Estimated Hash Rates for the RTX 2080 and RTX 2080 Ti

[whitesites]

I have done a little comparison of the hash rates for current cards to what the next generation might do.
How fast will the Nvidia 2080 1180 Hash Crypto Currencies

Granted I am focusing more on ALGOs that are Core and clock intensive and not Memory ALGOs like ETH.

[1714982028]

1714982028

[1714982028]

1714982028

[MATHReX]

Site not reachable.
You can post your comparison here too if you want and that will be much appreciated.

[whitesites]

Sorry the firewall on my server is a little thick.
Here is the article

Even though the estimates are all over the place we an make some predictions based on performance of past models.

What determines hash rate?

There are three things that mainly determine hash rates.  Core Count, Clock Speed, and Memory Speed.  Different Algos require different things. Coins like Ethereum are Memory instensive, meaning that Memory Speed makes the biggest difference in its performance.  Hence the reason a GTX 1070 hashes almost as fast as a GTX 1080 Ti.  But most other coins depend more on Core Count, and Clock Speed.  For the purpose of this discussion we are going to focus on the later, since Ethereum has become a saturated coin with little to no profits for miners.  

Nvidia GPUs Compared

If we assume that core counts can be used to determine hashing rates for future cards we get the below chart

GPU	GTX 1060	GTX 1080	GTX 1080 Ti	RTX 2080	RTX 2080 Ti
Core Count	1280	2560	3584	2944	4352
Clock Speed	1708 Mhz	1733 Mhz	1582 Mhz	??	??
Memory and Type	6GB GDDR5	8GB GDDR5X	11GB GDDR5X	8GB GDDR6	11GB GDDR6
Hash Rates				15% More	21% More
X16R	5 Mh/s	10.5 Mh/s	15Mh/s	12.0 Mh/s	18.1 Mh/s
Lyra2z	1.1 Mh/s	2.2 Mh/s	3.0 Mh/s	2.5 Mh/s	3.6 Mh/s
Lyra2REv2	20.3 Mh/s	46.5 Mh/s	64 Mh/s	53.4 Mh/s	77 Mh/s
Phi2	2.3 Mh/s	4.3 Mh/s	6.0 Mh/s	4.9 Mh/s	7.2 Mh/s

This is going by data from Whattomine.com  

We are not including any improvements due to the GDDR6 Memory, or Clock speeds and or overclocks.  This also ignores energy expelled when mining.  The above figures are also on the low side, as recent Software miners have been known to be up to 30% faster than the numbers above.  Example The miner I use gets a consistent 20 Mh/s on X16R running on a GTX 1080 Ti.

Hash Rates for Ethereum

Considering that memory speed and latency has such a huge impact on hashing ETH, and we do not know the specs of the memory going into the new cards other than they are GDDR6, we have no way to calculate the speed increase.  Even though we can make comparisons to cards like the Titan V ( 12GB HMB2 ) which does ETH at 79 Mh/s, that is HMB2, and not GDDR6.  However if use Total bandwidth to make these calculcations with the 1080 Ti at 484 GB/s and the Titan V at 652 GB/s we start to get ratios that represent the increase in hash rate for ETH.  The 2080 Ti with 384 bit  GDDR6 has a theoretical speed of 864 GB/s which is 32% faster than HMB2, means we should be looking at ETH speeds around 104 Mh/s.  But we have to take into account the Titan V has 5120 Cores vs the 4352 Cores of the 2080 Ti.  Which is a 17 Increase. So lets take 17% off that hash rate to get us a card that should hash ETH at about 86 MH/s.  

However thanks to the Enlargementpill a 1080 Ti can hit speeds of 55 MH/s on ETH.  ( Which sells used for under $550 on ebay )
and the 1070 Ti can hit speeds of 31 MH/s on ETH ( which sells for under $350 on ebay )

If the 2080 Ti can hit 86 MH/s that is a 56% increase over the 1080 Ti, and a 177% in crease over the 1070 Ti.
By those figures For Ethereum Mining the 2080 Ti might actually be a better value than buying a used 1080 Ti off ebay.
However as we all know the ASIC miners are likely to start taking over ETH coins, so buying a 2080 Ti for ETH mining would be risky.

If I was the CEO of Nvidia I would be trying to bribe the developers at Ethereum to tweak their coin to stay ASIC resistent, and help keep sales going of their GPUs.
Its not like anyone in the mining community would be against such a move.

Its better to focus on other ALT coins.

Is the 2080 Ti worth the money for mining?
Probably not.  Remember you can score used GTX 1080 Tis on ebay for about $550.  While the RTX 2080 Ti will likely sell for about $750 when it hits shelves.  That is a 33% price increase.  Yet the estimated hash rates will likely only be about 20% increase over the previous generation.  To be on equal ground with the previous generation Nvidia would have to price their RTZ 2080 Ti at $665, and that price has to include all taxes and fees, Remember almost nobody is paying tax for cards on Ebay.  

The fact that Nvidia's CEO outright said that sales to miners will be almost Zero, indicates that they already know the hashrates of the new cards, and they are disappointing.  Nvidia has also announced the discontinued production of mining specific cards.  Which indicates that miners had no interest in buying up cards that had no resale value to gamers in the event that Crypto crashed.

