Ethereum GPU mining?

[EightySix Four]

I want to learn more about this attack vector, this is the first time I've heard about this. Crazy.

This also kind of explains the wild swings I'm getting on my mining pool. I get the caching techniques but maybe in addition the mining pool servers be upgraded to handle the extra load until we eventually fork?

Early days for Ethereum yet

Indeed, there's reason to believe the attacker is a semi-famous bitcoin maximalist, which wouldn't be surprising in the least bit. It's an interesting attack vector for sure, but it also means the attacker has mixed economic incentives - hurt miners while keeping ETH prices up to sell what he's mining. Probably helping to minimize the damage.

State journaling should help deal with a huge part of the load making the attack more expensive then it is worth. Gas pricing for contract is supposed to look more like a market than a fixed pricing thing, but the user experience for miners on that front needs to be smarter or the pools need to start throwing their weight around w/regards to gas prices for contracts.

I've actually decided to spin up a Geth and Parity node inside of some VMs because I think network stability is something worth throwing CPU cycles at.

Onward

 

[VeryCharBroiled]

Indeed, there's reason to believe the attacker is a semi-famous bitcoin maximalist, which wouldn't be surprising in the least bit. It's an interesting attack vector for sure, but it also means the attacker has mixed economic incentives - hurt miners while keeping ETH prices up to sell what he's mining. Probably helping to minimize the damage.
Onward

its almost like messing with ethereum is his hobby.

 

[Madpacket]

Greg Maxwell?

 

[lienad216]

been seeing that problem with RAM/CPU use as well. It didn't happen when I installed the mist client on the PC just a few days ago but it happened the next time mist was started up.

Its bad but its good, in a way, because the commitment of the devs is going to be tested and as someone else has pointed out, people pay for that sort of testing

On the other hand, does one need to have mist/geth/parity running to mine? I think there was a mention earlier in the thread that this was necessary for solo-mining, but I don't seem to have problems mining without mist running (as part of a pool).

 

[EightySix Four]

been seeing that problem with RAM/CPU use as well. It didn't happen when I installed the mist client on the PC just a few days ago but it happened the next time mist was started up.

Its bad but its good, in a way, because the commitment of the devs is going to be tested and as someone else has pointed out, people pay for that sort of testing

On the other hand, does one need to have mist/geth/parity running to mine? I think there was a mention earlier in the thread that this was necessary for solo-mining, but I don't seem to have problems mining without mist running (as part of a pool).

No, it isn't necessary for mining in a pool. That said, it's helpful for the network if you're running Geth/Parity/EthereumJ. Every node adds some resilience. As of right now a lot of pools and nodes are/have switching(ed) to Parity. While it helped with this attack, it's nice to have a more mixed network so I'm running both in VMs with set maximum resource usage. If Geth gets slammed again it can't slow down parity or my mining operation that way.

Obviously not necessary for anyone and it is using power and resources to do, but if you've got dedicated boxes you don't work on, every bit helps

 

[Feld]

Wow thanks for the detailed response.

"Why is the attacker doing it? Well it appears the attacker is mining with a black list in place to ignore their computationally expensive transactions, cheating the system to mine faster than everyone else who is stuck with their junk transactions in the blocks."

I want to learn more about this attack vector, this is the first time I've heard about this. Crazy.

This also kind of explains the wild swings I'm getting on my mining pool. I get the caching techniques but maybe in addition the mining pool servers be upgraded to handle the extra load until we eventually fork?

Early days for Ethereum yet

https://www.reddit.com/r/ethereum/comments/55xh2w/i_thikn_the_attacker_is_this_miner_today_he_made/

Good discussion on it there.

 

[Raduque]

Now I'm getting super inconsistent GPU usage and hash rates. Is it related to the attack? GPU usage drops from 99 to 70, 60, 30, then up to 85, 96, 99, and then back down following the first progression.

OK figured it out. Running parity alongside the miner causes the weird GPU usage.

 

[Madpacket]

Ignore

 

[Elfear]

What's the cheapest way to convert ETH to USD? Right now I'm selling my ETH through Poloniex for BTC, sending my BTC to Coinbase, and selling the BTC for USD. Lots of fees along the way though and I'm wondering if there is a less expensive option.

 

[IEC]

What's the cheapest way to convert ETH to USD? Right now I'm selling my ETH through Polyniex for BTC, sending my BTC to Coinbase, and selling the BTC for USD. Lots of fees along the way though and I'm wondering if there is a less expensive option.

