Hi Claymore, thanks for this first update to XMR GPU miner. Since this update obviously focused solely on bringing functionality inline with your other miners, is it safe to say that your next step on this miner will be to try to increase speed for recent GPUs like rx 470/480s?  Hopefully, now that you released v10 of the ZEC miner, you will come back to this one to help those of us with 470/480s 

Phil, that's a good price for the 550w.  I've been watching the 1200w prices for over a month and they are ridiculously high. Amazon had it for $160 for a while but now it can't be found anywhere for less than $185!  I am kicking myself for not buying 2 when they were on sale for $120 a few weeks back.

Anyway, this good price on the 550w got me thinking... what potential issues would there be in running a 6-GPU rig off of 2 of these?  Wattage wise, it could easily be done on XMR and while a bit closer, could probably mine ETH too with enough headroom.  But, not sure what concerns there would be with running a rig off 2 separate PSUs.  Could PSU #1 run the mobo and 2 GPUs, and GPU #2 run the other 5 GPUs?

I suggest doing a search in this forum as there are many threads that have addressed this. I am not familiar with your CPU but even an old i7-2600k is capable of mining 200-250 h/s on Monero (XMR). Given its recent spike in price, this would add up to a small but decent amount over time. I have 3 K-series CPUs on it and combined they yield more than a bios-modded 470... so yes, it can be done, but probably not decently with low end CPUs.

Thanks Aries! that 1750 strap seems to have done the trick for my 1 MSI 470 with Samsung memory. I switched over to ETH on that rig (mainly cuz of XMR 9.6 being unstable but also to get the hash-rate by card) and that Samsung MSI is right in line with the other 4 MSI speed wise.  I don't think I will mess with modding the Nitro right now as there seems to be some debate as to how best to mod it with Samsung memory.

As to my happiness with that Nitro... if I could do it over, I would rather have had another MSI 470 4GB.  Firstly, that Nitro 8GB cost me $40 more.  In addition, as mentioned above, it is proving to be much more challenging to mod the bios... so as it stands, it cost me about 20% more than an MSI 4GB and is getting about 15% less hash-rate... not to mention sucking more power.  Granted, if I can get a good bios mod in it won't be as bad, but somehow I doubt I will ever make up for that extra cost. It would have to hash 33 on ETH to be proportionally as good as the MSI 4GB, and there is now way that is going to happen.  I bought this card on a recommendation from another thread when I was still very new to mining... before Aries and others helped me maximize effectiveness of my MSI's.  I was very frustrated back then with getting low 20's on those MSI's so kind of jumped the gun with the hopes of getting a better card.

Thanks Aries!  I had heard that 1750 strap mentioned elsewhere but much better to get the confirmation from and expert like you before trying it.

Out of curiosity, would this apply to all 470's with Samsung memory or just the MSI?  Reason I ask is that I have 1 Sapphire Nitro 470 8GB OC that also has Samsung memory and I have been hesitant to mod the bios on it as well.

Totally agree on Lee and his videos... really good stuff and I have learned a lot from watching them. He is also quick to respond to any questions, which is really cool.

Gotta love the fact that the $400 of poker winnings I sent to my Coinbase BTC wallet a few weeks back is now worth over $600.  Unfortunately, now kicking my self that I didn't send more winnings to BTC at that time, but who knew, right?  This is also the first day in the ~6 months that I have been in this that all 9 of the coins I hold are up in the last 24 hours... and all but 1 (XRP) is up 10% or more lol.  I only wish I had more significant holdings but can't have everything, right?

Hey Aries... have you purchased any MSI 470s lately?  The one I got from Amazon a couple of weeks ago has Samsung memory and the previous 8 I own all have Hynix.  This made me stop short of 1500-strap-modding that newest card since I wasn't sure if memory brand would have any impact on this modding process.

So, my question is, for you or anyone else that is strap-modding their 470's with the method used below shown by Lee... has anyone performed this same mod with a newer 470 that has Samsung memory in it?  Just trying to figure out if something different needs to be done for these or not.

https://www.youtube.com/watch?v=OP_zSP2H6ho

Yeah, I've checked that site as well but it does not seem to be as reliable as coinwarez. The 2 big issues I have seen is that I is usually WAY off on ETH difficulty (a spot check now of the 2 sites at the exact same time has CW with 92.6 and WTM at 85.4), which really skews profitability. Every time I've checked, CW has been right and WTM has been wrong, so there must be something wrong with how they are pulling ETH diff.  Also, the $ values that WTM uses also usually seem to be off by some degree. Based on these 2 factors, seems to me that CW is much more reliable.

