Pages:
Author

Topic: Baikal Giant N - Cryptonight, Cryptonight-lite FPGA/ASIC miner - page 10. (Read 32830 times)

newbie
Activity: 8
Merit: 0
why is every one calling the giant n a doorstopper ? 

https://www.nicehash.com/profitability-calculator#baikal-giant-n;USD;0.1;1;60;0;a22=20       i dont complain on this profit if you buy it now so an some one plz explane to me why every one is throwing shit on this miner atm ? and karbo still not gonna fork what i can read 

newbe to the baikal miner so plz can some one plz let me get some info from you guys

Once (if) ETN algo forks, which i heard they might next week block 250000 or something.. that is where most of the old cryptonight algo has is at so the profitability for the rest of the algo coins will plummet





so karbo isent strong enought in other words to keep the bk-n profitable?
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
I will publish an X16R bitstream which can do Ravencoin at an incredible rate; however these dynamically changing algorithms are extremely complicated to implement in FPGA's so it will be early 2019 by the time I have that one ready.

Why don't you do one at a time and store the result in sram. 1/16 the speed or less, but still faster than the gpu's right?

You can also make 256 kernels by combining 2 of the 16 hashing functions in one kernel. then you get 1/8th of the speed.
sr. member
Activity: 689
Merit: 253
why is every one calling the giant n a doorstopper ? 

https://www.nicehash.com/profitability-calculator#baikal-giant-n;USD;0.1;1;60;0;a22=20       i dont complain on this profit if you buy it now so an some one plz explane to me why every one is throwing shit on this miner atm ? and karbo still not gonna fork what i can read 

newbe to the baikal miner so plz can some one plz let me get some info from you guys

Once (if) ETN algo forks, which i heard they might next week block 250000 or something.. that is where most of the old cryptonight algo has is at so the profitability for the rest of the algo coins will plummet
newbie
Activity: 8
Merit: 0
why is every one calling the giant n a doorstopper ? 

https://www.nicehash.com/profitability-calculator#baikal-giant-n;USD;0.1;1;60;0;a22=20       i dont complain on this profit if you buy it now so an some one plz explane to me why every one is throwing shit on this miner atm ? and karbo still not gonna fork what i can read 

newbe to the baikal miner so plz can some one plz let me get some info from you guys
full member
Activity: 420
Merit: 184
Sorry, but it looks like you are too late. Only tiny market cap coins are left that are still running the original CN algo and it sure looks like ASICs have found them because the network hashrates (and difficulty) of two that I watch (ITNS & DERO) have gone up by >10x over the last two days.

Both of those coins are barely worth mining now, even with an ASIC, and both are going to hard fork at some point anyway; ITNS is scheduled for May 14 while DERO is "imminent."

This is very bad news. But then why does whattomine get this result? Is it outdated?:

[link deleted - cryptonight profitability on whattomine.com]

What is the result of mining in nicehash?.

It does look like you found a relatively high market cap coin still using CN - karbo - but even in the time you posted this reply it went from $16.50US per day at 20kH/s to around $11.50US per day... A good thing to remember is that any coin listed on whattomine.com should be assumed to be well known by the mining community, so you are unlikely to get a competitive advantage deploying your ASIC on it; the best situation for an ASIC owner is to only have to compete against GPUs and CPUs, not other ASICs.

copper member
Activity: 166
Merit: 84
How will the VCU1525 perform on Cryptonight?
Thanks,
Jeff

I haven't gotten to a full implementation of CryptonightV7 yet, but based on simulations, 8KH/s per VCU1525, so 64KH/s for an 8-FPGA rig.  Which means the ROI is not stellar but still better than a GPU.  There are much more profitable algorithms to mine with the VCU1525 (with < 100 day ROI's), but still the VCU1525 does reasonably well on cryptonight 7; hashing at 4 times a Vega 64, consider that the VCU1525 burns around 100W at 8KH/s, vs. >800W for 4 Vega 64's hashing at the same rate.  So in terms of price-per-electricity, it has a huge advantage.  Still, you'd make a lot more money mining neoscrypt, tribus, phi, X17, etc...


newbie
Activity: 14
Merit: 0
nicehash working for giant N ?
or wich pool use ?

I have been using Nicehash for my two Giant-N.  It worked flawlessly so far.  So far I made $610 on 2 x Giant-N.