Probably GDAX (part of Coinbase). That way you sell directly from ETH to USD.
https://www.gdax.com/

Be warned you have to generate a new ETH deposit address each session. So always click on your Deposit button --> ETH tab and use whatever the current address assigned to you is to make deposits.

 

[Elfear]

Probably GDAX (part of Coinbase). That way you sell directly from ETH to USD.
https://www.gdax.com/

Be warned you have to generate a new ETH deposit address each session. So always click on your Deposit button --> ETH tab and use whatever the current address assigned to you is to make deposits.

Thanks! I'll give that a shot.

 

[SimianR]

So I haven't mined Eth for a long time but decided to mine for a bit on my R9 290. When I fired up my miner, I'm getting really low hashrates compared to before, something like 15-16 MH's when I used to get closer to 30. Is this because of the new DAG size or is there another reason for this? I've also updated my drivers since I was last mining, not sure if that is also contributing to the lower rates.

 

[Bigbadwu]

anyone mining with a fury? just picked a sapphire nitro. I will like to know what hashrate to expect and how much I can undervolt it.

 

[Shmee]

IIRC a fury will get around 40-45 MH/s

 

[suklee]

What's the cheapest way to convert ETH to USD? Right now I'm selling my ETH through Poloniex for BTC, sending my BTC to Coinbase, and selling the BTC for USD. Lots of fees along the way though and I'm wondering if there is a less expensive option.

How about Gemini, has anyone traded Eth -> USD on there? I'm considering giving them a try..

https://gemini.com/fee-schedule/

https://gemini.com/transfer-fees/ (all "Free"?!)

 

[n0x1ous]

my Fury X tops out around 33/34

 

[lienad216]

There's a community forum on eth hash rates. I think its on the ethereum website forum but I'm not sure. A web search should yield the correct website. Perhaps you'll find more answers there?

 

[DrMrLordX]

My 290s at 900 MHz are still generating 25 MH/s so I don't know why one would be down in the 16-17 MH/s range.

 

[Madpacket]

anyone mining with a fury? just picked a sapphire nitro. I will like to know what hashrate to expect and how much I can undervolt it.

I bought a Fury Nitro a few weeks ago. At 1050Mhz core / 333Mhz mem it's mining at 30Mh. This is with a -72mv undervolt. The Nitro at these settings only consumes a tad more energy than my tweaked MSI RX 480 8GB Gaming X!

I plan on going lower to -96mv but this requires some ini modifications in Trixx first. These Nitro's are beasts. The fans don't even turn on while mining however I set a manual fan speed of 50% which lowers then wattage even more.

 

[Madpacket]

IIRC a fury will get around 40-45 MH/s

Edit: Oops, Nox already replied.

Fury X's top out at 33Mhz or so with a reasonable overclock / undervolt.

 

[DrMrLordX]

Geth users: are you seeing geth.exe chew up a ton of available storage space lately? Like, running it produced extraordinary constant disk access (nearly 100 MB/s) and caused a lot of free space to just . . . go away. I can't tell what data it was writing to the drive or where, so I'm not doing anything about it until I get a better grasp of what was going on.

This is with the geth from over a week ago, and the geth version released yesterday that actually works.

 

[Shmee]

Oops, my bad. I was thinking they were a good deal faster than Hawaii cards, but I guess not.

 

[DrMrLordX]

Okay, I did some poking around, and I've found that geth.exe/Ethereum Wallet is dumping a ton of data in the chaindata subdirectory (Windows: Users/<username>/AppData/Roaming/Ethereum/chaindata

Presently I have over 60 gigs in my chaindata subdir, what gives?

 

[Shmee]

Most likely the block chain? It would be nice to tell it to store it to a different drive though.

 

[Feld]

Okay, I did some poking around, and I've found that geth.exe/Ethereum Wallet is dumping a ton of data in the chaindata subdirectory (Windows: Users/<username>/AppData/Roaming/Ethereum/chaindata

Presently I have over 60 gigs in my chaindata subdir, what gives?

I read somewhere on reddit the last couple days that it has to do with all the data the attacker has been spamming into the blockchain.

You don't have to run Mist at all if you aren't solo mining. You can manage your existing Mist wallet with myetherwallet.com or switch to Jaxx. Then you don't need to download the blockchain at all.