The Good news
Many people have been liquidating their GPUs in anticipation of the 2080 and 2080 Ti.  For gamers this makes sense, but for miners this is pure speculation that the new cards would hash much faster then current gen.  Once its realized that the hashing increase of 2080 / 2080 Ti is not going to be that great, prices of used GTX 1080 Tis should increase.  Its the same reason people don't run out and build mining Rigs with Titan V cards.  The cost is too high and the ROI is not worth the premium paid for the hardware. 

The Bad news for Nvidia
Nvidia who for the past  year was in denial to investors of how big a portion of their business miners were.  Is about to see what happens to their stock price when their biggest customer (miners) suddenly stop buying their products.  On top of that most of us all bought our GPUs around January of this year, and they have a 3 year warranty on them.  So Nvidia gets to deal with Millions of RMAs that are bound to head their way over the next 2 years as Miners continue to burn up cards.

[cudapop]

The various discussions on the theoretical ETH hashrate for the new cards have generally been based on the memory bus width and memory data pin speed used on the different cards thus giving the memory bandwidth, then dividing by 8192 bytes (since each ETH hash will sample 8K bytes of the DAG).

So, in theory, the zero-latency hashrates would be computed as:

bus width (in bits) = memory capacity (in GB) * 32
data pin speed = 14 Gbp/s (if assumed same memory as used in Quadro RTX series)
bandwidth (in GB/s) = bus width / 8 * data pin speed
zero-latency hashrate (in MH/s) = bandwidth / 8192 * 1000

For example, using the leaked specs from AdoredTV's video on YT:

2080 Ti with 11 GB GDDR6 at 14 Gbp/s - memory bus width: 352 bit, memory bandwidth: 616 GB/s, hashrate: ~75 MH/s
2080 with 8 GB at 14 Gbp/s - bus width: 256 bit, bandwidth: 448 GB/s, hashrate: ~55 MH/s
2070 with 7 GB at 14 Gbp/s - bus width: 224 bit, bandwidth: 392 GB/s, hashrate: ~48 MH/s
2060 with 5 GB at 14 Gbp/s - bus width: 160 bit, bandwidth: 280 GB/s, hashrate: ~34 MH/s

(If you have different specs for the cards for memory capacity and memory speed, eg. 2070 with 8GB at 16 GB/s, just use them in the formulas and recompute)

This is all assuming the cards are not "gimped" out-of-the-box with "poor" latency (trrd, tfaw) settings for mining use (ie. similar to what was the case with out-of-the-box 1080s and 1080TIs with their GDDR5X).

[gameboy366]

Who cares about hashrate for eth now. Just get an E3 miner if you want to mine ethash coins. People bought it for $800 for 195 h/s.

[soothaa]

From what I saw the 2080ti was taking a huge core speed hit, something like 1500MHz range. If that's true I don't think core-bound mining will see any speedup at all at this rate.

[ThePunisher49X]

Buriedone a youtuber has a breakdown of each 20 series card for eth hashrates at the end of his stream of the nvidia promo event of these cards.

[abstractHaze]

so. to be honest. no one really knows so these cards are currently looking way overpriced?

[nsummy]

so. to be honest. no one really knows so these cards are currently looking way overpriced?

For mining in the current economy yes, especially considering the founders editions are more expensive than msrp.  If there is a resurgence in altcoins though these will be flying off the shelf. 

[nsummy]

Sorry the firewall on my server is a little thick.
Here is the article

Even though the estimates are all over the place we an make some predictions based on performance of past models.

What determines hash rate?

There are three things that mainly determine hash rates.  Core Count, Clock Speed, and Memory Speed.  Different Algos require different things. Coins like Ethereum are Memory instensive, meaning that Memory Speed makes the biggest difference in its performance.  Hence the reason a GTX 1070 hashes almost as fast as a GTX 1080 Ti.  But most other coins depend more on Core Count, and Clock Speed.  For the purpose of this discussion we are going to focus on the later, since Ethereum has become a saturated coin with little to no profits for miners.  

Nvidia GPUs Compared

If we assume that core counts can be used to determine hashing rates for future cards we get the below chart

GPU	GTX 1060	GTX 1080	GTX 1080 Ti	RTX 2080	RTX 2080 Ti
Core Count	1280	2560	3584	2944	4352
Clock Speed	1708 Mhz	1733 Mhz	1582 Mhz	??	??
Memory and Type	6GB GDDR5	8GB GDDR5X	11GB GDDR5X	8GB GDDR6	11GB GDDR6
Hash Rates				15% More	21% More
X16R	5 Mh/s	10.5 Mh/s	15Mh/s	12.0 Mh/s	18.1 Mh/s
Lyra2z	1.1 Mh/s	2.2 Mh/s	3.0 Mh/s	2.5 Mh/s	3.6 Mh/s
Lyra2REv2	20.3 Mh/s	46.5 Mh/s	64 Mh/s	53.4 Mh/s	77 Mh/s
Phi2	2.3 Mh/s	4.3 Mh/s	6.0 Mh/s	4.9 Mh/s	7.2 Mh/s

This is going by data from Whattomine.com  

We are not including any improvements due to the GDDR6 Memory, or Clock speeds and or overclocks.  This also ignores energy expelled when mining.  The above figures are also on the low side, as recent Software miners have been known to be up to 30% faster than the numbers above.  Example The miner I use gets a consistent 20 Mh/s on X16R running on a GTX 1080 Ti.

