Techfast is back with another great GPU deal
Gainward GeForce RTX 3070 8GB Phoenix (LHR) $1199
A quick comparison with 3070Ti, I think this is a great price and you don't have to deal with the notorious hot gddr6x
Gainward GeForce RTX™ 3080 10GB Phantom (Non LHR) $1788
Some extra stock added for the popular Gainward 3080, all previous remaining orders are being fulfilled today
People who ordered this 2 weeks ago still haven't received theirs. Could be quite a wait on this.
All fulfilled as of today. These are all in stock too.
damn, already sold out of 3080s? always too late :(
@luketechfast: Hey @luketechfast. I ordered two 3080 cards on the 13th of July. I received one of the cards on Wednesday the 28th. What happened to my other card? I was under the impression they would both be delivered together since the invoice listed the serial number for two cards. However, only one arrived. The post package also says 1/1 so it’s not like one got lost in transit. Can you find out what happened to my other card? TF4656932631. Thanks.
@luketechfast: Hi Luke, I am waiting for my 3070 and I sent multiple emails to the support. Any chance you can help me with ETA? Thank you
Mine got shipped as of 40 minutes ago
Mine just got shipped too :D
I ordered my card two weeks ago and today received a tracking number when this ad was posted.
Very upset about the way they treated us
@Justintime88: well luke did say that the card will ship between now and the end of the month. so its the 28th now. its late but within what he originally said so I'm not mad. the one im annoyed about is i-tech because they said 3-4 weeks when I ordered the card and that slowly turned into 8-9 weeks. it was a waste of time but at least they refunded me after telling me at week 6 it would be another month.
its hard to point fingers though it isn't always the retailers when these things happen. its only when they repeatedly or widespread take your money for a period of time and don't deliver you start to think now this is getting ridiculous can you not adjust your timeframe based on past performance?
@Justintime88: Why? What did you need them to do? They said they'd be shipped by the end of the month. That's exactly what's happened?
RTX 3070 is down to a 60% premium versus US MSRP (adjusted for GST), maybe a bit above 50% for AU MSRP (exchange rates have shifted since then).
Still not a buy IMO, but progress.
Waited this long, might as well wait for next gen lol.
Is that next year around Nov ?
Time until next-gen gets announced: Unknown…
Time until next-gen is readily available at non-scalper prices: Unknown + a lot longer…
I'm assuming next-gen won't be as bad.
NVIDIA will likely from the get go, make every card an LHR variant, so the mining performance will be heavily reduced from release. On top of that, with a big market like China having mining farms being banned, it also will contribute to stock not being nearly as in high demand. Plus, NVIDIA will have CMP (mining card) replacements.
@KARMAAA: While that's true, it will still be improved mining performance. The hope is that ETH is PoS by then.
@jasswolf: It's all about perf per watt when it comes to mining. So we'll just have to wait and see. Hopefully, it's not that profitable and thereby the incentive to mine is lost.
@jasswolf: Yea, hopefully by next year, but there's plenty of other coins to mine apart from Ethereum.
From leaks of the 4090 details, they'll be moving to a different node so they won't be sharing with consoles.
That said, nVidia is rushing the 4090 because AMDs RDNA3 is releasing Q3-Q4 and it doesn't look like nVidia was ready for such a quick jump between generations. It looks like it'll only be the absolute top end cards in limited numbers for quite a while (much like we haven't seen the 6600XT/3050 yet, but even higher end).
@freefall101: They don't share a node now with consoles. Both consoles use 7nm from TSMC. NVIDIA is using 8nm (a 10nm improvement node) from Samsung.
@KARMAAA: Ah, you're right, I confused a few things. Reportedly the RTX 30 Super will be coming from TSMC and AMD are already using TSMC, Samsung is just having supply issues.
@freefall101: Where is the rumor RTX 30 SUPER series is coming from TSMC? Literally no credible rumors have come out for that to be the case. The best we've heard is that an RTX 30 SUPER is a possibility and that a new chip called GA103 will be coming and it will still be from Samsung, GA102 and GA104 are both Samsung and currently are in the RTX 30 Series cards.
