Live data from Hacker News

Nvidia Limits RTX 3060 Hash Rate

tomshardware.com

411–420 of 520 posts

Re: Nvidia Limits RTX 3060 Hash Rate

#411

Earlier quoted context omitted.

The link shows paying in btc because they don’t give you what you mine. You rent your hash rate and people bid on it with btc. You’re not mining btc directly. I use NiceHash at the moment. Your gpu mines ethereum (or whatever the most profitable coin is as NiceHash has a profit switcher.)

Could you explain to me why someone would offer their hashing power on NiceHash vs. just directly working for a pool? The main cases I image are: 1. The miner doesn't want to deal with price fluctuations and wants to prevent arbitrage risk 2. It might be more profitable to outsource smallish coins that have to buy hashing power to stabilize their network (though that would probably also be covered by the profit switc…

I honestly don't understand who's buying the hash power on the other side. I think people start with Nicehash because it's PnP and easy to use. From there you can look at Minerstat which mines directly on pools and offers features like triggers and fancy graphs.

Setting up a miner yourself doesn't take much work but it won't have profit switching or extra cloud features that Nicehash or Minerstat have.

Re: Nvidia Limits RTX 3060 Hash Rate

#412

Earlier quoted context omitted.

Can someone explain how transactions are not insanely cost prohibitive for Bitcoin with power consumption figures like those? Those numbers would suggest something like >20 USD. Is it because most transactions in practice actually happen internally on exchanges that avoid putting every individual transaction on the blockchain? I have limited knowledge of Bitcoin so maybe not quite "explain like I'm five"... but close…

The whole point of Bitcoin mining is to produce a block , which contains a bunch of transactions. The Bitcoin network dictates that each block's SHA256 hash starts with a certain number of zeroes, so the only way to achieve this is to brute force the block data until you find the "winning" block which hashes to a number of zeroes. Once someone finds a winning block, they're rewarded with a number of Bitcoin. This sub…

> And to add to the "it costs 657.6kWh to process a transaction": energy is used to produce a block, which contains an arbitrarily defined number of transactions. Right now, Bitcoin Core limits each block to 1MB, which works out to about 2k transactions per block. If Bitcoin Core were to increase the limit to say 10MB, the energy used to produce a block doesn't change, but the energy used to process a transaction goes down tenfold.

Oh, it could be somewhat improved. Then it'd be "only" 65 kilowatt hours, compared to ~1 watt-hour for conventional payment networks.

(And, this assumes that the increase in bitcoin price doesn't cause more mining).

Re: Nvidia Limits RTX 3060 Hash Rate

#413
post #306

Earlier quoted context omitted.

Technology continues to advance - I don’t think there is enough evidence to say that we have hit a fundamental limit on computer/asic power/speed performance yet. 3nm will hit in 2024, but until then there will always be advances in chip design. It’s not just about speed, it’s also about thermals, power consumption and manufacture price to work out the TCO of mining.

> 3nm will hit in 2024, but until then there will always be advances in chip design. I'm not an expert on chip design, but I would have assumed that there isn't a lot of room for improvements of ASIC mining chips, given that they are probably a straightforward implementation of the hashing algo used in Bitcoin (and that a few thousand+ times), in contrast to CPU design which is more dependent on other computing/hardw…

Well even without chip design, for instance if you decrease unit cost, that also has the impact of changing the mining landscape because many more ASIC miners can enter the game.

Also remember an ASIC miner will generally use more than it costs in electricity each year, so a relatively small improvement in hashes / watt makes a big difference to overall profitability.

CPU design is more complex, but also has more people working in the space and has been going for longer. I just get sceptical if someone says this is as good as it is going to get because history shows usually those predictions end up being wrong. As a child I thought Super Mario 64 was as good as computer graphics were ever going to get...

Re: Nvidia Limits RTX 3060 Hash Rate

#414
post #400

Earlier quoted context omitted.

>Btc mining only works on asics. No, you can mine Bitcoin on your GPU or CPU. It will work.

Yes it will work. But it will be so slow that you won't ever do anything useful with it, so in practice it doesn't work. Go mine Ethereum or Monero instead.

They said it didn't work. That's all.

I was just correcting that. Bitcoin isn't unable to be run on GPUs or CPUs.

I think that's important to point out.

Re: Nvidia Limits RTX 3060 Hash Rate

#415
post #4

I have no sympathy for miners, but purely from software freedom perspective this is maximally evil: closed-source software that decides what you can use your hardware for. The CMP version is a straight-to-landfill product. As soon as it stops being profitable for mining it will have zero resale value, and become e-waste.

Try scanning a dollar bill and importing it in Photoshop.

Re: Nvidia Limits RTX 3060 Hash Rate

#416
post #386
post #285

Earlier quoted context omitted.