Hash Rates for Ethereum

Considering that memory speed and latency has such a huge impact on hashing ETH, and we do not know the specs of the memory going into the new cards other than they are GDDR6, we have no way to calculate the speed increase.  Even though we can make comparisons to cards like the Titan V ( 12GB HMB2 ) which does ETH at 79 Mh/s, that is HMB2, and not GDDR6.  However if use Total bandwidth to make these calculcations with the 1080 Ti at 484 GB/s and the Titan V at 652 GB/s we start to get ratios that represent the increase in hash rate for ETH.  The 2080 Ti with 384 bit  GDDR6 has a theoretical speed of 864 GB/s which is 32% faster than HMB2, means we should be looking at ETH speeds around 104 Mh/s.  But we have to take into account the Titan V has 5120 Cores vs the 4352 Cores of the 2080 Ti.  Which is a 17 Increase. So lets take 17% off that hash rate to get us a card that should hash ETH at about 86 MH/s.  

However thanks to the Enlargementpill a 1080 Ti can hit speeds of 55 MH/s on ETH.  ( Which sells used for under $550 on ebay )
and the 1070 Ti can hit speeds of 31 MH/s on ETH ( which sells for under $350 on ebay )

If the 2080 Ti can hit 86 MH/s that is a 56% increase over the 1080 Ti, and a 177% in crease over the 1070 Ti.
By those figures For Ethereum Mining the 2080 Ti might actually be a better value than buying a used 1080 Ti off ebay.
However as we all know the ASIC miners are likely to start taking over ETH coins, so buying a 2080 Ti for ETH mining would be risky.

If I was the CEO of Nvidia I would be trying to bribe the developers at Ethereum to tweak their coin to stay ASIC resistent, and help keep sales going of their GPUs.
Its not like anyone in the mining community would be against such a move.

Its better to focus on other ALT coins.

Is the 2080 Ti worth the money for mining?
Probably not.  Remember you can score used GTX 1080 Tis on ebay for about $550.  While the RTX 2080 Ti will likely sell for about $750 when it hits shelves.  That is a 33% price increase.  Yet the estimated hash rates will likely only be about 20% increase over the previous generation.  To be on equal ground with the previous generation Nvidia would have to price their RTZ 2080 Ti at $665, and that price has to include all taxes and fees, Remember almost nobody is paying tax for cards on Ebay.  

The fact that Nvidia's CEO outright said that sales to miners will be almost Zero, indicates that they already know the hashrates of the new cards, and they are disappointing.  Nvidia has also announced the discontinued production of mining specific cards.  Which indicates that miners had no interest in buying up cards that had no resale value to gamers in the event that Crypto crashed.

The Good news
Many people have been liquidating their GPUs in anticipation of the 2080 and 2080 Ti.  For gamers this makes sense, but for miners this is pure speculation that the new cards would hash much faster then current gen.  Once its realized that the hashing increase of 2080 / 2080 Ti is not going to be that great, prices of used GTX 1080 Tis should increase.  Its the same reason people don't run out and build mining Rigs with Titan V cards.  The cost is too high and the ROI is not worth the premium paid for the hardware. 

The Bad news for Nvidia
Nvidia who for the past  year was in denial to investors of how big a portion of their business miners were.  Is about to see what happens to their stock price when their biggest customer (miners) suddenly stop buying their products.  On top of that most of us all bought our GPUs around January of this year, and they have a 3 year warranty on them.  So Nvidia gets to deal with Millions of RMAs that are bound to head their way over the next 2 years as Miners continue to burn up cards.

A.  Ethereum is already pretty much ASIC resistant.  The E3 mines at 190 MH/S at 760W.  Not surprising that a machine that can mine only 1 algorithm marginally beats a 2 year old gaming graphics card that is half the price.

B.  The CEO's comments on mining sales have nothing to do with hashrates of the new cards.  Even if the 2080ti tripled the performance of of the 1080ti I doubt many miners would be buying.  His point was that with altcoin prices and the availability of cheap GPUs on eBay, it doesn't make sense currently for miners to purchase new cards.

C.  Miners aren't Nvidia's largest customers by a long shot.

[whitesites]

A.  Ethereum is already pretty much ASIC resistant.  The E3 mines at 190 MH/S at 760W.  Not surprising that a machine that can mine only 1 algorithm marginally beats a 2 year old gaming graphics card that is half the price.

B.  The CEO's comments on mining sales have nothing to do with hashrates of the new cards.  Even if the 2080ti tripled the performance of of the 1080ti I doubt many miners would be buying.  His point was that with altcoin prices and the availability of cheap GPUs on eBay, it doesn't make sense currently for miners to purchase new cards.

C.  Miners aren't Nvidia's largest customers by a long shot.

If the 2080Ti trippled the performance of the 1080ti ( in mining that is ), I would be liquidating all my GPUs and prepare to upgrade to 2080tis, as used 1080ti still sell for $550 on ebay, and the 2080ti will be priced around $1100.

I guess we will see how many GPUs Nvidia sells when Miners are no longer interested.

[shater]

A.  Ethereum is already pretty much ASIC resistant.  The E3 mines at 190 MH/S at 760W.  Not surprising that a machine that can mine only 1 algorithm marginally beats a 2 year old gaming graphics card that is half the price.

B.  The CEO's comments on mining sales have nothing to do with hashrates of the new cards.  Even if the 2080ti tripled the performance of of the 1080ti I doubt many miners would be buying.  His point was that with altcoin prices and the availability of cheap GPUs on eBay, it doesn't make sense currently for miners to purchase new cards.