@freefall101: NVIDIA have had a 2 year cadence for a while, they're ready. The way these rumours tend to play out is that AMD gets really excited and pounds their chest, only to find that all the advancements they've made generally hold them at parity (even with a node advantage).
AMD are getting closer again though, and making architectural advancements of their own, so you'll see NVIDIA using a more advanced node (current rumour is a TSMC 5nm process, either N5 or N5P), and that will allow for a substantial clock jump, as well as much as 2x density compared to Samsung 8N.
And all of that is without considering DLSS as a performance accelerator, but NVIDIA needs to bring a bigger cache bump compared to what they did with Ampere (AMD currently has the advantage there, and it really came to the fore below 4K).
@jasswolf: It's the same cycle year after year, and I'm not sure why people haven't realized this yet.
AMD talks about a revolution in performance, but after release the top cards end up just slightly eclipsing last-gen nvidia cards, and people act like that's the end of nvidia. Nvidia responds by releasing a mid/mid-high card (like the 3060ti) that deflates them a bit.
Then soon after AMD releases a genuinely good middle and low-middle range that blows everything else out of the water, so nvidia responds with a mid-generation update that brings them close to parity in the middle.
Then nvidia releases their next generation top cards, which blows everything out of the water again, and the cycle starts anew.
Which means that the best cards to buy are always AMD low-mid/mid cards (RX580 etc), nvidia mid/mid-high cards from the mid-generation update (2060S etc), and nvidia top cards (3080 etc) at launch. Everything else is going to be outclassed within a year because they're waiting to launch the products they actually care about.
AMD doesn't care about high performance cards, nvidia doesn't care about low performance cards, so buying them is like getting a kebab at a pub.
@Jolakot: I mean the RX6000 series is good value at all price points until RTX (RT + DLSS) factors in, and NVIDIA hasn't had enough demand to dip below the 3060 for desktop cards, yet.
I'd imagine we'll see something like an RTX 3050 Super appear in the next series of launches, and with any luck, a card below that. The 3050 is available for mobile though.
Solid strategy, though it's about 14 months away at this stage.
We'll learn more about next-gen architectures (there may be two from NVIDIA) in March-April next year.
LEt's skip 30 Series if they dont offer MSRP.
3070 non LHR possibly would make you back some money while others wait for the price to fall. However non LHR at this price.. doesn't make sense to me. For those who are wondering, 3070 is better than 3080 non LHR for mining Etherium in your non gaming time
And then there's most of the population, who want nothing to do with crypto.
Its their loss.
Yes, most of the population don't want anything to do with crypto at the moment, that will change.
In the 90s, most of the population didn't want anything to do with the internet, dial up modems were just too technical, internet was for the geeks only, look at it now?
@techlead: And then ISP's came in and crushed all the small BBS just as the governments or corporations will do to small time miners. Won't take much, small changes in the environment policies, or maybe the governments simply decide to push their own crypto.
Either way I'd still prefer to offer my computers downtime to those enterprises which help people as a whole, like SETI/BOINC for example.
@Ebany: What can governments or corporations do to crush small time miners? All I'm doing is running a computer with 6 or 12 graphics cards attached to it 24/7?
Are they going to make it illegal to turn on a computer 24/7? That doesn't sound like a "small change" to me. As for deciding to push their own crypto, well, good luck, introducing their own crypto doesn't automatically ban other cryptos.
@techlead: They're going to regulate what kind of power devices should be consuming for a given function, and so long as there are emissions associated with that, then they're going to fine and charge you for wildly exceeding that, as you will with cryptomining.
As they rightly should, because the blockchain isn't the future of money, it's the future of online transactions. There's a difference.
Eventually they'll be tallying up the silicon you're wasting on dead end computational tasks.
They're going to regulate what kind of power devices should be consuming for a given function, then they're going to fine and charge you for exceeding that.
So power rationing? How do they know what device is consuming the power?
That's pretty draconian, they don't even do that in China haha. I doubt we'd ever see restrictions like that here in our lifetimes.
@techlead: It's already happening, and not just in China. It's a huge problem with TVs, computers, tablets and phones.