I think it’s important to note that the 3060 is not for sale yet. Nvidia is telling everyone up front that this is how it’s going to work. The analogy is improved like this: Ford announces in advance that their upcoming truck requires a certain usage, and if you want whatever special commercial usage, you need to pay a premium. You might be grumpy about the upcharge, but this is a very different situation from them c…

The key difference is a manufacturer exerting control over what you're allowed to do with property that is supposedly "yours". That the terms were announced ahead of time will mean little once this becomes commonplace.

This has been commonplace for a long time. See consoles, phones, etc.

Re: Nvidia Limits RTX 3060 Hash Rate

#417

Earlier quoted context omitted.

> The CMP version is a straight-to-landfill product. As soon as it stops being profitable for mining it will have zero resale value, and become e-waste. Like literally every other piece of Bitcoin or crypto mining technology. It’s straight to the landfill every 6 months. Bitcoin generates 100 GRAMS of ewaste for every transaction. Every 2 transactions consumes as much energy as driving a Tesla from SF to NY [edit for…

No. It's time to push back on Bitcoin. It is disgraceful to watch people such as Elon Musk and most of the VCs put money ahead of the environment.

Given the energy requirements and how it's kind of morphed into this beast of greed, I'm inclined to agree....but how to pull this off?

Fiat bans cannot really work well against just doing math. Taxation schemes/financial regs tamping down on bitcoin speculation, what you can spend it on? By design I don't think you can just pump money into bitcoin in order to "deflate" its value without similarly consuming a lot of energy?

Re: Nvidia Limits RTX 3060 Hash Rate

#418
post #80

I wonder if open source drivers will eventually circumvent this limit?

Sadly, the last time I checked the benchmark for the nouveau project ( https://www.phoronix.com/scan.php?page=article&item=nvidia-n... ), it wasn't even close to being "usable" in terms of performance (need to expect about 10x slowups in games).

It’s because you need to load firmware on the gpu to make it run fast and that firmware is illegal to redistribute and impossible to reimplement since it’s signed. There is nothing the nouveau project can do short of finding an exploit in the cards crypto.

Re: Nvidia Limits RTX 3060 Hash Rate

#419
post #259

Earlier quoted context omitted.

I checked your numbers against [0]. Road distance NY -> SF = 2900 miles [1]. Model S mileage is 290Wh/mi [2]. That is 840kW for Tesla against 657kW for a single transaction. This is all order-of-magnitude stuff but... wow. [0] https://digiconomist.net/bitcoin-energy-consumption/ [1] https://www.google.com/search?hl=en&q=road%20distance%20san%... [2] https://www.zeviecars.com/browse-cars/tesla-model-s

Can someone explain how transactions are not insanely cost prohibitive for Bitcoin with power consumption figures like those? Those numbers would suggest something like >20 USD. Is it because most transactions in practice actually happen internally on exchanges that avoid putting every individual transaction on the blockchain? I have limited knowledge of Bitcoin so maybe not quite "explain like I'm five"... but close…

Oh that’s inflation.

The cost is socialized across the block reward. So long as there’s more new money coming in than block reward paying electric bills on the way out, the cost of a transaction is socialized efficiently.

Elon’s $1.5B investment only lasted a total of 4 weeks. It’s already gone. It’s in the hands of Chinese coal produces now.

Re: Nvidia Limits RTX 3060 Hash Rate

#420
post #413

Earlier quoted context omitted.

> 3nm will hit in 2024, but until then there will always be advances in chip design. I'm not an expert on chip design, but I would have assumed that there isn't a lot of room for improvements of ASIC mining chips, given that they are probably a straightforward implementation of the hashing algo used in Bitcoin (and that a few thousand+ times), in contrast to CPU design which is more dependent on other computing/hardw…

Well even without chip design, for instance if you decrease unit cost, that also has the impact of changing the mining landscape because many more ASIC miners can enter the game. Also remember an ASIC miner will generally use more than it costs in electricity each year, so a relatively small improvement in hashes / watt makes a big difference to overall profitability. CPU design is more complex, but also has more peo…

> for instance if you decrease unit cost

That doesn't diminish the hashing power of the equipment you already bought though (and the topic was obsolete equipment).

> so a relatively small improvement in hashes / watt makes a big difference to overall profitability

Yeah, that's a fair point.

> I just get sceptical if someone says this is as good as it is going to get

And I'm skeptical if people blindly extrapolate past trends, assuming that they will continue to hold. I wasn't saying that it won't get better from here on out, just that the extreme growth that was mostly fueled by quickly going through the existing nm process steps can't be sustained, and that from now on it will follow the same slowish pace as all chip miniaturization.

Post reply on HN