News for hackers. Stuff that matters.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @10:48AM (#32912398)
I have 6 digits worth of these chips. They've been running fine for years in all sorts of weather conditions. Mining does not destroy cards.

The paste goes bad, but is easily replaced (majority of our cards have been running just fine with OEM old paste). We don't have fans on the cards, but if we did those would go bad and easily replaced. But the cards themselves, really don't go bad very often.

If you're thinking about picking some up, don't worry about the mining part, but definitely use that to try to get a discount.

Re:In 2019 RX470 were $70, RX580 $100 by bravetraveler on Tuesday September 20, 2022 @10:57AM (#32912542)
One thing I find kind of interesting about this -- the folks dumping cards on the market had farms. Those running mining farms are playing a numbers game

This involves under-clocking and under-volting the cards, while also ensuring they're cooled well.

I think an argument could be made that mined-on cards are more well cared for than those owned by your average gamer.

There's not some 'destroy the ASIC' instruction. Work is work.

Mined on cards undergo fewer duty cycles, I wonder how the calculus works out between one long mild duty cycle compared to many more that are also more extreme.

Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @11:13AM (#32912830)
Most large farms don't have the ability to properly tune cards for best ROI. At scale, it is too hard because each card is a snowflake in terms of performance. In order to tune, you're basically pushing the card to the limits of underclock/undervolt... which means the card crashes. Therefore, most large farms actually undertuned, for stability, which is the least abusive on the cards.

If the card crashes due to tuning, the machine crashes, and the rest of the cards are taken offline until the machine reboots. In order to change settings, you have to reboot the entire machine too. It is a pain in the ass to find optimal tuning at scale.

There is no publicly available software to 'autotune' cards at scale. It took me a long time to first learn how to tune cards (by hand tuning thousands of cards and getting a feeling for things) and then once I figured that out, I built software to do this. It required changes in the AMD driver to get it to output in the kernel messages, which card was crashing. It required updating the mining software itself. It was a nutty insane project, but it worked well. We hit our target hashrate / power usage.

Re:In 2019 RX470 were $70, RX580 $100 by Melatonic on Tuesday September 20, 2022 @01:12PM (#32914756)
You should really pivot and start selling that software. Package it up with a nice pretty bow and wrapper and it could be a very lucrative product.
Re:In 2019 RX470 were $70, RX580 $100 by hailwren on Tuesday September 20, 2022 @11:24AM (#32913005)
What are you doing now? Did you shut the project down entirely?
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @11:29AM (#32913086)
Some is shut down, some is mining other coins. Kind of a wait and see time right now.
Re:In 2019 RX470 were $70, RX580 $100 by Ruphin on Tuesday September 20, 2022 @02:24PM (#32915829)
With Ethereum no longer in the game the total USD issuance for GPU mining probably dropped over 80%. Is your strategy to just wait out for miners to capitulate until mining is profitable again? Are you not worried that this will crater the value of your hardware?
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @02:59PM (#32916279)
We bootstrapped the company on mining, knowing full well that it would eventually end, and our business model goes beyond just mining ETH.

That was another reason to go with the RX4/5 series... fastest route to ROI.

Re:In 2019 RX470 were $70, RX580 $100 by Kerbonut on Tuesday September 20, 2022 @12:49PM (#32914416)
Considering the current state of PoW mining, it doesn’t seem to me like there is much advantage to holding secrets on auto tuning and diagnostics… (correct me if I’m wrong), would you consider releasing any of that work to the community? Auto tuning and diagnostic information would be incredibly useful for almost everyone.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @01:15PM (#32914798)
My knowledge is very specific to just these cards.

The only 'secret', which isn't a secret, was that I realized that since this is hardware, we know the maximum settings (hash/watt) that should be possible. Therefore, I set the cards to that best setting and then tune down from there. This is the opposite train of thinking than most other miners.

Most people think, let's start low and then tune up from there to make them 'faster'. The cards crash when they can't handle the settings, so it turns out that tuning down is a better way to tune since they stop crashing when they are stable... and thus don't need tuning from there. There are 3 different sets of 'knobs' to tweak, so I had to build an algo to adjust the knobs in the right order to tune things down. I just had the concept of 'current -> next settings'.

Temperature and power fluctuations can make the cards crash too... so by always tuning down, you're always heading towards more stability instead of instability. Since neither of those could be controlled, machines would reboot randomly all the time.

The software I built was a golang daemon that ran on each machine and watched for these crashes and modified the tuning of each card individually. The daemon is pretty cool as it is effectively a task runner. I had different tasks to configure and monitor the machines as well. The machines are all independent, idempotent and self-healing workers. Reliably distributing the software to 20k different workers, is a fun challenge. There are a ton of unit tests, so that helped a lot.