Upcoming TVs and monitors will come with hardware restrictions in place for brightness while in SDR mode to encourage people to use them apporiately (both for power consumption and their own health).
@jasswolf: That's a hardware restriction for those components, very similar to the hashrate limiters in GPUs. How can this applied to computers which is a collection of different components? You mean hashrate limiters on computers themselves? That's quite difficult to achieve and I'd like to see them try.
In the meantime, I will continue to mine with my 3080s and enjoying the profits.
There's always a risk in everything, including regulation risk for crypto, this existed since the beginning and if I didn't enter the crypto space because of it, I would have lost millions, so let's see what happens when it actually happens.
@techlead: I'm talking about policy to better define what typical system usage looks like, and then energy companies getting mandates from governments on how to charge for excess (and where to send the money). If people want to suck down way more power than an individual user should be, they can pay for the excess pollution and then some.
In terms of crypto, I wouldn't be ruling out speculative risk to the level of watching everything essentially disappear. Regulation will completely water down your holdings, and I wouldn't be shocked if anything resembling a buy back scheme would merely seek to assist mum-and-dad type investors recover losses, rather than aid the absurdity that's existed in that market for years now. I think they'd cap that out, and frankly, I'd welcome it.
@techlead: Fair enough, but I'd hope at the very least you spend some of those millions on solar panels.
@Ebany: I already have solar panels. When I purchase a new property I will install them if it doesn't already have it. Tesla roof solar panels looks good.
What price are you getting the 3070? 3070 hashes at 58Mh/s while the 3080 gets 100Mh/s maybe more if you tweak it.
I have the 3080s in my rigs, bought them for $1600-1800 around launch, they've all paid themselves back, its pure profits after power costs for me. :D
Power to hash is high. Check minerstat
If only it was Gigabyte or MSI gaming trio
Cool story.
Aside from DLSS and Raytracing, I don't see why anyone would buy these over a 6700 XT.
They have better supply, are cheaper with better quality cards, and way more GPU memory…
Are they actually cheaper? I haven't seen this
At the moment it appears to be around the same price, but better quality cards and more availability.
https://www.scorptec.com.au/product/graphics-cards/amd/88757…
This Sapphire Nitro is always a good AMD card, Hardware Unboxed has reviewed the previous versions favorably.
You can get a 6700xt for $1050, which is $150 cheaper.
Eg: https://www.bpctech.com.au/catalog/product/view/id/38436/
And: https://www.ple.com.au/Products/645884/MSI-Radeon-RX-6700-XT…
Or even $1029 here: https://www.ebay.com.au/itm/MSI-Radeon-6700-XT-MECH-2X-12G-/…
The extra $150 is worth it for raytracing and DLSS
If the MECH is anything like the 5000 series it isn't worth buying. The cooler is horrid, it's only good if you're gonna liquid cool it or something.
If you're an Adobe user then Nvidia and cuda cores are a must unfortunately. Especially with Premiere there's a substantial difference in gpu accelerated performance compared to AMD.
Bingo mate
DLSS and raytracing are pretty big reasons to get a Nivida card! Especially for a slight premium over AMD cards
DLSS sure, but AMD has already announced and made their alternative version open source, thus will be more readily adapted by game developers.
Source: https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&c…
Plus Raytracing is available in a very small sample of games, and already has many replacements on the way that may even make it redundant due to the significant frame rate penalty.
The thing is FSR has a upper limit to quality improvement since it's a spatial technique. For instance, if you reduce the shading by 25%, then the best it can have is 75% of the quality of the native resolution.
DLSS 2.0+ is only limited by the amount of pixels a monitor or TV can produce since it's a temporal upscaler, so theoretically, it has the same limits as native resolution in terms of an upper bound.
FSR is very good as a spatial upscaler, but it's primitive compared to DLSS 2.0+. Both will improve, but it's hard to not pick DLSS over FSR, especially when the source image is a low resolution like 720p or 1080p.
@KARMAAA: DLSS sure is the superior technology, with FSR being technically 'dumb' in comparison.