ZEC is only most profitable for older cards... I guess 390s, 290s, etc.  But for 470/480s, ETH & XMR have been more profitable than ZEC for quite a while now.  Simply plug your system stats into this site and you will see it.  ZEC has made a surge in profitability over the past few hours mainly due to lower difficulty but it is still clearly below the other 2, at least for the 470/480 crowd.

http://www.coinwarz.com/cryptocurrency

For ETH & XMR, rx 470's are probably the consensus winner... they hash slightly slower than 480s but consume less power and cost less, so the ratios are the best, at least for most people/systems.  For ZEC, it's a bit more confusing since older cards seem to get more hashes on it (but probably also suck a lot of power). Check out the ZEC thread for better recommendations on that. As for 470's, I prefer the MSI 4GB Gaming.  Super quiet, easy to mod the bios, look great and run pretty cool. Many people in this forum love that GPU so check it out.

Yeah I noticed that ETH went to the top this morning with its rise in price and the spiking difficulty of XMR.  For my 10 cards on 2 rigs, currently showing $10.20 profit for ETH and down to $8.55 on XMR. I will probably switch over my 6-card rig to ETH as that is a pretty decent % better for ETH. That would give me about a 55/45 ETH/XMR ratio since I have 3 CPUs churning XMR at a combined h/r of a bit more than a modded 470. 

Will be interesting to see what happens when Claymore releases v10 of his XMR GPU miner (should be any day now).  I guess it will really depend on how much more speed this version will provide us with 470/480s. Not expecting a huge increase but a marginal amount, along with better stability, would be nice.  I have had many problems with 9.6 on my rig #2 for some reason, which is yet another reason to move it to ETH, at least until we see what v10 XMR miner can do for 4xx cards.

Logged onto Bittrex both today and yesterday and got notification e-mail both times (with gmail)

The benefit is that ETH is currently more profitable to mine than XMR.  For my 10-GPU set-up, XMR profit per day has dropped to $9.52 due to skyrocketing difficultly, but ETH has now jumped ahead of it to $10.23 per day due to diff drop and rise in price.

Granted, if you are of the opinion that XMR has a much brighter future than ETH and your are holding what you mine, then my all means, continue mining XMR.  But you have to realize that a lot of miners are mining most profitable coin then quickly converting it to BTC, so if you are not holding ETH or XMR or ZEC, you should always be mining the coin with the highest profitability ratio.

The interesting thing is that when I checked this last night, ZEC was very close to the others, but now, even though the price is up a bit, profitability plummeted to $5.79 since the difficulty is up like 70% vs last night. Crazy that diff can fluctuate that much!

Thanks HashZilla... I am beginning to wonder/fear that this is a hardware (riser or GPU) issue, since my other rig is running fine for the most part (only that 1 fan issue in almost a full week running XMR). Yesterday, rig #1 ran fine with 1200/1900 from about 10am to 8pm, then a random reboot. Happened again just after midnight, then again at 4:30am... so I lost 4-5 hours of mining time there since I was sleeping.

I did check the log files for what you mentioned but all 6 GPUs are under that 1750ms mark. most are in the 1500-1600 range, only 1 is usually running about 1720-1730 (this is the 1 that I have not strap-modded yet since i noticed it has Samsung memory while all my other MSI 470's have Hynix memory).

I did notice that on 1 of the crashes, the log file shows a ridiculous mark of like 65000ms on GPU #1 about 1 minute before it crashed.  But, when I checked another log file, I didn't see anything like that so who knows what is causing this. I am well under my 1200W power limit so that's not the issue.  If I see this same behaviour today I may switch it over to ETH and see what happens.  At least that might narrow down if it's a hardware or miner issue.