If I have the energy, I may rip out the tasks from the daemon, turn it into a library and open source that. It is kind of a fun project that could be useful for others trying to manage large scale individual workers. Tasks could easily be 'apt install' or monitoring utilities. I even bundled node_exporter into the binary, so that we could monitor the machines with prometheus.

Re:In 2019 RX470 were $70, RX580 $100 by arcticbull on Tuesday September 20, 2022 @12:48PM (#32914390)
The planet thanks you for shutting down. Don't worry, I'd say the same for any tobacco executive.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @01:02PM (#32914631)
Our power is unused hydro, in a geographic location full of other data centers for all the large tech companies.

This comment is just your continued trolling with your anti-crypto trope. Please try another tactic.

Update: I got downvoted to -4 on this one after getting several upvotes. Makes me think that articbull has multiple accounts and is just gaming the system.

Re:In 2019 RX470 were $70, RX580 $100 by robocat on Tuesday September 20, 2022 @04:04PM (#32917047)
> downvoted

You are very likely mistaking normal HN herd behaviour - plenty of people are going to think bullshit to “unused hydro”.

Also “Please don't comment about the voting on comments. It never does any good, and it makes boring reading.” and “Please don't post insinuations about astroturfing, shilling, bots, brigading, foreign agents and the like. It degrades discussion and is usually mistaken. If you're worried about abuse, email [email protected] and we'll look at the data.” - https://news.ycombinator.com/newsguidelines.html

If you assume that HN is full of good people, and respond accordingly, you will learn more and have more fun. If you assume HN is full of bad people causing trouble, why would you stick around? Good communities are built upon trusting cooperators and learning the valid signals of trust.

Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @04:34PM (#32917412)
Right after that update, I went from -4 to +2.

> plenty of people are going to think bullshit to “unused hydro”.

Maybe... but the way I went up and then suddenly went down to -4, was super odd.

> If you assume that HN is full of good people, and respond accordingly, you will learn more and have more fun..

I've been around here, with an account, since 2009. It is the only community I participate in any more. I believe that HN is full of great people and that the content is about as good as you're going to get these days. This is why accounts such as articbull are becoming boring to me. All it is is anti-crypto hatred that is often not fully factual. Equating me to a tobacco exec is just icing on the cake. Sorry for calling that out.

Re:In 2019 RX470 were $70, RX580 $100 by mattwilsonn888 on Friday September 23, 2022 @03:48PM (#32956263)
Not only have I debated this person multiple times, but on a search about settlement a HN thread came up and he had the top comment citing the infamous Mora et al study which makes the naive assumption that required hash for a proof of work network scales linearly with transaction volume, rather than block volume (which come at a constant rate no matter what the total hash is).

This person is a successful troll, and a pathetic thinker.

Re:In 2019 RX470 were $70, RX580 $100 by arcticbull on Tuesday September 20, 2022 @01:05PM (#32914668)
There's no such thing. That power should be connected to the grid instead of wasted or put to productive use, not to mention all those GPUs that are about to head for the landfill after a long life of guessing random numbers and warming up the outside. Wasting the power for money disincentivizes storage and grid connection.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @01:31PM (#32915011)
You are hilarious, your crypto hatred has clouded your judgement. Since you have so much karma here, you don't mind saying silly things and losing points left and right.

Transmission of power is expensive and lossy. These data centers are in a very remote location that isn't heavily populated. That's why they are there to begin with! The power would absolutely go to waste otherwise. That's also why we picked that area... power was cheap because it wasn't being used.

Our GPUs are 'new old stock'. They were already produced, sitting in warehouses not doing anything, and would have gone to waste anyway.

Update: Similar with the downvotes...

Re:In 2019 RX470 were $70, RX580 $100 by arcticbull on Tuesday September 20, 2022 @01:46PM (#32915213)
> You are hilarious, your crypto hatred has clouded your judgement. Since you have so much karma here, you don't mind saying silly things and losing points left and right.

First question, where do you think I got most of the karma? Sentiment here is broadly against crypto and has been for years (although it comes and goes). The pro side is loud, but enjoys far less support than you might imagine.

> Transmission of power is expensive and lossy.

No, it isn't. Transmission of power is super efficient. The EIA says the US lost only 5% of its electricity to T&D over the 2016-2020 period. [1] Expensive it may be, but that's not a reason not to do it. A single long power line from Boston to LA would only see ~25% loss from one end to the other. The whole Ohm's law thing (P=IV), we use transformers that are 99%+ efficient to raise the voltage to ~345kV and big fat wires to reduce current and resistance and hence transmission line losses.

> That's also why we picked that area... power was cheap because it wasn't being used.

