|
Was on my vacations while my small rig with 2 vega64 lc and 1 vega56 air was running happily with 5 extra nvidia cards.
vega 64 lc are running 1408/1160@875mv, vega56 is running 1408/1148@875mv.
miner is cast-xmr, perfectly stable for a week.
vega64lc are averaged to 2070h/s, vega56 is averaged to 2000h/s.
I didn't retest power consumption but it is low, my estimation is 140-150 watt for vega56 and 150-160watt for vega64lc
Now I'm back and happy that cast-xmr supports nicehash )) didn't wanna bother with etn mining directly ))
|
|
|
|
|
You already killed profitability of this coin with your aggressive marketing. So your improvements worth zero. ))
|
|
|
|
if I can run 11x Vega on ASUS B250 I can achieve 2 KHash/s on each gpu ?
(Windows with the latest updates removed cards limitation you can run 12 AMD card, but ASUS supports 11 AMD)
with current drivers - no. you need to wait for fall creators update compatible drivers with compute option for Vega activated. |
|
|
|
|
you should get 2kh/s with vega56 and 2.07kh/s with vega56 with right tuning. 150-160watt.
|
|
|
|
cryptonight is both memory and compute hard - the same as with zec.
10% is compute hard. 90% wasted waiting for memory. My cryptonight mod is running full speed @ 50% tdp. Most of the gpu is not used. but speed is rising when you overclock gpu core. Both amd and nvidia. The same is for zec. For ETH hashrate has upper limit bounded with memory and if you o/c gpu clock above some value there is no hashrate increase |
|
|
|
|
guys, don't bother arguing me ))
I said about "stable" - coins - that you can mine during months without jumping the pools, exchanges, etc. Now it is eth, xmr and zec. Maybe some extra dagger or equihash coins. All others need constant attention and time. It is not suitable for all ))
And don't tell me about other coins - it was me here to point you smartcash some time ago - it was not very trivial to find and rather profitable for a while. Just an example ))
|
|
|
|
All memory-easy algos will end this way ))
wrong, Small memory easy algos could be included in dual miners. I am working on a cryptonight dual miner. Claymore is in the mud, to time to give him some competition. cryptonight is both memory and compute hard - the same as with zec. You won't get good results )) |
|
|
|
Should i buy 3x 1080ti or 4 1070ti ? any idea ?
Buy vega You go with shitty vega you troll. Yeah quite shitty the vega with 2048 bit memory bus and memory easy algos getting ASIC'ed by x10 Baikal with a mere firmware update, good buy the ti with its awesome 352 bit bus. 1. It's thread about mining specific on 1080ti. Please stop spaming it with shitposts. 2. As you can see there's nothing released by Baikal right now, there's rumours that it'll be included in next batch(not firmware update). 3. What Your precious Vega can? Mine monero with 400W at wall or getting 3-4MH/s in ETH by huge increase in energy used? 4. As been said before. It's 1080ti specific. If You don't own one, don't have anything to say, take your '2/3 words nothing-to-contribute posts' and go to forum spam section  Vega is doing 2.1kh/s xmr with only 150watt or 46mhs eth with the same 150watt. Or 45mhs eth+220lbr with 210wat. This is stable coin profits and you don't need to always seek new coin to mine like with 1080ti where zec is only stable now. Vega is cheaper, more energy effective and ~equal gross profit comparing to gtx1080ti. I have both. nVidia is the only ZEC and new shitcoins queen. |
|
|
|
So, I'm trying to pinpoint an issue with STAK-AMD.
On my test rig with 4 vega 56, it will lock up hard after 10 to 60 minutes in the mining.
One of the cards will do 0 hash. The miner is locked up. Stopping it hangs windows with a driver error.
Rising the HBM and GPU voltage from 905 to 930 will stabilize the card, but then something else happens: All cards stop mining after some time (again, 10 to 60 minutes).
At that point it's possible to stop the miner, restart the cards and go for a little while until the same happens.
If I take out the "problematic card", laving only 3 of the others, the same happens: all cards stop mining.
Now the catch: I get cast-xmr and it's been mining for hours with no issues. I can also mine at about 7850 h/s on 4 vega, at 700 W from the wall.
Using the blockchain drivers from Aug.23.
Registry undervolting. 905-905 gpu/hbm and 1408/1100 freq on gpu/hbm.
Reducing the memory speeds does nothing. Stak will hang again.
Now, anyone has stak-amd running well on similar hashrates or are you all using some other miner ?
What config did you use with xmr-stak-amd? Vega56 doesn't like high intensities with xmr-stak. cast-xmr is way better for vega56 if you are not concerned with closed-source risk )) |
|
|
|
Feel free to remove MyrGroestl and Skein from list of selling mods. Baikal included them in X10 miner. Probably 2 more should be removed with first X10 firmware upgrade.  Yes, this is the reason _sp gets rid out of it. All memory-easy algos will end this way )) |
|
|
|
We are talking about speed 1900-2000h/s and power usage like 130-150watt. |
|
|
|
I have 2 vega56, does vega64 bios increase hashrate?
only if you will oc memory to 1100 and higher. And the increase will be small for xmr. For ETH - much higher. |
|
|
|
Thanks for this, bought a few vega 56's just to play with your miner. I get 1950 H/s but it drops to 1750 every time I start the second vega. Any tips? I wanted to make sure crossfire is disabled but I can't see this option anywhere. This is running on an older AMD 990 FX board with latest updates on Windows 10 and the beta blockchain amd driver.
