ZCash/ZEC GPU mining

[thilanliyan]

^Yeah, I have no idea how a 1080 + 960 is putting out 750S/s. I've heard the 1080 can get to ~500S/s though, so maybe the program is reporting incorrectly. Even then though...a 960 can't get to ~250S/s can it?!

 

[[DHT]Osiris]

If a 1080 normally peaks at 500S/s and the 960 is benching at about 35% of a 1080, I'd expect more of a combined around 650-700. 150-200S/s would be more reasonable.

I can only assume that equihash is either measuring things differently, nicehash is reporting things differently, or some combination of the two.

 

[Accord99]

I would expect that too, since it's no-kidding 100%'d for the entire time it's mining vs gaming's natural peaks and valleys. Having said that, I know the 1080's gonna be pretty efficient due to being a newer arch, although it is powerful. I wish I had an actual wall-meter to test this heh.

The GTX 1080's power meter is pretty accurate, you can use a tool like GPU-Z or Afterburner to monitor the power % and multiple that by the TDP to give a reasonable wattage use value for the video card.

 

[[DHT]Osiris]

The GTX 1080's power meter is pretty accurate, you can use a tool like GPU-Z or Afterburner to monitor the power % and multiple that by the TDP to give a reasonable wattage use value for the video card.

So it looks like during mining, it spends 95%+ of its time between 80% and 100%, with a few peaks above and valleys below. @180W TDP, and assuming 90% power usage (which seemed average) that should mean it's probably averaging 165Wish.

Interestingly, the 960 holds dead steady at 73-74%, so it's using about 90w.

It looks like the equihash miner (eqm.exe) is showing 135 Sols/s for the 960, unfortunately right now the 1080 is mining via ccminer (lyra2v2 algo) so it's showing in MH and not Sols, but it's steady at 47.2 MH/s.

 

[cytg111]

Whats the ideal mining rig here, hardware, return on the dollar after electricity?

 

[poofyhairguy]

Whats the ideal mining rig here, hardware, return on the dollar after electricity?

One you can game with so as to get real value out of it. Mining isn't the profit center it was six months ago.

 

[Shmee]

Impressive results with the 1080, looks like more recent drivers/miners have helped it a lot for mining. Has anyone tried mining eth with a 1080 recently to see if the situation there is better now?

 

[DrMrLordX]

On the GCN side, it looks like there's a dev war going on between Claymore and Optiminer.

Claymore released v10.0 boasting 340 Sols/S for a 390x, while Optiminer is on v1.3.1 which boasts 285 Sols/S on a 290x. I had problems with Optiminer a few versions ago, but I haven't tried it recently . . .

 

[thilanliyan]

I'm getting 300s/s on an unmodded bios 290 at 1100/1250 using Claymore, so I guess he's winning?

 

[DrMrLordX]

Probably. Though when I updated from Claymore 9.3 to 10.0 I got no noticeable upgrade in performance. It didn't hurt anything though. Maybe I am not using enough intensity.

I see no reason to mess with Optiminer at the moment. It used to be the only game in Linux town, but not anymore . . .

 

[poofyhairguy]

Probably. Though when I updated from Claymore 9.3 to 10.0 I got no noticeable upgrade in performance. It didn't hurt anything though. Maybe I am not using enough intensity.

I got an extra 20 out of my 480 with Claymore 10. Intensity set to 8.

 

[thilanliyan]

My 290 is actually faster under Optiminer by about 20 S/s, but it's a lot less stable, can't reconnect after network loss, and the watchdog command doesn't seem to do anything in Windows anyway.
One thing I have noticed is that Claymore V10 seems to disconnect my internet for about 30-40 secs every hour or so.

 

[DrMrLordX]

I got an extra 20 out of my 480 with Claymore 10. Intensity set to 8.

Not too bad. I sense that a lot of development is shifting away from Hawaii and older cards.

My 290 is actually faster under Optiminer by about 20 S/s, but it's a lot less stable, can't reconnect after network loss, and the watchdog command doesn't seem to do anything in Windows anyway.
One thing I have noticed is that Claymore V10 seems to disconnect my internet for about 30-40 secs every hour or so.

I have had problems with Optiminer being randomly dropped by the pool. It'll mine locally, but the pool won't acknowledge results. Sometimes it happens after an hour or two, or sometimes it takes a day or more to manifest. I haven't tried 1.3.1 to see if it is faster on my Hawaii cards than Claymore 10.0

 

[thilanliyan]

randomly dropped by the pool. It'll mine locally, but the pool won't acknowledge results. Sometimes it happens after an hour or two, or sometimes it takes a day or more to manifest. I haven't tried 1.3.1 to see if it is faster on my Hawaii cards than Claymore 10.0

At least in the miner reported results, Optiminer 1.3.1 was faster than Claymore 10 on my 290s (slower on my 480) but I didn't run it long enough to get an idea of whether the pool was seeing the same hashrate.

 

[thilanliyan]

I ran Optiminer for longer and it's definitely faster on the pool side, by a good 40-50S/s. However, it crashes after several hours of mining. It didn't even go to the watchdog. It just ended the program. Need a restart script maybe.
On Claymore, for some reason when I add up the miner rate, it is never equal to what the pool is showing (pool shows about 30-40 S/s lower than what it should be), even with the 2% deducted.

 

[suklee]

Claymore 11 is out : https://bitcointalk.org/index.php?topic=1670733.0

 

[Shmee]

Cool. I'll have to give it a try.

 

[thilanliyan]

Fury users will see great speeds from what I understand with V11.

 

[DrMrLordX]

I have V11 on Win10, but it was not available for Linux when I checked. Hopefully it'll become available soon.

 

[Shmee]

My fury would not even hash with the drivers I have. I use that rig for gaming as well, so having 15.12 is silly.

 

[DrMrLordX]

Does 15.12 even support Fury?

 

[thilanliyan]

Claymore 11.1 is out. Didn't really do anything for my 290s...actually I think they are slightly slower than V11. Anyway, I'm using Optiminer for my 290s (it's about 20S/s faster), and CM for the 480 as it is a bit faster for that. Not sure what I'm doing wrong with CM for the 290s...not sure why it's so much slower than Optiminer.

 

[Madpacket]

11.1 shows a very small bump for 480's. Fury's and Hawaii are a bit faster as well. Very small gains though. All Windows 10 using drivers from Nov 2016..

 

[poofyhairguy]

I still can't use anything past 9.3 on my rig of mixed old and new AMD GPUs.

I might just take out the 460 (the only new GPU in there) as the speed boosts since then might negate the loss of that card.

 

[thilanliyan]

I still can't use anything past 9.3 on my rig of mixed old and new AMD GPUs.

I might just take out the 460 (the only new GPU in there) as the speed boosts since then might negate the loss of that card.

Have you tried running an instance per GPU (with .exe renamed and in a different directory)? That's the way I do it (2x290 + 1x480), but then I only have 3 cards so it's easy.