It should be used to do something productive, not generate e-waste.

> Our GPUs are 'new old stock'. They were already produced, sitting in warehouses not doing anything, and would have gone to waste anyway.

Or, prices would have gone down and more people would have bought them. I can't see 3060s getting scrapped due to lack of demand when they were going for $850 a pop.

[1] https://www.eia.gov/tools/faqs/faq.php?id=105&t=3

Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @02:13PM (#32915668)
> Sentiment here is broadly against crypto and has been for years.

Agreed. However, I'd hope that 'against crypto' was less of one dude ranting endlessly, which is really boring and should get down voted.

> Transmission of power is super efficient.

We're not talking about efficiency. Transmission is expensive. Which goes beyond efficiency. You have to factor in the cost of the lines themselves, which are based on distance. I've seen these costs first hand, have you?

> It should be used to do something productive, not generate e-waste.

Oh yes, please define productive for me.

> Or, prices would have gone down and more people would have bought them.

They weren't for sale to the public. Nor would they ever have been.

> I can't see 3060s getting scrapped due to lack of demand when they were going for $850 a pop.

Huh? This whole thread is about RX4xx, which is what my cards are.

Re:In 2019 RX470 were $70, RX580 $100 by Reason077 on Wednesday September 21, 2022 @03:48AM (#32921943)
> "A single long power line from Boston to LA would only see ~25% loss from one end to the other."

It would actually be less than that using a modern HVDC transmission line. Losses for HVDC are around 3% per 1000 km of cable, and around 0.7% for the converter stations at each end. So around 15% loss for a theoretical Boston-to-LA HVDC.

Re:In 2019 RX470 were $70, RX580 $100 by dumbfounder on Tuesday September 20, 2022 @05:01PM (#32917752)
It’s the storage that’s the problem, not the transmission. Have talked with several in this business, their take is that it allows plants to have extra capacity to expand to meet peak needs, instead of flirting with disaster. Most plants are hard to spin up and down, so having an alternative way to monetize the energy creation is super important.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @10:38PM (#32920284)
In the past, it was aluminum smelting pots, which can't go cold or they take a huge amount of effort to spin up again. They built huge dams just for this [0].

When those plants shut down, the power, which was transmitted directly to the plants, had nowhere to go. Sending it somewhere else requires building high voltage transmission towers, which is very expensive to build as well as maintain.

Coinmint [1] took over the old Alcoa plant and is now pulling power from that dam. As it was explained to me by the head electrician when I visited, Coinmint is doing the dam a favor by having a constant draw of power to replace what was once the smelters. They also are paying for the electricity, which helps with maintenance of the dam, which is effectively a national resource.

[0] https://en.wikipedia.org/wiki/Moses-Saunders_Power_Dam

[1] https://www.coinmint.one/

Re:In 2019 RX470 were $70, RX580 $100 by Reason077 on Wednesday September 21, 2022 @04:11AM (#32922087)
> "It’s the storage that’s the problem, not the transmission... Most plants are hard to spin up and down"

This is true for nuclear and some types of fossil-fuel power plants. But for hydro, all you have to do to vary the power output is adjust the amount of water running through the turbines. Plus, hydro plants that have their own storage lakes/dams have built-in storage: any water you're not using for generation can be saved in the lake for when it's needed.

(OK, some rivers have minimum flow requirements so that fish don't die and water can be used downstream etc...)

Re:In 2019 RX470 were $70, RX580 $100 by schoen on Tuesday September 20, 2022 @03:50PM (#32916881)
According to the HN moderators, if you suspect someone is manipulating votes in some way, you're supposed to e-mail the moderators to have them look at it rather than making an accusation in your post.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @04:36PM (#32917448)
Sure. Apologies for calling that out in my update. That said, I don't feel it is worth @dang's time to waste on validating if someone with 24k karma is using fake accounts. As soon as I posted that update, the karma on the comment reversed.
Re:In 2019 RX470 were $70, RX580 $100 by superjan on Tuesday September 20, 2022 @02:44PM (#32916100)
I don’t particularly like crypto but I don’t think we need to rehash the same debate again. Isn’t it interesting to hear what this guy is going to do? Passing judgement does not improve the conversation.
Re:In 2019 RX470 were $70, RX580 $100 by NikolaNovak on Tuesday September 20, 2022 @03:02PM (#32916338)
>>I think an argument could be made that mined-on cards are more well cared for than those owned by your average gamer.

What strikes me though is people are rarely discussing mileage, only nature of work.

The video card in my gaming computer works, on average, maybe 1hr/week (some weeks more, most weeks 0). So right out of the gate, my video card has been working basically 168 times less long than any mining card running 24x7. We are talking more than two orders of magnitude difference in "mileage", before we discuss how well or poorly it has been taken care of.

