. Buying all I can under 20 cents (~0.0002) as I think it has the potential to at least hit $1.00 at some future point, if not more. With the new ~21 mil coin cap, that is only around a $21 mil market cap, which is not too hard for an alt to achieve these days.

Now I need to check out the game aspect and see what the real potential is.