You can just use the regular Claymore 9.4 miner and add arguments to mine SIA. That's what I've done from the beginning since I don't want to use dwarfpool.

Sorry for bumping but I'd really appreciate some help with this. Thanks!

Hello,

I have two Asus RX 470 4GB in my daughter's PC. They are the same model and I've flashed the same BIOS to them both.

However, one of them refuses to go past 23MH/s. It doesn't matter which settings I use in Claymore or Afterburner, it just doesn't. It's also stuck at the same 55c temperature while the other is at 65c (it's hot in her room and not so good cooling). No matter what I do it's stuck. The other hashes at just above 27MH/s.

I've tried increasing the page file, I've tried changing power limit, memory timings and memory clock. It's the exact same speed at 1650MHz memory clock as it is at 1950MHz. I've also messed around with the -dcri setting.

Any ideas? Thank you!

No, perhaps not increasing the core clock for ETH but it matters for other coins. And since it's simply not working at all, I'd like some help in figuring out what the reason could be.

When it comes to power limit, however, it makes quite a bit difference. I know it consumes more power, but in Windows 10 and +20 powerlimit I got 165 MH/s and 600 MH/s SIA (I think). However, in SMOS I get 144 MH/s without lowering -dcri more.

Is there any way to increase power limit and increase core clock speed? I have 6x Sapphire RX 570 Pulse Mini's and SMOS defaults to 1106 core clock, even though the GPU BIOS is 1244 and I've set 1350 in SMOS. Memory can be overclocked fine.

I'm not familiar with linux if there's anything I need to do that isn't available from the dashboard. I should note that adding "-powlim 50" to the Claymore setting doesn't do anything. Neither does "-cclock 1350".

When I was using Windows 10 with 6 GPU's Windows usually crashed at least every 3 hours or so, needing a hard reset. I'm sure it was a software problem but after like 20 hours of testing I gave up and tried SimpleMining instead. So far, it's been working great, except for not being able to overclock the core clock. I should be able to (SMOS supports it), I'm just not sure why it's not doing it. Memory overcklock works though.

I have all six cards on risers. I'm pretty sure it'd be fine to have one or two without risers and I don't think it should impact performance. But I haven't tried so I can't say really.

The iGPU is set to be the primary display. It has a monitor connected to it. I do have a dummy in one of the RX 570's though.

It's the same rig, I just switched to SimpleMining. I'm having some issues with overclocking in SimpleMining though but it'll probably get sorted soon - or I'll have to flash the cards in Windows first I guess.

I'm having another issue. While my 6x Sapphire RX 570 Pulse Mini's work fine with SimpleMining, it seems that the core overclocking doesn't work.

If I leave it as nothing, it'll run at 1106MHz even though base clock is 1244MHz. It doesn't matter if I change it from the OC settings or directly in the miner config.

Memory clock, however, works just fine to overclock.

Any ideas?

Edit: It looks like it's setting the correct value at boot but then changes back to 1106MHz as soon as it starts mining.

I have the Asus Prime Z270-P and 6x Sapphire RX 570 Pulse Mini's. I got it working with Windows 10 after quite a while. However, SimpleMining just works out of the box, so I'm using that instead.

Since you already have 4G enabled, try setting PEG and PCI Gen to Gen 1. That's really all I had to do to get Windows 10 to boot (it didn't boot at all with more than 3 otherwise).

With SimpleMining, however, I needed those two settings set to "Auto" otherwise it wouldn't work properly.

I do have it set to Gen 1, I'll try setting it to auto in the morning and report back. All other things you suggested is already how I'm set up.

Edit: I tried it before I went to bed yesterday and it's been running great for the last 7 hours, so it seems that it did the trick. Thank you!

I'm having issues getting any proper speeds with SimpleMining.

Without OC but 1500 mem straps my 6 Sapphire RX 570 Pulse Mini are hashing at about 24MH/s ETH + about 470 MH/s SIA each in Windows 10.

However, with SimpleMining I get these weird speeds that fluctuates a lot:

Next status update looks like this:

Like I said, they all work properly in Windows 10.

I should mention that I'm using the same settings for Claymore in both Windows and SimpleMining of course.

The dashboard also reports some weird numbers for the GPU's:
These also fluctuate up and down and never seems to be stable at all.