Re:In 2019 RX470 were $70, RX580 $100 by LoveMortuus on Tuesday September 20, 2022 @05:26PM (#32918021)
Most of the "damage" that happens to a GPU is due to thermal expansion, which happens at to a more extreme degree if you're cold booting a game and then from a hot gaming state back down to cold. Where are with mining it's usually running constantly at a more or less stable temperature, thus it experiences much less of stress from the thermal expansion side of things.
Re:In 2019 RX470 were $70, RX580 $100 by belval on Tuesday September 20, 2022 @04:03PM (#32917033)
My understanding is that it's similar to a car, changing state (starting/stopping) is much more damaging that running the same thing, at the same temperature/voltage for extended period of time.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @04:39PM (#32917491)
I've rebooted GPUs hundreds of times to find the optimal tuning parameters. Failures after a reboot were very few and long term over the entire collection of GPUs, we've seen very few failures. Definitely within operating parameters.
Re:In 2019 RX470 were $70, RX580 $100 by Obi_Juan_Kenobi on Tuesday September 20, 2022 @02:26PM (#32915868)
They usually undervolt/clock the GPU, but not the memory. VRAM temps can be an issue and only some miners will try to address that properly with copper shims, etc.

The 3090 is notoriously bad for mining because it has memory chips on both sides of the board, leading to more heating in one spot and poor cooling for the 'back side' chip. Other cards tolerate it much better, but it can be an issue.

So while overall mining cards should be good, there are some specific issues to watch out for. A failed memory chip is not an easy fix, requiring a reball/replacement of that chip.

Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @03:08PM (#32916406)
100% accurate. ethash is memory bound. I've definitely seen 'memory go bad'... but at our scale, the percentage failure rate on that is in the single low digits. Additionally, we have cards that run just fine for years on end.

Maybe, as you suggest, it is just something that is more card specific.

Re:In 2019 RX470 were $70, RX580 $100 by synapse26 on Thursday September 22, 2022 @03:30AM (#32935773)
I fully agree with your point. However, this only applies when miners are buying new cards. I live in Vietnam, crypto mining/usage is pretty big, during the last crypto craze a few months (a year?) ago, all miners are willing to buy old cards, used cards, even broken cards at high price, I sold my 570 for $300.

Furthermore, flashing custom mining BIOS, what I’m told is that every miner here has to do or it wouldn’t be profitable, voids warranty.

Re:In 2019 RX470 were $70, RX580 $100 by captainbland on Tuesday September 20, 2022 @11:30AM (#32913096)
Yeah if I had to guess, the main concern with very long periods of load is going to involve moving parts rather than anything else. So pretty much the fan is likely to go but not much else.
Re:In 2019 RX470 were $70, RX580 $100 by npteljes on Tuesday September 20, 2022 @11:50AM (#32913411)
My logic would be that these cards are for consumer usage, and mining means 24/7 100% load. So, a year mining would be like 5 years of gaming, and even then you gamed 5 hours a day, every single day! So, mining might not destroy cards, but it's heavy use.

I think the used goods buyers' attitude is similar to those in the car market. Never mind the miles, if a car had heavy users, like a pizza company's delivery vehicle, it looks different in the market than a garage-stored family car used to visit the relatives.

Re:In 2019 RX470 were $70, RX580 $100 by hansvm on Tuesday September 20, 2022 @12:04PM (#32913646)
Counterpoint, heavy use mostly doesn't apply to solid-state devices (much how the expected lifespan of an incandescent bulb is independent of its previous usage), and to the extent it would matter the effect is probably dwarfed by not being stuck in a gaming PC full of dust and experiencing constant thermal shock as the games ramp up and down.

To the pizza car analogy, when I was in that game I put a hard 60k miles/yr on such vehicles, but I also regularly change the oil, keep great tires and brakes on them, replace bushings and other suspension components at the first sign of issue, etc... Partly because driving that many miles makes issues easy to spot due to familiarity and temporal coherence and partly because it is cheaper to fix many such things sooner rather than later after they've jostled something else out of place. Nobody ever asked if it was a pizza car. I listed it as well maintained and got top dollar each time because it was and their inspections showed that to be true.

Re:In 2019 RX470 were $70, RX580 $100 by falcolas on Tuesday September 20, 2022 @12:51PM (#32914455)
> heavy use mostly doesn't apply to solid-state devices

Sorry, but this is pretty much incorrect. Copper migrates, even in solid state devices, and those copper migrations cause breaks and shorts, which cause errors and failures. The smaller the scale, the faster this happens. There is absolutely a MTBF even for solid-state devices.