However this comparison is redundant, because DLSS can only be 'chosen' in the games it's developed for, which when developers now have an open source alternative may shrink over time.
Open source is a huge deal in this respect, it means anyone can develop it and it's compatible with any hardware. This basically secures the future of FSR over DLSS, and was a step of strategic brilliance from AMD.
If DLSS requires significant development time and effort for games over FSR, I can see developers either needing to be incentivized to use it by Nvidia, or just avoiding it for FSR in the hope Nvidia will add compatibility for FSR in the future.
@thedean: DLSS is also now available for any developer to use, just download the SDK and in fact it's more available than FSR since it's directly available to put into any Unreal Engine game via the Unreal Engine Store, whereas FSR needs to be manually added by the developer. The same thing will happen soon for DLSS in Unity. DLSS now has a general model too, so it no longer needs to be trained per game. Any advantage FSR "had" is gone, other than being available for most if not all cards. NVIDIA will work out a way to use a liter version of DLSS through FP32 or FP16 and it'll really be game over.
FSR is nothing like DLSS though. FSR just upscales each frame individually basically, whereas DLSS takes what is happening in the frame and previous frames into account. The DLSS SDK is now free for all to use as well.
FSR might be implemented more, but it's significantly worse so who really cares? DLSS can make games look better than native resolution (see Doom Eternal - 1440p + DLSS to 4K looks better than native 4K).
What are these "many replacements" for raytracing that you speak of btw?
What are these "many replacements" for raytracing that you speak of btw?
Unreal Engine 5's Lumen provides a large majority of the functionality, quality and features that are currently implemented using HW raytracing, completely in software. CryEngine is also introducing it's competitor to Lumen shortly as well, which uses full software raytracing with similar quality, but without the same performance impact that actually tracing rays every frame requires.
Not to mention, raytracing as implemented on RT Cores is in some ways technically inferior over classical GI and reflection techniques, as employed by UE5 and CryEngine. For example, RT in NV has a really bad problem with persistence and "after glow" as light sources change in a scene. This leads to smeary looking reflections in fast moving objects, and a low quality smudge of light persistence after lights are turned on and off in a scene. This problem is fundamental to the RT technology as implemented on RT Cores. The RT Cores aren't fast enough to get enough samples per frame, so they collect samples across multiple frames and then average them out to reduce noise, leading to the issues I mentioned earlier.
Now this isn't to say RT is dead in the form of RT Cores, but I think a lot of game developers and gamers will be more than happy with the level of quality that UE5 and CryEngine will provide with their next generation lighting engines, which is going to significantly impact the value proposition of "real" Ray Tracing, especially when it comes with such huge performance impacts. Lumen can run at full resolution without any need for "hacks" like DLSS to claw back performance.
Also CUDA, NVENC 265 encoder and Nvidia Broadcast are very good utilities that some people can't live without if they have applications that take advantage of those.
I absolutely agree here, NVENC is amazing if you need it, such as running a plex server or streaming game play. Also Nvidia voice I use daily from a laptop is amazing.
However, if you do none of these things, I think the AMD card represents much better value for money at this moment. Will be very interesting to see how to prices play out after the ETH hard fork in 9 days destroys mining profit and cards likely flood the second hard market.
@thedean: People usually keep a card for a few years, so you might not need it now but some day you could want to have it
You'll only see a flash sale of mining gpus if ethereum drops below $1000, still plenty of money to be made post eip1559
@ln28909: A likely cut in profit of between 25% and 50% is not entirely 'plenty of money'.
The London hard folk is designed to significantly reduce transaction fee's by burning a portion of the fee (that would normally go to miners).
If they had occurred over the last year, 76% of the newly created ETH from transaction fees would have been burned. That's a huge amount that won't be going to miners.
Source: https://www.coindesk.com/why-ethereums-london-upgrade-matter…
Personally I think the majority of cards showing up on the market will be worn our 5600xt's and 8gb 570's.