I don't use Nicehash for any other rigs though... but for Cryptonight it seems good.



How will the VCU1525 perform on Cryptonight?

Thanks,

Jeff
newbie
Activity: 82
Merit: 0
Any news of change of algorithm to CryptonightV7 of Karbo (KRB), Electroneum (ETN), Bytecoin (BCN), Dinastycoin (DCY) and DigitalNote (XGN)??.

Are there other coins with cryptonight algorithm?.

Is it still viable to buy a baikal for about $ 1000?.

Best regards.

Sorry, but it looks like you are too late. Only tiny market cap coins are left that are still running the original CN algo and it sure looks like ASICs have found them because the network hashrates (and difficulty) of two that I watch (ITNS & DERO) have gone up by >10x over the last two days.

Both of those coins are barely worth mining now, even with an ASIC, and both are going to hard fork at some point anyway; ITNS is scheduled for May 14 while DERO is "imminent."



This is very bad news. But then why does whattomine get this result? Is it outdated?:

http://whattomine.com/coins?utf8=✓&adapt_q_280x=0&adapt_q_380=0&adapt_q_fury=0&adapt_q_470=0&adapt_q_480=1&adapt_q_570=0&adapt_q_580=0&adapt_q_vega56=0&adapt_q_vega64=0&adapt_q_750Ti=0&adapt_q_1050Ti=0&adapt_q_10606=0&adapt_q_1070=0&adapt_q_1070Ti=0&adapt_q_1080=0&adapt_q_1080Ti=6&factor%5Beth_hr%5D=1500.0&factor%5Beth_p%5D=1500.0&factor%5Bgro_hr%5D=348.0&factor%5Bgro_p%5D=1260.0&factor%5Bx11g_hr%5D=117.0&factor%5Bx11g_p%5D=1020.0&cn=true&factor%5Bcn_hr%5D=20000.0&factor%5Bcn_p%5D=60.0&factor%5Bcn7_hr%5D=20000.0&factor%5Bcn7_p%5D=60.0&factor%5Beq_hr%5D=4110.0&factor%5Beq_p%5D=1140.0&factor%5Blrev2_hr%5D=384000.0&factor%5Blrev2_p%5D=1140.0&factor%5Bns_hr%5D=8400.0&factor%5Bns_p%5D=1140.0&factor%5Bbk14_hr%5D=26100.0&factor%5Bbk14_p%5D=1260.0&factor%5Bpas_hr%5D=10200.0&factor%5Bpas_p%5D=1260.0&factor%5Bskh_hr%5D=285.0&factor%5Bskh_p%5D=1140.0&factor%5Bn5_hr%5D=220000.0&factor%5Bn5_p%5D=480.0&factor%5Bl2z_hr%5D=420.0&factor%5Bl2z_p%5D=300.0&factor%5Bxn_hr%5D=31.8&factor%5Bxn_p%5D=1140.0&factor%5Bcost%5D=0.1&sort=Profitability24&volume=0&revenue=24h&factor%5Bexchanges%5D%5B%5D=&factor%5Bexchanges%5D%5B%5D=abucoins&factor%5Bexchanges%5D%5B%5D=bitfinex&factor%5Bexchanges%5D%5B%5D=bittrex&factor%5Bexchanges%5D%5B%5D=binance&factor%5Bexchanges%5D%5B%5D=cryptopia&factor%5Bexchanges%5D%5B%5D=hitbtc&factor%5Bexchanges%5D%5B%5D=poloniex&factor%5Bexchanges%5D%5B%5D=yobit&dataset=Main&commit=Calculate

What is the result of mining in nicehash?.

Regards.
full member
Activity: 420
Merit: 184
Any news of change of algorithm to CryptonightV7 of Karbo (KRB), Electroneum (ETN), Bytecoin (BCN), Dinastycoin (DCY) and DigitalNote (XGN)??.

Are there other coins with cryptonight algorithm?.

Is it still viable to buy a baikal for about $ 1000?.

Best regards.

Sorry, but it looks like you are too late. Only tiny market cap coins are left that are still running the original CN algo and it sure looks like ASICs have found them because the network hashrates (and difficulty) of two that I watch (ITNS & DERO) have gone up by >10x over the last two days.