C.  Miners aren't Nvidia's largest customers by a long shot.

If the 2080Ti trippled the performance of the 1080ti ( in mining that is ), I would be liquidating all my GPUs and prepare to upgrade to 2080tis, as used 1080ti still sell for $550 on ebay, and the 2080ti will be priced around $1100.

I guess we will see how many GPUs Nvidia sells when Miners are no longer interested.

If 2080Ti hypothetically mines at 150MH/s for around $1,100, then I do think is a game-changer.

[cudapop]

150 MH/s on ETH would require a memory bandwidth of around ~1.2 TB/s (terabytes per sec), the announced memory specs for the RTX 2080Ti of 11 GB at 14 Gbps would mean a bandwidth of around 616 GB/s for a single 2080Ti. So it looks more likely that the theoretical, zero-latency hashrate would be around ~75 MH/s.

Maybe in another algo and not Ethash, assuming the algo can make use of the paralllel issue of INT32 and FP32 instructions in Turing's CUDA cores (cf. to what was introduced with Volta), as well as an algo that needs to do matrix multiplations and thus use the Tensor cores. For example: Bytom's Tensority algo would benefit from the new architecture since it can use 3 features: INT8 DP4A instructions (assuming Turing still retains this from Pascal), WMMA (warp matrix multiply accumulate) using Tensor cores, and parallel issue INT32 and FP32 (since Tensority uses 8-bit values as input in it's matrices and thus can use the FP32 to multiply 8-bit values in parallel with INT8 DP4A on the INT32 cores). Not sure if it would triple it vs. a 1080Ti though, need to test it first.

The Groestl hash function also has a matrix multiply function in it so maybe it might benefit from the Tensor cores. Not sure though since I've only really tested CUDA kernels for Ethash and Tensority.

[Nexigen]

If you compare the number of CUDA cores than you will see only 20% of the difference. I don’t think that memory in these cards is much faster than in GTX1000 series. That’s why I think that RTX 2070 will give 36-40 mh for Ether mining.

[Longsnowsm]

Following... Once you guys get your hands on some cards please post your results here.  Would love to see the comparisons.  That is a lot of money per gpu so it will have to be killing the numbers to justify the cost.

[cudapop]

It's definitely possible that the new RTX cards could have poor memory latency (trrd, tfaw) settings for ethash mining use. After all, we already have existing examples of this: eg. the GTX 1080 with it's 320 GB/s memory bandwidth has a theoretical zero-latency hashrate of around ~39 MH/s, but without something like the ethlargement-pill it won't get near this hashrate.

So the idea is that the new RTX cards, with their much higher memory bandwidths, would allow for the higher ethash hashrates (as computed from the bandwidths), either in stock form assuming they are not "crippled with non-mining-friendly" latency settings, or with some equivalent of the eth-pill.

As for the CUDA core count and it's effect on ethash rates, if you download the source of ethminer and modify the "dagger_shuffled.cu" file so that instead of loading DAG entries from global memory it loads it from a structure in shared memory (ie. just to test what the cores are capable of hashing at assuming memory latencies which are comparable to the L1 cache of the device), you'll find that even something like a GTX 1060, which would normally hash at 20+ MH/s , will now be capable of 50+ MH/s with the latencies of the shared memory instead of the much slower global memory on the device. So even a device with just the core count of a GTX 1060 can achieve a much higher hashrate if it had memory with much lower latencies.

So it's possible that even with just the 20% increase in core count in the new RTX cards, _if_ it has mining-friendly latency settings (or we get an equivalent of an eth-pill), the new cards might be capable of hashing as fast as their theoretical memory bandwidths would allow.

[gerask0]

linustechtips gets this card one of the first, likely.
but he rarely test card in mining
I will try to ask him for some results

[Metroid]

ethlargement-pill it won't get near this hashrate.

devs of that program will probably sell a functionality to do the same as gtx 1080ti for rtx 2080ti owners, i dont think they will release it publicly like they did for the gtx 1080 owners.

[lewaeve]

Following,hope someone can get the first test card.....

[ummasek]

In my opinion, hashrate wil not be as simple to say.
1. More cuda
2. More memory bandwitch
3. faster clockrate
For me it`s seems to be better hash for less power/hashrate and more space in your rig.
IS it good for all of us?
We`ll see when cards comes for us
We just need them fast and lots of them.

[rockmoney]

Was hoping to see a bit more solidity in the speculated stats for the new RTX 2080, especially since the Founders Edition is currently available on NVIDIA website for $799 as a pre-order, and all of the specs have been released. Will keep an eye out for more info, but I imagine these cards will be harder to find at first as time goes on, mainly because the two other editions are already sold out on NVIDIA website.

I am also seeing 11GB editions of this card on second and third party websites. I assume this 11GB version would be the one to get for ETH mining, as it will certainly remain relevant for much longer than 8GB versions (again, for ETH mining).

Anyway, thanks for all of the info so far, and please post any new info as it becomes available! 

[infofront]

Is it technically possible to do any mining on the new ray tracing cores?

[Wananavu99]

Just wait for AMD Navi 7nm, that should force Nvidia to lower prices.  For a possible 15-20% hashrate increase on current prices, dependent on electrical costs, ROI could be minimum or even negative.  It's bear market right now and some coins are honestly thinking of moving away from PoW. 

[KevinMiles]

The Turing new cores are ASICS. Imagine someone write a miner to use Tensor and RTX cores. As a result it will be possible to mine while gaming with the Cuda cores. I bet some sophisticated miner will pop up soon in order to utilize the nvidia new marvel.