So... both rigs ran fine overnight with the 1242/1900 settings, but then this morning, rig #2 crashes and reboots again. Same settings, same cards, generally same setup, but rig 2 is just unstable. Got it running again but now I am getting sporadic "low difficulty shares".  Does anyone know what the main causes of these are?  Also, what problems do they cause?  I've heard of people getting banned from dwarf for too many of these so I don't want that to happen, but don't know what to do to fix it either. Right now at least it I've only seen 2 in about 25 minutes of running but not sure what "too many" is either.

Well, after initial good results on both my rigs with 9.6, I now have problems with both.  Woke up this morning to rig #2 (6x RX 470s) crashed and all fans blasting. This happened the night before with core/clock settings of 1200/1950, so I had lowered it to 1200/1920.  It ran all day fine then just a few minutes ago, I check it and I had the same "Open CL hung in thread 10 & 11" issue that had occurred the night before.  Now I have it set at default 1242/1900, so we will see how that goes.

On Rig #1 that had been running 4x rx 470s for over 4 days straight at 1242/1900, it just experienced that sporadic fan blasting issue that others have encountered. So, I had to reboot and reset that one as well.

Bottom line, there definitely appear to be stability issues with the 4xx series cards, so hopefully this is something that can be addressed in the forthcoming v10. 

I know this belongs on the CPU thread but I have also seen weird things going on with Claymore CPU XMR Miner 3.5.  I have 3 K series i7's mining it and on 2 of my 3 PCs, the hash-rate has inexplicably dropped by about 10-15%.  This is after restarting and rebooting several times and there is nothing else running CPU wise, yet for some reason it's not as effective.

Really looking forward to the GPU miner update. Thanks in advance Claymore!

As a miner of ETH (not currently due to XMR profitability) this is a bit disturbing, and I don't really even agree with the premise or impact of what they are proposing here. But, I also am holding a decent # of ETH coins in advance of the move to POS, so I am somewhat leveraged either way.  But... speaking of the move to POS... isn't that going to happen in about a month anyway?  I heard in several places they were expecting that to happen around Feb, which would render this whole block reward chopping almost useless if true, right? So, they want to kill the last month of ETH mining?  Not sure but maybe that POS move isn't happening that early after all.

The smoke is a huge deterrent for me since I have horrible smoke allergies. I can maybe do 1 or 2 short casino trips a year and that's all I can handle.  The other thing to factor in with casinos is the "local factor".  Maybe this isn't as big a problem in larger areas like Vegas but have seen some pretty sketchy stuff in Reno... like locals/regulars getting preferential treatment in both cash games and tourneys.  I get why they do it... those people are their life-blood but the same rules should apply to everyone on the floor and that is just not the case all the time in casinos.

With online poker at least (not into other online gambling), most average poker players can do really well, mostly because there are sooooo many donkeys and fish playing it.  The only real thing to factor is player strength and what I like to call "donkey reward factor".  As I said, almost everywhere online the fields are weak to super-week so no real concern there. But the donkey factor is what you really need to pay attention to. This is basically the tweaking that the site does to the code to favor the horribly bad players so they won't bust their bankrolls right away. The thing is, all sites do this, but to varying degrees, which is the key.  Sites on the Revolution network for instance (like Intertops & Juicy Stakes) have horribly bad players but only an average donkey reward factor, so, for decent players, it's quite easy to win a lot there. The only problem at all with that network is that it has shrunk over recent years, but you can still get some decent action there and hopefully it will grow. Other sites run the gamut, so you just have to try them. But, bottom line is, it is almost always easier to win online vs casinos, unless you find a super juicy live game but that is not exactly easy.

In this miner, the only intensity throttle I know of is the "-a" setting.  For instance, -a 4 is max power (for certain cards), -a 1 is standard, -a 2 is lower power mode.  Hopefully, when Claymore releases what I assume will be v10 of this miner, he update this aspect as well since I think most people prefer the way intensity is handled on his ETH & ZEV miners, but we will see.  At minimum, if he stays with the -a options, there should probably be one more specific for recent cards.

Good data Carlo... my 6x470 rig was running solid over 24 hours at 1200/1950.  So yesterday I decided to try bumping to 1200/1975. It seemed fine but then I woke up this morning and found that it had crashed overnight.  Also, there was not really any notable h/r increase when I pushed them from 1950 to 1975, so I have them set back to 1950 now.  But, I will watch this closely given your comment about 1920.  No sense risking more down-time, especially for little or no h/r gain.