@thedean: yes eth mined per day will drop by around 25% of what it is today, still $6ish dollar before cost for 3080, there are people used to mine when profitability is like $1/card. As a miner, I would not consider selling unless Eth price goes below $1000
@thedean: If I was a betting man I'd wager than all these Gainward 3080's will ship the day after EIP1559 hits the main net.
At the moment they are making a nice profit for who ever has them. It's such a good business model.
A likely cut in profit of between 25% and 50%
Impossible considering that only transaction fees are being reworked. Static block reward will still be 2 ETH and uncle rewards will be unaffected.
Looking at data for the last 60 days shows that on average the transaction fees only account for 16% of the total rewards.
Keep in mind that transaction fee will be replaced by a base fee + tip with only the base fee being burned so the cut in profits may end up being even lower.
| New supply from mining | Tx rewards | Total rewards | Tx rewards % | |
|---|---|---|---|---|
| 29/05/21 | 13,526.31 | 3,095.52 | 16,621.84 | 18.62% |
| 30/05/21 | 13,435.63 | 2,644.97 | 16,080.59 | 16.45% |
| 31/05/21 | 13,391.06 | 3,078.77 | 16,469.83 | 18.69% |
| 01/06/21 | 13,461.88 | 2,835.01 | 16,296.88 | 17.40% |
| 02/06/21 | 13,570.56 | 2,966.32 | 16,536.89 | 17.94% |
| 03/06/21 | 13,374.75 | 3,057.51 | 16,432.26 | 18.61% |
| 04/06/21 | 13,611.25 | 2,606.23 | 16,217.48 | 16.07% |
| 05/06/21 | 13,356.50 | 1,837.06 | 15,193.56 | 12.09% |
| 06/06/21 | 13,588.19 | 1,560.43 | 15,148.62 | 10.30% |
| 07/06/21 | 13,403.25 | 2,102.20 | 15,505.45 | 13.56% |
| 08/06/21 | 13,478.25 | 2,852.44 | 16,330.69 | 17.47% |
| 09/06/21 | 13,413.88 | 2,373.72 | 15,787.59 | 15.04% |
| 10/06/21 | 13,375.44 | 2,000.59 | 15,376.03 | 13.01% |
| 11/06/21 | 13,449.06 | 2,281.28 | 15,730.34 | 14.50% |
| 12/06/21 | 13,397.13 | 1,912.18 | 15,309.31 | 12.49% |
| 13/06/21 | 13,435.25 | 1,466.30 | 14,901.55 | 9.84% |
| 14/06/21 | 13,498.00 | 1,929.35 | 15,427.35 | 12.51% |
| 15/06/21 | 13,575.88 | 2,035.56 | 15,611.44 | 13.04% |
| 16/06/21 | 13,326.88 | 1,937.69 | 15,264.57 | 12.69% |
| 17/06/21 | 13,514.00 | 2,110.16 | 15,624.16 | 13.51% |
| 18/06/21 | 13,298.94 | 2,085.18 | 15,384.12 | 13.55% |
| 19/06/21 | 13,457.75 | 1,486.77 | 14,944.52 | 9.95% |
| 20/06/21 | 13,455.44 | 1,710.91 | 15,166.35 | 11.28% |
| 21/06/21 | 13,201.25 | 4,129.55 | 17,330.80 | 23.83% |
| 22/06/21 | 13,486.00 | 6,598.90 | 20,084.90 | 32.86% |
| 23/06/21 | 13,412.25 | 2,152.92 | 15,565.17 | 13.83% |
| 24/06/21 | 13,334.25 | 1,909.34 | 15,243.59 | 12.53% |
| 25/06/21 | 13,179.19 | 2,461.55 | 15,640.74 | 15.74% |
| 26/06/21 | 13,413.31 | 1,489.40 | 14,902.71 | 9.99% |
| 27/06/21 | 13,435.38 | 2,486.77 | 15,922.14 | 15.62% |
| 28/06/21 | 13,525.88 | 2,250.97 | 15,776.85 | 14.27% |
| 29/06/21 | 13,353.81 | 2,244.09 | 15,597.91 | 14.39% |
| 30/06/21 | 13,457.31 | 3,057.20 | 16,514.51 | 18.51% |
| 01/07/21 | 13,238.38 | 2,147.39 | 15,385.77 | 13.96% |
| 02/07/21 | 13,424.50 | 1,683.14 | 15,107.64 | 11.14% |