As I understood it, problems could arise if you're drawing more power from the cables than they are rated for.

In my case the GPU is 6-pin so 75W from the 6-pin cable and, I assume, the rest from the riser which is connected with a Sata cable.

Hello,

I'm waiting for my 006C risers to arrive but have a lot of new 007S risers lying around (not from this seller but this kind: https://www.aliexpress.com/store/product/Latest-version-ver-007S-board-PCI-E-PCI-E-Express-1X-to-16X-Riser-Card-USB/1627923_32766321782.html).

Would it be possible to use them for about a week until the others arrive? Of course, I would be using a single SATA cable to power each riser. The PSU is an EVGA G2 1300W.

I'd be running 6 Sapphire RX 570 Pulse Mini. The specs claim that it will use <150W (one 6 pin) but I could of course under volt it to use even less.

Thank you for your input.

I'd rather go with GPU's. But thank you for the suggestion.

I already have a Baikal Giant but am not interested in getting more ASIC's.

Sure, the RX 570 is cheap so it seems good to begin with. And we've actually begun by ordering six Sapphire RX 570 Pulse Mini's for the first test-rig. But how much can you undervolt it and what would the hash/watt be? That's the big and really important questions here.

We have the money but we don't have the power, that's why it's really important with hash/watt - and also the price for the GPU. As per my example in my first post - the GTX 1070 Katana use half of the power and has about half of the hash rate of the GTX 1080 Ti. So they're almost the same when it comes to has/watt. However, since two GTX 1070 Katana's are more expensive than a single GTX 1080 Ti, it's not as good anyway.

Right, I'll only be able to do 2+2+1+1 in that case since I have two molex cables left over that I don't use from my PSU that I have on my gaming PC. I'm assuming they'll work since the two PSU's are the same model but different wattage. One is EVGA G2 750 and the other (for the mining rig) is EVGA G2 1300.

I understand that it's safer to go with the 6-pin risers. I'll order those. However, I have molex risers incoming already - which I always thought as temporary because I could get them fast. Question is if I can use them for a week.

I've read that three risers per cable will melt it when it comes to molex. That's why I was thinking two risers per cable and three cables.

You say that some cables melted when the room was hot enough - can you remember how hot the rooms was when they melted - or take a guess at least?


Now, I don't know much about electricity but pretty much everyone who has an AsRock BTC motherboard and use the two molex connectors on the motherboard claims that it works absolutely fine with all six GPU's and unpowered risers. I don't doubt you (I'll order the 6-pin risers to be safe!), but it seems to me that only two molex connectors on the motherboard nets you another max 100W, but that's apparently still enough for six GPU's on that motherboard. Or am I missing something?

This is so much great information that I had no idea about. Thank you so much, very informative and important!

In other words, you're saying that 6-pin connectors on the risers are needed. However, if that's the case, then how come it seems that most miners (well, that I've read about, perhaps not "big-time miners" though, but they do just fine with molex on their risers. Most cards use more power than the Pulse Mini. Of course, under volting is an option that many do so that would explain it.

Let's say the card uses 150W and since it's being powered by 6-pin cables, it also draw 75W from the PCI slot. Obviously that's not a good idea for molex cables then. However, if I under volt each card to make sure it never goes above 125W (lower of course to have some head room), it should be fine until the new 6-pin risers arrive, correct?

For example, if I under volt it enough to get to 110W max, would it still draw 75W from the 6-pin cable and then another 35W from the molex cable?

Thank you again for all your answers, you're extremely helpful!

Edit: Are you sure I can't use two molex risers on a single cable? Most people I've read about claim that it works just fine but that three is too many.

The molex cables are connected to the "PERIF1" and "PERIF2". It looks like 6-pins on the PSU but the other end of the "PERIF1" and "PERIF2" are molex cables.

The 6-pin cables that we're talking about are the ones that are used on GPU's. On the PSU they are connected to the ones on the right (the VGA1-6). So there are 6x 6-pin cables. Molex cables, however, only have two connectors on the PSU (the "PERIF1" and "PERIF2").

He was talking about splitting those 6 cables to make them 12 in order to use them for both the GPU's and risers.

I hope that makes more sense.