[zenstrive]

If Volta is used as reference, cryptonight v7 hashrate is still lower than Vega 56

[revenant2017]

The Turing new cores are ASICS. Imagine someone write a miner to use Tensor and RTX cores. As a result it will be possible to mine while gaming with the Cuda cores. I bet some sophisticated miner will pop up soon in order to utilize the nvidia new marvel.

This is also what i have in mind with Tensor Cores. If someone could create and optimized Tensor cores for mining and also the speed of GDDR6, the possibility is the mining hashrate can be higher than past architectures. The sweet spot is at RTX2080 base on it's price/performance.

[giagge]

I think the RTX and tensor core is much faster to Xilinx fpga .

[vuli1]

no one got the new gpu yet, to give us the results.

[markiz73]

In the near future, the test results should be released, but they will not reflect all the capabilities of the new generation of video cards.
After a few months, the developers optimize the software for the new GPU architecture and memory and then we will get a real result.

Here are a few tests of  RTX 2080 Ti & RTX 2080
https://videocardz.com/77983/nvidia-geforce-rtx-2080-ti-and-rtx-2080-official-performance-unveiled

[R0land]

The ethash speed is limited by the bandwith.
With a GDDR6 384-bit memory bus you have ~768 GB/s  bandwidth --> ~90 Mh/s eth.
With a GDDR6 256-bit memory bus you have ~512 GB/s  bandwidth --> ~60 Mh/s eth.

[cudapop]

The memory bandwidths you listed above are for 16 Gbps GDDR6. The RTX cards ship with 14 Gbps speed memory, bandwidths are: 2080Ti (352 bit bus) = 616 GB/s, 2080 (256 bit bus) = 448 GB/s.

[nitrobg]

Those 14Gbps chips can probably be overclocked to at least 15Gbps - just like most GDDR5 can reach 9Gbps. Perhaps it is the lowest 100% stable speed for all chips.

[cudapop]

Yes that's true, but when using the memory bandwidth values to estimate Ethash hashrates, the value obtained is actually a theoretical zero-latency hashrate. But since Ethash takes each DAG sample from a pseudo-random location, latency actually has a substantial effect. So one conservatively assumes that the overclocking is used largely to make up for the actual latencies as compared to the theoretical zero-latency hashrate.

So for example, a 192-bit bus GTX 1060 with 8 Gbps memory has a theoretical zero-latency Ethash hashrate of around ~23.4 MH/s. An overclock to 9 Gbps would raise this to a theoretical zero-latency rate of ~26.4 MH/s. But due to the actual latencies involved, what we see in reality is hashrates closer to the ~23 MH/s value.

[RivAngE]

I think the RTX and tensor core is much faster to Xilinx fpga .

Let me try calculate that...

For starters we ignore the RTX cores, these are doing a very specific kind of job and I don't expect them to have any use in mining.
 - The Nvidia's Tensor cores for a 2080ti have an FP16 (aka half precision) computation performance of 110 TFLOPs!
 - Xilinx most powerful FPGA can offer 10,948 GFLOPs, or almost 11 TFLOPs (source: https://www.xilinx.com/products/technology/dsp.html#solution)

Xilinx doesn't mention which model resulted in this performance, only that it's from the UltraScale+ family. The fastest available FPGA for mining is this one https://store.mineority.io/sqrl/cvp13/ which costs $6,370 before tax and if you import it in Europe.... I don't want to think about it. This FPGA miner uses the Virtex UltraScale+ VU13P.

Nvidia's Tensor cores can also offer mixed precision computation which, to be honest, I have no idea what it is, if Xilinx offers this or if it matters for mining!

So as far as pure computation performance per $ goes, the RTX 2080ti is putting everything else miles away... and people call the RTX series overpriced
I wish however someone could go into more detail about how important the above numbers are and which algorithms would benefit more from this because I'm sure TFLOPs are NOT the only factor.


For comparison, 1080ti offers 13 TFLOPs FP32 (aka Single Precision), shader GPU only. I don't know if the 1080ti could compute FP16 but if it could that'd be 26 TFLOPs (double the FP32).
The 2080ti shader GPU offers 16 TFLOPs FP32 on top of the Tensor core which I mentioned above.

PS: That post took me more than half an hour to gather all these numbers from valid sources! omg

[cudapop]

Note that the 110 FP16 tflops of performance of the tensor cores is in one specific operation only: 4x4 matrix multiplication and accumulate. That's all tensor cores can do. It's essentially an ASIC which is designed to take two 4x4 matrices of FP16 values and multiply them, accumulating the result into a 4x4 matrix of FP32 values. You can't use a tensor core for anything else except to do matrix multiply and accumulate.

That's why it's rated at such a high tflops number: because it's hardware has been designed to do only matrix multiply and accumulate, it has no other functional use beyond that, you can't reprogram a tensor core to perform any other operation. Think of a tensor core like an S9 ASIC, but instead of doing sha256 all it does is 4x4 matrix multiply and accumulate.

On the other hand, Xilinx logic cells can be reconfigured to perform different operations, it's completely flexible, hence VHDL/Verilog development. In fact, using something like Xilinx's SDAccel you can write a C++/OpenCL program and have it built into a bitstream to run on a Xilinx FPGA.

[RivAngE]

Note that the 110 FP16 tflops of performance of the tensor cores is in one specific operation only: 4x4 matrix multiplication and accumulate. That's all tensor cores can do. It's essentially an ASIC which is designed to take two 4x4 matrices of FP16 values and multiply them, accumulating the result into a 4x4 matrix of FP32 values. You can't use a tensor core for anything else except to do matrix multiply and accumulate.