| 03/07/21 | 13,333.56 | 1,521.22 | 14,854.78 | 10.24% |
| 04/07/21 | 13,452.56 | 1,278.97 | 14,731.54 | 8.68% |
| 05/07/21 | 13,417.13 | 1,755.97 | 15,173.10 | 11.57% |
| 06/07/21 | 13,459.50 | 4,818.44 | 18,277.94 | 26.36% |
| 07/07/21 | 13,379.81 | 3,678.96 | 17,058.77 | 21.57% |
| 08/07/21 | 13,455.88 | 3,458.55 | 16,914.42 | 20.45% |
| 09/07/21 | 13,328.06 | 2,630.17 | 15,958.23 | 16.48% |
| 10/07/21 | 13,451.75 | 1,626.45 | 15,078.20 | 10.79% |
| 11/07/21 | 13,343.63 | 1,253.04 | 14,596.66 | 8.58% |
| 12/07/21 | 13,253.44 | 2,679.64 | 15,933.07 | 16.82% |
| 13/07/21 | 13,179.19 | 2,928.88 | 16,108.07 | 18.18% |
| 14/07/21 | 13,366.63 | 4,347.92 | 17,714.54 | 24.54% |
| 15/07/21 | 13,426.56 | 4,182.96 | 17,609.52 | 23.75% |
| 16/07/21 | 13,254.88 | 3,927.15 | 17,182.03 | 22.86% |
| 17/07/21 | 13,213.50 | 2,665.88 | 15,879.38 | 16.79% |
| 18/07/21 | 13,311.31 | 2,562.80 | 15,874.12 | 16.14% |
| 19/07/21 | 13,332.38 | 2,751.10 | 16,083.48 | 17.11% |
| 20/07/21 | 13,216.81 | 3,131.49 | 16,348.30 | 19.15% |
| 21/07/21 | 13,427.75 | 2,487.72 | 15,915.47 | 15.63% |
| 22/07/21 | 13,348.06 | 3,156.56 | 16,504.62 | 19.13% |
| 23/07/21 | 13,233.38 | 3,701.58 | 16,934.95 | 21.86% |
| 24/07/21 | 13,442.56 | 2,216.47 | 15,659.03 | 14.15% |
| 25/07/21 | 13,315.88 | 1,990.42 | 15,306.29 | 13.00% |
| 26/07/21 | 13,210.38 | 3,878.50 | 17,088.88 | 22.70% |
| 27/07/21 | 13,106.94 | 4,849.17 | 17,956.10 | 27.01% |
| Average | 16.15% |
@HomeAlone: A very over simplification to be honest.
Firstly, block fee is 2 gewi, not 2 ETH.
Secondly, at this point we don't know how the gas will be split between base fee and tip, however a portion of the fee (which equates to under 50% of total excluding tips / benefit) will be burned.
This graph that's provided does not accurately represent the post london fork splits, which will be different to do the reclassification from MV to Base Free + Tip.
Will be a wait and see situation, but personally I think my estimate is conservative.
A very over simplification to be honest.
Historical data showing transaction fee portion of the rewards (the very thing which the fork will effectively rework/burn) is an oversimplification?
Firstly, block fee is 2 gewi, not 2 ETH.
2 Gwei is only 0.000000002 ETH. That's worth like 1/1600th of a cent.
With roughly 6400 blocks per day that would make the total guaranteed daily rewards shared by all miners around 0.0000128 ETH or 4 cents. That's the most ridiculous things I've heard all day and I even stopped by Reddit earlier…
The Static Block Reward is 2 ETH, but you don't have to take my word for it. You can very easily verify this yourself. Just go to https://etherscan.io/blocks, pick any number of random blocks post Feb 2019 Constantinople Hard Fork and check the reward breakdowns. I guarantee you they will all show Static Block Reward as 2
Secondly, at this point we don't know how the gas will be split between base fee and tip, however a portion of the fee (which equates to under 50% of total excluding tips / benefit) will be burned.