This entire slide deck is worth the perusal, if only for the "whoah, that actually happens?!" moments. But here's a relevant slide for this claim:

https://c3.ndc.nasa.gov/dashlink/static/media/other/Observed...

The money quote: "Commercial semiconductor road maps show component reliability timescales are being reduced to 5-7 years." (with attribution)

Another fascinating slide is this one, which shows how an IC can start to have a timing issue after being in service for a time.

https://c3.ndc.nasa.gov/dashlink/static/media/other/Observed...

Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @12:57PM (#32914564)
We've definitely seen 'memory go bad'. But, our failure rate is in the single digits, across a huge sample set across multiple years and in relatively extreme conditions (4 seasons, outdoors).

If you're buying something for like half of retail and rolling the dice on it, I think it is a fair bet. Even if you only get 1-2 more years out of a card that has been used before (for any purpose), then you've done well.

Re:In 2019 RX470 were $70, RX580 $100 by d_tr on Tuesday September 20, 2022 @01:47PM (#32915223)
Half of retail (which is still high) for a mining card with expired warranty? No thanks.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @02:07PM (#32915584)
Ok, $10, but you have to pay for shipping and min order is 10k. ;-)

For clarity, because @derac is confused below... this is a joke. Hence the winky face.

Re:In 2019 RX470 were $70, RX580 $100 by derac on Tuesday September 20, 2022 @02:01PM (#32915495)
Your incentive is to propagandize this point. you have 100k+ cards to sell on the market, by your own admission.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @02:06PM (#32915565)
It is amazing how crazy HN can get. I'm not propagandizing anything.

If we sell them for pennies on the dollar or e-waste them, doesn't matter to me.

We ROI'd these cards already.

Re:In 2019 RX470 were $70, RX580 $100 by seattle_spring on Tuesday September 20, 2022 @04:07PM (#32917091)
Dude it's nuts how often people sling the "shill" insult on HN. One of the laziest insults by far, and I'd guess incorrect very near 100% of the time.
Re:In 2019 RX470 were $70, RX580 $100 by derac on Tuesday September 20, 2022 @02:20PM (#32915781)
Whoa crazy to point out someone's conflict of interest. CrAzY

You are sitting on 10m+ of stock and you don't care about the money. OK bud guess I must be schizophrenic, damn.

You are literally trying to sell your cards to a sister comment...

"Ok, $10, but you have to pay for shipping and min order is 10k. ;-)"

Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @02:27PM (#32915884)
There is no conflict of interest, I'm not trying to sell cards here on HN.

You are literally making stuff up now. The 10m+ of stock is worthless if nobody would buy it. If anything, it'll cost us to e-waste it. So, the price ceiling, if we were to sell it, is simply the cost e-waste and moving it around.

That was a winky face... a joke. I was kidding because the OP was unhappy with the statement of half price, which I also made up as a random baseline.

Re:In 2019 RX470 were $70, RX580 $100 by derac on Tuesday September 20, 2022 @02:31PM (#32915927)
The goal is changing the public discourse. Thus the word propaganda. You wouldn't be propagandizing this point of view if you didn't think you would change people's minds. You are fully aware that this forum has an outsized influence on the tech/programming community.

Edit: to me it's clear that you are acting in bad faith since you called me crazy for pointing out an obvious conflict of interest.

Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @02:45PM (#32916101)
1. I didn't call you crazy, I said "it is amazing how crazy HN can get."

2. Tell me exactly how would it benefit me to change public discourse around the long term reliability of GPUs? The most I can think of is that it might drive up the prices of used cards a tiny bit in a very saturated market, but I doubt it. Especially since I'm just one data point and I'm only posting here on HN. Plus, as I said, I don't care what our cards sell for. Our baseline cost at this point, is to offset any price of e-waste and shipping.

3. I shared my findings about long term reliability because this is a tech community and I thought they'd be interested in it! OMG, I'm such a horrible person for thinking this way! That is sarcasm.

4. Allowing people to realize that mined on GPUs are actually ok (not all, but generally better than common thought) might mean that a few more find useful homes and not end up as e-waste. That's a good thing! I even suggested that people use it as a way to negotiate the price down!

Re:In 2019 RX470 were $70, RX580 $100 by derac on Tuesday September 20, 2022 @03:44PM (#32916804)
Fwiw, I don't think you're a bad person. I think you have a conflict of interest and they way you are responding to criticism is weird. You also handwaved the copper migration issue. Meh. Carry on. Personally, I took that as you calling me crazy. If that wasn't your intention, fair enough.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @04:48PM (#32917603)
Your criticism is to say that I'm spreading propaganda, which is untrue and it is fair for me to respond to that.

Now you're at it again with the criticism. I'm not handwaving anything. I'm stating what I saw after having the fairly rare experience of running 100k+ GPUs over multiple years in harsh environments.