That's why it's rated at such a high tflops number: because it's hardware has been designed to do only matrix multiply and accumulate, it has no other functional use beyond that, you can't reprogram a tensor core to perform any other operation. Think of a tensor core like an S9 ASIC, but instead of doing sha256 all it does is 4x4 matrix multiply and accumulate.

On the other hand, Xilinx logic cells can be reconfigured to perform different operations, it's completely flexible, hence VHDL/Verilog development. In fact, using something like Xilinx's SDAccel you can write a C++/OpenCL program and have it built into a bitstream to run on a Xilinx FPGA.

Oh... I see! Good explanation!
Then I guess that a specific algo has to build which would take advantage of this operation. A software developer who creates miners probably wouldn't be able to take advantage of this core for the current algos.

[cudapop]

There are two current algos I am aware of which use matrix multiply: Tensority and Groestl.

I posted some details on them wrt Turing's new features in this post: https://bitcointalk.org/index.php?topic=4948083.msg44769341#msg44769341

[goxed]

Here's the sauce
https://www.computerbase.de/2018-09/nvidia-geforce-rtx-2080-ti-test/6/#abschnitt_ethereummining_ist_nicht_ueberragend

[Meteorite777]

Here's the sauce
https://www.computerbase.de/2018-09/nvidia-geforce-rtx-2080-ti-test/6/#abschnitt_ethereummining_ist_nicht_ueberragend

If that's true; and RTX 2080Ti only hashes at 50 Mh or less that would make it just at fast as a GTX 1080Ti with the ETHLargement Pill, correct? That would make it a massive flop for mining ETH atleast.

[revenant2017]

Here's the sauce
https://www.computerbase.de/2018-09/nvidia-geforce-rtx-2080-ti-test/6/#abschnitt_ethereummining_ist_nicht_ueberragend

If that's true; and RTX 2080Ti only hashes at 50 Mh or less that would make it just at fast as a GTX 1080Ti with the ETHLargement Pill, correct? That would make it a massive flop for mining ETH atleast.

If they developed an EthLargement pill for the newer Turing architecture, we might speculate atleast a 50% increase in hashrate at max or around 84mh/s.

[Race-Coin]

Would it be worth to upgrade from 1080ti?

[abstractHaze]

Would it be worth to upgrade from 1080ti?

sure.

[Meteorite777]

If they developed an EthLargement pill for the newer Turing architecture, we might speculate atleast a 50% increase in hashrate at max or around 84mh/s.

The 1080Ti and EthLargement Pill works by editing the memory timings of GDDR5X memory specifically though; and 2080Ti uses GDDR6 memory so I'm not sure if it would apply in the same way. By that same argument they could make an Ethlargement Pill for 1060/1070 and straight GDDR5 Memory too but they haven't. I'm not well-versed enough in the implentation to argue either for or against but yes you could be right.

[RivAngE]

If they developed an EthLargement pill for the newer Turing architecture, we might speculate atleast a 50% increase in hashrate at max or around 84mh/s.

The 1080Ti and EthLargement Pill works by editing the memory timings of GDDR5X memory specifically though; and 2080Ti uses GDDR6 memory so I'm not sure if it would apply in the same way. By that same argument they could make an Ethlargement Pill for 1060/1070 and straight GDDR5 Memory too but they haven't. I'm not well-versed enough in the implentation to argue either for or against but yes you could be right.

But they've made it for "older cards" too (3:00 of interview: https://youtu.be/ZLTRYp_kCYg?t=180), it's a private tool though and they sell them to companies.
Now if they'll release a 2080ti tool for free... I don't know... but they could if they wanted to.

[swogerino]

The data are not that impressive considering the price of a Rtx 2080 ti which is over 1000 dollars and the mining rewards from Gpu mining computers right now it is not worth to spend that amount of money for a card who only mines at less than 50 mhs , it may be the best card mining ethereum yet but not worth to spend over 1000 dollars for a single card.

If you want to sell it after some time , still it is a fifty-fifty successful situation in my opinion.

[szczur333]

after crypto prices falling

2x RX570 card can be bought for about 1600zł (PLN) here (its about 435$)

hashes about 55-60mhash easy

used even 20% cheaper

even with pill - miners will not buy 2xxx series for 1000$ O_O

[netto7]

Would it be worth to upgrade from 1080ti?

for gamer it sure

for miner i dont think so

[jazz1984]

Would it be worth to upgrade from 1080ti?

for gamer it sure

for miner i dont think so

I think it's too early to draw conclusions about the effectiveness of new video cards. To begin with, we need to know the hasrate and power consumption on different algorithms, and then we can make a desition.

[Marvell2]

Would it be worth to upgrade from 1080ti?

for gamer it sure

for miner i dont think so

I think it's too early to draw conclusions about the effectiveness of new video cards. To begin with, we need to know the hasrate and power consumption on different algorithms, and then we can make a desition.

the 2080 seems the same speed as a 1080ti with 24% less power consumption
that might be worth the $150 premium

[RivAngE]

Who is mining Ethereum anyway? (obviously still many people... ) I prefer the x16r and HEX hashing algorithms. I've found also Renesis which is using their own algo and have promised to have forks ready to implement the moment an ASIC or FPGA has been found on their network, but they're still a small team unlike the teams that support x16r and HEX.