I think the technical documentation and developer discussions on GitHub have covered it all pretty thoroughly, but I suppose if your primary source of information is CoinDesk articles then I can certainly understand why you don't understand.
This graph that's provided does not accurately represent the post london fork splits, which will be different to do the reclassification from MV to Base Free + Tip.
I did mention to "Keep in mind that transaction fee will be replaced by a base fee + tip with only the base fee being burned so the cut in profits may end up being even lower."
In other words it shows a worst case scenario of up to 16% cut.
Will be a wait and see situation, but personally I think my estimate is conservative.
The average block reward over the past 60 days was 2.49531537181068
50% cut would make it 1.24765768590534
25% cut would make it 1.87148652885801
It's literally impossible to have a block with a reward lower than the 2 ETH Static Block Reward.
Therefore your estimate is impossible.
@ln28909: Eth below 1k is a definite possibility, but I wouldn't count on it. The only problem is that Eth 2.0 is who knows how long away, once it's implemented at least it will reduce the mining demand of gpus (hopefully).
@bbqwer: True. Although don't you think it will reduce the demand even a little bit?
Currently Eth I assume makes up the majority of gpu mining across the POW blockchains, surely that disappearing would have some effect on the demand for mining.
@Dubious speculation: It will reduce demand for sure. but for how much, we dont know.
similar to BTC , and soon ETH. The next coin will pop up.
Crypto are getting more and more world known now, demand will keep going up.
And ETH PoW has been delayed since 2017, may be it can get delay again. We will soon find out Q42021 or Q12022
@Dubious speculation: Even if ETH2.0 is implemented, there are plenty of other cryptos to mine with GPUs apart from Ethereum.
@techlead: Alts apart from ETH are profitable now because they have a minuscule fraction of the miners on the network.
Once ETH 2.0 comes in, if a tiny minuscule fraction of the miners move from ETH to Raven or other, there won't be any profitability unless those coins see a massive, past Mars, price soar.
I personally don't see that happening. I believe ETH 2.0 is the end of GPU mining, but man was it good while it lasted.
@studentl0an: Crypto is here to stay, I don't think ETH2.0, when implemented is the end of GPU mining.
Time will tell, let's enjoy the ride in the meantime.
I've been in crypto since 2013, I've seen many naysayers slain in my time. Bitcoin will never be over $1000 USD. Bitcoin will never be over $10000USD. Bitcoin will never be $20000USD ever again. Bitcoin is a scam. Bitcoin will go to zero. Ethereum is a scam (I thought this when it ICO'd, that's why I didn't buy it until it went up way above its IPO price T_T regrets…) hahaha
@techlead: Yes it's here to stay, which is why ASIC mining and/or proof of stake is the future. I don't think you'll find many people arguing your case when they can see the writing on the wall with ETH removing GPU mining and not a single other coin could be remotely profitable if they have to share block rewards with all the miners moving from ETH to whatever.
GPU mining is old inefficient tech that has no place in the future of crypto in the same way lead acid batteries don't belong in electric cars for more efficient lithium tech (which will be phased out once solid state graphene batteries are commercially available).
@studentl0an: I think demand for GPU mining will be around for at least the next decade. Although ASICs are starting to pop up for Ethash (Ethereum's hashing protocol) and other protocols currently mined with GPUs, but the value proposition is not there yet.
This is why I didn't touch ASIC at all. I started as a GPU miner in 2013 and still am. ASICs have no salvage value, while GPUs do and in good times, their value even appreciates even after years of mining. Look at the RX 580 8GB, those cards are selling for $400-500 used, that's ludicrous!
There will certainly be changes to the demand in GPU mining after ETH2.0 is implemented, but its not as dire as your prediction, the transition away from GPU mining would be more gradual and during that time, it would still be quite profitable.