Let's turn that around, do you have any experience with this at all? Or do you just criticize people?

It also doesn't benefit me in any way to share information like this. I'm only sharing this information now because we're pretty much done with this phase of the business, so I can talk about things a bit more openly, and it seems relevant to share on a thread about RX cards.

Re:In 2019 RX470 were $70, RX580 $100 by derac on Tuesday September 20, 2022 @05:15PM (#32917879)
I didn't say it was not ok to respond. Calling people crazy is the weird part.

You did handwave that, we can agree to disagree.

I didn't even say you are or aren't doing anything nefarious. I pointed out a conflict of interest (incentive to propagandize) and you responded by calling me crazy. Now you are saying "All I do is criticize". Boohoo to you sir.

Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Wednesday September 21, 2022 @10:44AM (#32925897)
Good luck with everything. I wish the best for you.
Re:In 2019 RX470 were $70, RX580 $100 by seattle_spring on Tuesday September 20, 2022 @04:20PM (#32917253)
"Assume good intent." This comment, as with all of your other ones in this chain, blatantly violate HN guidelines. Accusing people of "shilling" or "spreading propaganda" only makes this forum worse.
Re:In 2019 RX470 were $70, RX580 $100 by uhhyeahdude on Tuesday September 20, 2022 @01:08PM (#32914709)
Thanks for that slide deck! Many treasures to be found within.
Re:In 2019 RX470 were $70, RX580 $100 by falcolas on Tuesday September 20, 2022 @03:00PM (#32916291)
Any time I share it, I go through it again. It's so dang fun.
Re:In 2019 RX470 were $70, RX580 $100 by vineyardmike on Tuesday September 20, 2022 @12:44PM (#32914329)
Counter counter point, chips break down over time even though they’re solid state. The flow of electricity through a chip will slowly degrade the metal traces inside it. At the nanometer scale, it doesn’t take much to ruin a logic gate. This is called electromigration.

You can care for that car (or chip) all you want, but chips aren’t cars. It’s a slow process, and may not happen on the timescale of mining, but it happens.

https://en.wikipedia.org/wiki/Transistor_aging

https://en.wikipedia.org/wiki/Time-dependent_gate_oxide_brea...

Re:In 2019 RX470 were $70, RX580 $100 by AceJohnny2 on Tuesday September 20, 2022 @12:48PM (#32914402)
> heavy use mostly doesn't apply to solid-state devices

It kinda does. The effects are small, but increase as the fab node gets smaller.

https://en.wikipedia.org/wiki/Electromigration

Re:In 2019 RX470 were $70, RX580 $100 by ithinkso on Tuesday September 20, 2022 @01:26PM (#32914939)
> (much how the expected lifespan of an incandescent bulb is independent of its previous usage)

Either I'm reading this sentence incorrectly or there is something else I don't understand. How can that statement be true?

Re:In 2019 RX470 were $70, RX580 $100 by rowanG077 on Tuesday September 20, 2022 @03:20PM (#32916535)
It's possible if you have a completely flat failure rate. For every time unit spend it has x% failure rate. An example is a piece of rock on the side of the road. Everyday it has a small chance to be smashed by a car. But that chance doesn't increase or decrease over time. Unless external factors change

Whether it's true for light bulbs I don't know.

Re:In 2019 RX470 were $70, RX580 $100 by HenriTEL on Tuesday September 20, 2022 @04:57PM (#32917701)
I guess the bulb does not degrade over time, it just break at some point when some random events occur.
Re:In 2019 RX470 were $70, RX580 $100 by ithinkso on Tuesday September 20, 2022 @05:23PM (#32917983)
I don't think that's true, the filament evaporates (sublimes?, sublimates?) from the heat and eventually it is too thin to support itself and breaks. That's how I always understood the main failure mode of incadescent bulbs and that's why the above statement confuses me
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @11:57AM (#32913524)
There is no such thing as 100% load in ETH GPU mining. The bottleneck is the memory controller, not the GPU itself. That's why 5+ year old cards are still more ROI efficient than the latest greatest cards. The only way to speed things up is to speed up the memory.

Other algos take advantage of more of the GPU side of things as well. ProgPoW, if it had passed, would have done that. From what I understand, RVN also does as well... and uses more power too.

Re:In 2019 RX470 were $70, RX580 $100 by garaetjjte on Tuesday September 20, 2022 @12:04PM (#32913654)
On the other hand 24/7 usage means no thermal cycling, probably eliminating failures caused by solder joint cracking.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @12:44PM (#32914340)
The machines reboot a lot for auto tuning. Plenty of thermal cycling that happens in all temperatures. Still didn't see any issues.