Personally I've lost my interest in the long-term sustainability of Ethereum because I think its progression is lagging and ARK will soon surpass Ethereum's technology and usage... and also since mining and instant-selling ETH isn't profitable either, I just stay away from it altogether.

[agapes]

ok finaly i got my RTX 2080 today
i put on nemos miner start benchmarking, i will put the benchmark chart as soon as benchmarking done
but for now seems  look 1080 Ti maybe a bit more on some algorithms but on some of them is less thars shame

https://imgur.com/a/Euzycaa

this is half of the benchmark
still need more benchmark ill put about 45min

[agapes]

ok finaly benchmark done!
this is so far all algorythm and hashrate for RTX 2080   (Card is in normal mode not overclocked),

that's really disappointing
$1 a day max
also, no any equihash(zcash) miner supported yet


https://imgur.com/a/dCbI0wn

if you have any question or need hashrate for any algorithm let me know to i test it

but again dont buy this card wasting your money

[RivAngE]

I'm sorry for all of you guys who bought the non-ti version...
It's a 1080ti with two extra cores for AI factions and Ray Tracing, from which cores we might or more likely might not see a mining usage.

Nvidia should have named their products differently,
2080ti should be named 2080
2080 should be named 2070
2070 should be named 2060

If in a year they release a 2090 then I suppose I'm right.

(Btw I've preordered the 2080ti FTW3 and I'm not sad at all! I want it mainly for gaming and if it can mine at least decently, that's a bonus!)

[revenant2017]

ok finaly benchmark done!
this is so far all algorythm and hashrate for RTX 2080   (Card is in normal mode not overclocked),

that's really disappointing
$1 a day max
also, no any equihash(zcash) miner supported yet


https://imgur.com/a/dCbI0wn

if you have any question or need hashrate for any algorithm let me know to i test it

but again dont buy this card wasting your money

Quite disappointing though because the fastest possible ROI for 1 RTX 2080 is 600+ days or almost 2 years. At that time, the card is already near it's end of lifespan.

[grendel25]

So far, what I'm seeing is that the 2080 Ti is and will continue to be an improvement while people still make a value argument regarding what is better: Pay less for 10-series or pay 2x as much for 20-series.

The problem with that logic flow is that it makes valuation fallacies.  Ultimately, people should pay what they can afford but just because something costs 2x more doesn't mean that GPU farms won't acquire 20-series for the meager percentage gains because the gains can and most likely will outpace upfront costs in the long run as well as preventing losses when new opportunities present themselves.

So if I were to replace a $500 1080Ti with a $1K 20-Series card.  I paid $500 more at one time on one day.  From here on out, I gain some arbitrary percentage more in performance as well as power efficiency which matters more or less to different people paying different amounts for energy.  Unless you are on some sort of deadline, it doesn't matter if the 10-series or 20-series ROIs first.  Once they ROI, the 20-series will forever hold the profitability high-ground.  On top of all that, it is just as likely that there could be some unique development that only the 20-series could benefit from.  Some RT specific algo perhaps. 

Now, I'm not going to run out and instantly replace 10-series rigs.  However, I am watching the 20-series market closely as I expect any rational person would.

[RivAngE]

Interesting overclocking video here: https://youtu.be/FpmDa5VetME
've preordered the EVGA 2080ti FTW3 Ultra and I seem to have chose the right manufacturer!

[navarthelol]

So far, what I'm seeing is that the 2080 Ti is and will continue to be an improvement while people still make a value argument regarding what is better: Pay less for 10-series or pay 2x as much for 20-series.

The problem with that logic flow is that it makes valuation fallacies.  Ultimately, people should pay what they can afford but just because something costs 2x more doesn't mean that GPU farms won't acquire 20-series for the meager percentage gains because the gains can and most likely will outpace upfront costs in the long run as well as preventing losses when new opportunities present themselves.

So if I were to replace a $500 1080Ti with a $1K 20-Series card.  I paid $500 more at one time on one day.  From here on out, I gain some arbitrary percentage more in performance as well as power efficiency which matters more or less to different people paying different amounts for energy.  Unless you are on some sort of deadline, it doesn't matter if the 10-series or 20-series ROIs first.  Once they ROI, the 20-series will forever hold the profitability high-ground.  On top of all that, it is just as likely that there could be some unique development that only the 20-series could benefit from.  Some RT specific algo perhaps.  

Now, I'm not going to run out and instantly replace 10-series rigs.  However, I am watching the 20-series market closely as I expect any rational person would.

You forgot to say *IF* the 20-series ROIs, then it will hold the high ground. Today the yields are pretty terrible, but there are other innovations round the corner that could make GPUs obsolete; the acorns from squirrel labs; the FPGAs quietly being bulk produced and preorders placed in the background; the fact that POW has a questionable future e.g. Ethereum. These are all imminent dangers affecting profitability within the next 12 months, Acorns are meant to be shipping <1 month.

So I think arguing that paying 100% more for a 10% increase on a card that *might* do ROI in 10 years is a valuation fallacy. But anyway, if you have a fully operational set of rigs  I could understand purchasing some additional 2080TIs and expanding that operation, but I can't understand your logic for selling all your existing 1080 TI and replacing with 2080 TI. Maybe you entered the game at the right time and those 1080TI have already paid themselves off so the loss wouldn't be "as bad", but I still think it's pretty nonsensical to be buying hardware that you know has a bad cost:hash ratio- a GPU is never as profitable as the first day you turn it on, so economically speaking in a best case scenario, this time next year your 1080TI and 2080TI will have the same profitability as today.