@techlead: What are you basing your opinion off? Do you really think a new coin will come along that will be as profitable as ethereum yet only be mineable on generic GPUs? I find something NQR about someone who said they have been mining bitcoin since 2013 yet created an account on this website just a few years ago. It's like you want to appear as an authority on GPU mining yet somehow didn't make 100x their money since 2013 if you had that many GPUs? If that's the case obviously you made some really massive mistakes not to be a multi millionaire by now who doesn't spend their time on bargain forums, and perhaps your crystal ball needs calibration.
It really seems like you're just hoping GPU mining will somehow survive (despite a widespread consensus saying otherwise and giving their reasons for it) rather than having a rational basis for it.
Your reasoning as it stands can be debunked with a single word 'bitcoin'. Not a single GPU miner left. It can not be profitable, even with salvage value from GPU mining if ASICs are 100x better at hashing. It's only ethereum that is keeping GPU mining alive and it has an expiry date of when the blockchain moves to proof of stake at the end of this year or start of the next.
@studentl0an: Its up to you whether you believe me or not. I'm definitely not an authority, I'm learning something new every day and this space is changing so fast, what worked a few years ago might no longer work anymore. So we all have to be nimble and constantly look at the changing landscape to adapt.
I did an AMA, so you can find out more about my personal experience in crypto, not just in mining. https://www.ozbargain.com.au/node/613895
I didn't 100x my money by the way. Its waaaay more than that. :)
@techlead: You are knowledgeable, which is why I can't understand why you think GPU mining will continue when everyone else I've spoken to with your level of knowledge strongly predicts the end of GPU mining soon for the reasons I have given. You just seem to have a hunch that goes against all market predictions.
Having said that I don't believe anyone who would have made the money you say you make would spend the amount of time on this website that you do. I think that's a bad way to live life if you have the means to spend it in infinite better ways.
@studentl0an: I like bargains lol, that's why I'm on this site, I also like to open people's eyes to the world of crypto.
Let's see how it plays out, my rigs are still mining away, not going to stop that anytime soon.
I'm not always right, but my strategy has worked out so far. Take what other people say with a grain of salt, including what I say, apply your own analysis to it and come to a conclusion on your own. Just like all those people who have been calling for $20k or even $10k USD Bitcoin in July, Twitter was full of them for 8 weeks, I analyzed their comments and reasoning and concluded I do not agree. I know friends who panic sold under $30k USD, some even sold all of their crypto at a loss and sworn to never buy it again, against my advice (but its fine, they are free to do whatever they want with their own crypto).
I'm not a contrarian for the sake of going against the popular view, its just my analysis do not agree. Just like people who claim China's mining ban is the end of crypto. I see it as a very positive development. Also, cheap Chinese cards will not flood the western second hand market because of costs to transport and important them. So those $1200 3080s, ain't going into Australian computers anytime soon.
I think GPU mining has some ways to go yet because heaps of other coins can be mined profitably. Most newbies use Nicehash, at its core all it does it mine the most profitable coin and exchange it to Bitcoin or Ethereum or whatever you choose periodically. Until profits using Nicehash goes to near $0, there is going to be demand for GPU mining. People using Nicehash probably don't even know what EIP1559 is and they don't need to. If profitability in mining ETH goes down, Nicehash will switch them to something else seamlessly.
Also keep in mind, if we get a continuation of the bull run, ie, Bitcoin to over $100k or $250k USD and ETH to $10k USD and its sustained for a few months, mining profitability will be through the roof and demand for non-LHR GPU cards will skyrocket.
@techlead: PoS will eventually replace PoW, but that's a long time away. ETH may disappear soon, but I agree that it's still a good idea to mine other coins even if not as profitable.
@badjazy: Oh, I see. Well when ETH2.0 is implemented, there will not be any Eth mining at all, it will become PoS. It doesn't mean GPU mining will disappear though.
There's a world of difference between ETH may disappear soon and ETH mining may disappear soon haha. :D
@thedean: If you are referring to the London hardfork implementing EIP1559, its not going to destroy mining profits for Ethereum lol.
@techlead: Not replying to you specifically, but this thread stream
https://github.com/ethereum/EIPs/pull/3675
end of ETH mining now has it's EIP underway
TBH I forget Raytracing exists half of the time even enabled in a small selection of games.
watch this get ozbargained