PSU's, on the other hand, are mostly hand soldered in China and have all sorts of issues.

Re:In 2019 RX470 were $70, RX580 $100 by Scramblejams on Tuesday September 20, 2022 @01:56PM (#32915364)
Who/what do you like for PSUs for operating in that harsh environment?
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @02:50PM (#32916182)
We've run a ton of different PSU's. Everything from super cheap no-name chinese stuff to refurb/tested high end HP server class. None of them are super reliable.
Re:In 2019 RX470 were $70, RX580 $100 by PartiallyTyped on Tuesday September 20, 2022 @03:15PM (#32916478)
I trust mining GPUs a lot more than gamer owned GPUs tbh. Better ventilation, severe undervolting, not even maxing the gpu itself and the fans are not actually spinning that fast.
Re:In 2019 RX470 were $70, RX580 $100 by Gordonjcp on Tuesday September 20, 2022 @02:51PM (#32916193)
The garage-stored car will be a bag of shite, though. You'd be far better going for the delivery vehicle.
Re:In 2019 RX470 were $70, RX580 $100 by yieldcrv on Tuesday September 20, 2022 @01:13PM (#32914766)
undervolted use because not every processing component needs to be used

very different than a car

but valid questions to ask to whoever you are buying from used

Re:In 2019 RX470 were $70, RX580 $100 by simias on Tuesday September 20, 2022 @12:50PM (#32914432)
Yeah I don't really understand this worry about buying cards used for mining. If you're worried about wear I'd still rather take a miner's card than one that was used by a heavy gamer and routinely pushed to the limit.

I feel like people dislike cryptominers (IMO reasonably, and for all sorts of reasons) and want to punish them all the way by preventing them from liquidating their stocks of cards, but I don't think it's a good reason to make stuff up.

I wouldn't buy a card that I know was used for mining because I'm effectively boycotting miners, not because of some made up technical excuse.

Re:In 2019 RX470 were $70, RX580 $100 by Melatonic on Tuesday September 20, 2022 @01:16PM (#32914814)
Last crypto crash years ago I bought a used Titan XP for an insane price after the miners started having to dump them super cheap. Still running it today and it has chewed through even the craziest games. They probably bought it for way more so I don't think it is really effective to "boycott" them by not buying the card if you are the one getting a super good deal.
Re:In 2019 RX470 were $70, RX580 $100 by SilverBirch on Tuesday September 20, 2022 @01:06PM (#32914677)
The only problem I have with miners is that I don't understand the economics. My understanding is that air conditioning in data centres is a non-trivial cost. So the question is pretty much a balance of: Is air conditioning so expensive that it's cheaper for miners to blast these cards at high temperatures and accept a failure rate, or is the maths behind (cost * mean time between failures) so high that the miners are actually selling off good used equipment.

You could tell me that either of those things dominate, I don't know. But I would want to have atleast an idea of this before I bought a used miners GPU.

Having said that, I wouldn't buy a gamer's old GPU.

Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @01:22PM (#32914899)
The economics are simple: capex + opex. As long as you have an idea of ROI and can lower the cap+op as much as possible, you're making money. Otherwise, you're buying the coin directly.

You can't run the cards at high temps or they crash. They also have built in thermal protection where they shut down at about 85c (depends on the settings).

Miners also undervolt/underclock the cards. Lower power is better for opex in the equation above. =) This means they aren't run as hard as you might think. ethash as an algo is also memory hard, so the GPU portion isn't 100% at all. That's why older cards are still in use and more ROI efficient.

Re:In 2019 RX470 were $70, RX580 $100 by hot_gril on Tuesday September 20, 2022 @01:01PM (#32914621)
I'd not be worried about wear but about whatever meddling was done with the firmware (edit: or hardware). Maybe an experienced miner knows how to easily return it back to stock, but who knows who you're dealing with. Someone else can take the chance.

Also, PC gamers are the last group I would expect to successfully boycott something. They'll install some spyware-ridden game launcher just to get a $5 discount on a game, while complaining about it. What discount will they resist for a mined-on GPU? $50? Assuming they even know it was mined on.

Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @01:22PM (#32914902)
Easy to reset the firmware (vbios) to stock.
Re:In 2019 RX470 were $70, RX580 $100 by ok123456 on Tuesday September 20, 2022 @01:36PM (#32915077)
Hardware fuses.
Re:In 2019 RX470 were $70, RX580 $100 by hot_gril on Tuesday September 20, 2022 @12:19PM (#32913906)
I believe you out of sheer respect for your auto card tuning project, but in general someone with $1M of a product to sell has a conflict of interest with giving good advice about its quality.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @12:42PM (#32914306)
I have a lot more than $1m (new) of these cards. 6 digits of CARDS, not value.