-edit-
one more question -- is the performance bad with 4gb RAM? I saw you listed 8GB as the requirement. I have 16GB in the machine I'm testing it on now.
Use latest version - it supports multiple gpus in one mining window. You will need adjust bat file accordingly. By the way I was trying to mix cast_xmr for vega56 and xmr_stak_amd for vega64lc and got the same problem - after starting second process (cast or stak) the first miner losts the speed. So I have to choose to use cast_xmr or xmr_stak only. |
|
|
|
Today I got my first VEGA 56.
The hashrate is max at 1940H/s but frequently it drops down to 1870H/s and then rises to 1940H/s back and so on.
Average hashrate (15min) is 1915H/s.
I am just wondering, because my Vega 64s with the same setup/mod have quite stable hasrates around 1980H/s +/-3H/s
This my setup: Vega 56 with Vega64LC bios to 1408/1100@905mV (regmod) and August 23th blochchain drivers
"gpu_threads_conf" : [
{ "index" : 0, "intensity" : 2016, "worksize" : 8, "affine_to_cpu" : false },
{ "index" : 0, "intensity" : 1736, "worksize" : 8, "affine_to_cpu" : false },
],
Any ideas?
Thanks and cheers
make sure you have fan RPM enough to keep memory cool. and gandalf's cast xmr does better with vega56 than xmr_stak
Can you provide a link please? It is a parallel thread here https://bitcointalk.org/index.php?topic=2256917.0disclaimer - it is closed source and with devfee. But Wolf0 examined its internals and said it is interesting. Will hope it doesn't have some virus )) |
|
|
|
Today I got my first VEGA 56.
The hashrate is max at 1940H/s but frequently it drops down to 1870H/s and then rises to 1940H/s back and so on.
Average hashrate (15min) is 1915H/s.
I am just wondering, because my Vega 64s with the same setup/mod have quite stable hasrates around 1980H/s +/-3H/s
This my setup: Vega 56 with Vega64LC bios to 1408/1100@905mV (regmod) and August 23th blochchain drivers
"gpu_threads_conf" : [
{ "index" : 0, "intensity" : 2016, "worksize" : 8, "affine_to_cpu" : false },
{ "index" : 0, "intensity" : 1736, "worksize" : 8, "affine_to_cpu" : false },
],
Any ideas?
Thanks and cheers
make sure you have fan RPM enough to keep memory cool. and gandalf's cast xmr does better with vega56 than xmr_stak |
|
|
|
This sounds great and that is low voltage at 905 only but does this require very low temperature or would air work maybe also. HBM2 temp gauge shows much higher temperate then gpu temp sometimes so I wonder if its a problem.
I think over 80c the card is not stable at high clocks. Its liquid cooled so I presume its a factor, if I can keep HBM2 in the 60's will that be low enough to get the same results I wonder or it needs lower
I set my Vega56 with Vega64LC bios to 1408/1148@900mV and it is happy stable mining xmr 2kh/s. I set target temp 50 degrees, fan is 2400-2800rpm, hbm temp is 65-67 if we can trust HWINFO64 ... |
|
|
|
If the pool accepts stale shares, it's a bad configured pool, you're getting paid for work which is obsolete... You'd better check why your pool is paying less with many coins comparing to other respected pools even if they have larger fee like yiimp. Also I didn't receive any explanations or even excuses for all lost several day mining work with SweepStake coin on your pool. |
|
|
|
This miner has a serious problem for me. It can work only if there is no my nVIDIA 1070 on the system (with nVIDIA I mining ZCash). Then it says a blockchain driver was not detected. When I removed the nVIDIA 1070 then it works well. What is the problem ?
you should change opencl platform to 1 or 2. Check --help how to do it. If im not frong, here described it about 7 times, you need to remove nvidia schedule task.
Wrong answer |
|
|
|
Hi,
I have got a question...
Actually I have one with 3 VEGA 64, mining XMR 1408-1100@905. Power consumtion for the whole rig is around 520Watt.
I have a 850W PSU "evga 850 g2 supernova".
Now my question: Does it work out to add a 4th VEGA to the rig with that one 850W PSU?
Thanks and cheers
Oskar
Only if it is Platinum-rated. Otherwise, if gold rated, 80% rule of 850 is 680. 680-520 = 130 headroom for another Vega, I wouldn't do that on mine. Thanks. I expect around 700W for the hole rig with 4 Vegas, these are just 20W too much...  The efficiency of the PSU is min.92%@220V This PSU has only 6 pci-e power cables? How are you going to have 8 cables for 4 vega's? And what if it becomes profitable to mine eth or dual? Then your psu will be not enough |
|
|
|
|
Show Posts