Just my 2 cents as I have often seen in crypto land that when someone makes a bad investment, they often try to get others to share in that.

[TasosRiv]

On top of what navarthelol said, it also feels like Nvidia had PLANNED to make the 20 series cards bad for mining.

They added a bunch of expensive hardware which is probably not going to be helpful for mining except if a specific algo for Tensor cores is produced, but I doubt it since there are professional Volta cards around with much more power on them, it wouldn't make sense for a dev team to risk the stability of their coin like that.

Nvidia is focused on gaming and since AMD is miles behind Nvidia in gaming, only they might produce a GPU which is meaningful for mining.
I've bought a 2080ti but I'll use it for both gaming and mining, I can't see a reason for anyone to build a farm with these cards.

[john2231]

On top of what navarthelol said, it also feels like Nvidia had PLANNED to make the 20 series cards bad for mining.

They added a bunch of expensive hardware which is probably not going to be helpful for mining except if a specific algo for Tensor cores is produced, but I doubt it since there are professional Volta cards around with much more power on them, it wouldn't make sense for a dev team to risk the stability of their coin like that.

Nvidia is focused on gaming and since AMD is miles behind Nvidia in gaming, only they might produce a GPU which is meaningful for mining.
I've bought a 2080ti but I'll use it for both gaming and mining, I can't see a reason for anyone to build a farm with these cards.

I think there are some developers that they can develop a software to  improve the mining speed like gtx 1080 ti they made etherpill which improves the hashrate 70% so even Nvidia has plan to make a cards that bad on mining POW coins OhMyGodCompany can help to develop a software to improve hashrate.

[TasosRiv]

On top of what navarthelol said, it also feels like Nvidia had PLANNED to make the 20 series cards bad for mining.

They added a bunch of expensive hardware which is probably not going to be helpful for mining except if a specific algo for Tensor cores is produced, but I doubt it since there are professional Volta cards around with much more power on them, it wouldn't make sense for a dev team to risk the stability of their coin like that.

Nvidia is focused on gaming and since AMD is miles behind Nvidia in gaming, only they might produce a GPU which is meaningful for mining.
I've bought a 2080ti but I'll use it for both gaming and mining, I can't see a reason for anyone to build a farm with these cards.

I think there are some developers that they can develop a software to  improve the mining speed like gtx 1080 ti they made etherpill which improves the hashrate 70% so even Nvidia has plan to make a cards that bad on mining POW coins OhMyGodCompany can help to develop a software to improve hashrate.

I think this was an exceptional case because what the ETHpill did was to change the way some data was passed through the memory, the GDDR5X memory had some special needs as it seemed!
Judging by the hashrate difference of a 1080ti with ETHpill and a 2080ti's, I think the GDDR6 is scaling fine and there isn't much room for improvement.

My only hope for 20-series cards to become important for mining is for a a new algo which combines kernels which some are memory-intensive, some GPU-intensive, some require matrix operations (AI specific operations, taking advantage of Tensor cores) and if a Ray Tracing core would be detected, then use these to boost the hashrate somehow.

I'm not a developer so I don't know how accurate or possible are the things I've mentioned!

[TheLifeOfAMiner]

I have both cards, the RTX 2080 and the RTX 2080 TI.  Check out this page where I shared my findings so far.

https://bitcointalk.org/index.php?topic=5042514.0

Happy Mining!

[agapes]

ok i got my RTX 2080 Ti today too
and this is picture of bencharking for both 2080 and 2080Ti

https://imgur.com/a/1loLc2i

im super sad that waste my money on these 2 cards

[TasosRiv]

ok i got my RTX 2080 Ti today too
and this is picture of bencharking for both 2080 and 2080Ti

https://imgur.com/a/1loLc2i

im super sad that waste my money on these 2 cards

Good one but many algos are ASIC mined, so they bring no profit. I'd like to see how it'd hash on HEX algo (XDNA coin)

[agapes]

ok i tested both card with new version Z-Enemy 1.22

this is result
https://imgur.com/a/0NaN4KJ

[TasosRiv]

ok i tested both card with new version Z-Enemy 1.22

this is result
https://imgur.com/a/0NaN4KJ

Hm... thanks for sharing but this is very strange. My 1080ti hashes for 20-21 MH/s. I run the miner at 24 Intensity with overclocks of +300 memory and + 90 core (about 1770 MHz I think?) at 80% power.
I obviously don't know the overclock limits of a 2080ti but maybe the default intensity is bottle-necking you.

[Ajes]

ok i tested both card with new version Z-Enemy 1.22

this is result
https://imgur.com/a/0NaN4KJ

Hm... thanks for sharing but this is very strange. My 1080ti hashes for 20-21 MH/s. I run the miner at 24 Intensity with overclocks of +300 memory and + 90 core (about 1770 MHz I think?) at 80% power.
I obviously don't know the overclock limits of a 2080ti but maybe the default intensity is bottle-necking you.

z-enemy x16r 1.22 32bit

my 1070ti's: http://prntscr.com/l54zaw

but x16r hashrate, can go up and down depending on combined difficulty of the block-algos..

[revenant2017]

ok i tested both card with new version Z-Enemy 1.22

this is result
https://imgur.com/a/0NaN4KJ

Thanks for sharing this. Quite disappointed with the performance on x16r. Can you run a test on other algo as well? Excluding Lyra and Equih?

[badbart]

Won't the miners have to be updated to take full advantage of the RTX?