We have zero cards for sale right now. Nobody wants them because they are heat sinks and not fans.

Re:In 2019 RX470 were $70, RX580 $100 by chungy on Tuesday September 20, 2022 @12:23PM (#32913986)
Conflict of interest doesn't mean wrong, but is an indicator that one should take the claim dubiously until it can be backed up.
Re:In 2019 RX470 were $70, RX580 $100 by lampshades on Tuesday September 20, 2022 @11:17AM (#32912889)
Not the same card but five years ago I bought a refurbished GTX 1080 Ti that was previously used for mining. Still running smoothly to this day, no issues whatsoever.
Re:In 2019 RX470 were $70, RX580 $100 by PartiallyTyped on Tuesday September 20, 2022 @05:46PM (#32918222)
Same but around 3 years ago. Likely the best purchase I have made.
Re:In 2019 RX470 were $70, RX580 $100 by hydroev on Tuesday September 20, 2022 @11:42AM (#32913292)
but the perception is already that gpu used in mining are tough to sell as buyers know most of the GPUs bought in the past few years were most likely to have been used in mining.

seeing more and more people just waiting out to buy the newer cards so it seems like the price needs to come down a lot more to convince people to buy a gpu used in mining.

just my two cents based on observing reddit threads

Re:In 2019 RX470 were $70, RX580 $100 by Waterluvian on Tuesday September 20, 2022 @12:14PM (#32913836)
FWIW: "the paste can be replaced easily" is 100% true for like 1% of GPU consumers. As long as you replace the paste before resale, it's great. Otherwise the buyer might perceive that they have a card that's dead or slowed way down because "mining destroyed it."
Re:In 2019 RX470 were $70, RX580 $100 by hot_gril on Tuesday September 20, 2022 @02:04PM (#32915532)
The seller can replace the thermal paste first if it's so easy. It's in their interest to fetch higher prices. Not my job.

I was just thinking today about many years ago when I test drove a used car and heard whistling from the roof. Saleswoman said it was easy to fix. It must've happened during every test drive and possibly dissuaded some buyers, so why hadn't the dealer already fixed it? I bought the car, and guess what, the whistling wasn't fixable.

Re:In 2019 RX470 were $70, RX580 $100 by nathanvanfleet on Tuesday September 20, 2022 @11:31AM (#32913121)
Just wondering if, for a second, you considered that repasting or replacing fans is a large effort for normal folks who just want a video card?
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @11:36AM (#32913198)
This sounds like the perfect opportunity to buy a retail off the shelf card with a warranty!

What kind of question is that.

Re:In 2019 RX470 were $70, RX580 $100 by derefr on Tuesday September 20, 2022 @11:46AM (#32913358)
Presumably they bought these cards, took the fans off them, hooked up a custom water-cooling setup, etc.; and when selling them, are taking the pile of fans they shucked and putting them back on the cards with fresh paste.
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @11:53AM (#32913458)
We bought the cards directly from an OEM, with heatsinks on them instead of fans.

We use a case based cooling system. ie: an enclosed box with fans. 12 cards in a box. This is lower power, fewer fans (5 instead of 12), and better cooling than putting fans on each card. Also easier to replace a fan on a case than a card.

At my scale, water cooling doesn't work and isn't cost efficient. Air is a surprisingly great cooling system.

Re:In 2019 RX470 were $70, RX580 $100 by Melatonic on Tuesday September 20, 2022 @01:22PM (#32914904)
What does the actual heatsink look like one the card? Wondering if I can do this on my home system assuming I have good airflow on my case
Re:In 2019 RX470 were $70, RX580 $100 by latchkey on Tuesday September 20, 2022 @01:40PM (#32915138)
It looks just like a heatsink. Big aluminum bladed block.

It might work, I don't know... airflow on the box itself wasn't enough, but there were 12 cards in the box, we also had to have external airflow as well. A single card probably wouldn't be too much of an issue though.

That said, we aren't parting out single cards... it would be more like if you wanted to buy 5-10k of them. ;-)

Re:In 2019 RX470 were $70, RX580 $100 by mindw0rk on Tuesday September 20, 2022 @12:00PM (#32913580)
It's not just the thermal paste that goes. There are more parts that are affected by the constant heat. Capacitors in the first place
Re:In 2019 RX470 were $70, RX580 $100 by thfuran on Tuesday September 20, 2022 @12:38PM (#32914234)
Yeah, electrolytic capacitors derate due to time under load, exponentialy accelerated by temperature increases.
Re:In 2019 RX470 were $70, RX580 $100 by sfmike on Tuesday September 20, 2022 @01:19PM (#32914856)
what are you mining these days?