Both of those coins are barely worth mining now, even with an ASIC, and both are going to hard fork at some point anyway; ITNS is scheduled for May 14 while DERO is "imminent."

newbie
Activity: 82
Merit: 0
Any news of change of algorithm to CryptonightV7 of Karbo (KRB), Electroneum (ETN), Bytecoin (BCN), Dinastycoin (DCY) and DigitalNote (XGN)??.

Are there other coins with cryptonight algorithm?.

Is it still viable to buy a baikal for about $ 1000?.

Best regards.
copper member
Activity: 166
Merit: 84
nicehash working for giant N ?
or wich pool use ?

I have been using Nicehash for my two Giant-N.  It worked flawlessly so far.  So far I made $610 on 2 x Giant-N.

I don't use Nicehash for any other rigs though... but for Cryptonight it seems good.
copper member
Activity: 166
Merit: 84
However, some algorithms like Timetravel8, Timetravel10, X11Evo, X16R, X16S, they change the algorithm every block.  An FPGA can rapidly reconfigure itself to optimize itself each block; an ASIC cannot.  Therefore these algorithms are truly ASIC-proof, in that no ASIC can beat an FPGA (ever) on those coins, unless the ASIC itself is just designed as an FPGA which would be a waste of money since you can just buy ready-made FPGA's.

Correct me if im wrong but i think that algos like X16R still can be attacked with ASICs, you dont have to implement all of the algos that this algo is cycling through you just have to pick the once that are for example 3-4 times even more faster on ASIC compared to a GPU with a much less electricity consumed then you mine when blocks with the implemented algos are coming.

You should re-read the X16R white paper.  X16R uses 16 algorithms on every single block, done in sequence, in a different order.  You need to understand how pipelining works.  An ASIC capable of all 16 algorithms would not be able to pipeline the algorithms because the pipeline requires the order to be the same.  Titanic multiplexers might solve the problem for Timetravel8 and Timetravel10, but X16R has the extra change that some times you do one algorithms 3 times in the same set, making an FPGA the only method that could run the algorithms in a pipeline.

Regarding the VCU1525; the price is $3995 - $4600 USD, this is an incredible deal considering the FPGA on the board costs $40,000.  Since Xilinx makes the FPGA they can sell the board for whatever price they want.  I am trying to make a deal with Avnet+Xilinx to establish a custom channel for crypto miners to buy VCU1525 at short lead time.  Stay tuned in the next few days.

member
Activity: 163
Merit: 11
 For example, the VCU1525 can make a decent profit even while mining Myriad-Groestl (Baikal X10 ASIC), so even competing against a dedicated ASIC, the FPGA can still mine with profit, although the ROI will not be maximized (ROI when competing against ASIC's is 1-2 years; ROI when mining non-ASIC coins is 50-150 days).
any idea where one can easily buy a VCU1525? digikey has very, very long lead time.

How it possible to pay this fpga around 1000 usd when the italian distributor make the crazy price of 4000€ plus taxes i speak about the vcu1525
full member
Activity: 162
Merit: 100
supportXMR.com
 For example, the VCU1525 can make a decent profit even while mining Myriad-Groestl (Baikal X10 ASIC), so even competing against a dedicated ASIC, the FPGA can still mine with profit, although the ROI will not be maximized (ROI when competing against ASIC's is 1-2 years; ROI when mining non-ASIC coins is 50-150 days).
any idea where one can easily buy a VCU1525? digikey has very, very long lead time.
newbie
Activity: 95
Merit: 0
A lot of people have asked me about the FPGA rig that I am building, and make a long story short, I am hoping eventually to release bitstreams for a couple of very common FPGA boards; in this fashion, the average person can just buy one to 30 of these 'stock' FPGA boards, connect them to their PC by USB cables and PCIe-to-USB3 cards, and start mining with the publicly available bitstreams for a 2% fee.  The ROI on high end FPGA boards right now is 15 to 90 days depending on the algorithm and the board.  This setup is almost immune to 'forking', and in my opinion, GPU's will gradually be replaced by FPGA's and I believe stock-hardware FPGA mining with publicly available bitstreams will replace the current set up of stock-GPU's with publicly available mining software.

When the transition from GPU's to FPGA's is complete, true-ASIC rigs will not be that attractive.  They will offer only a moderate hash rate increase, for high risk.





Very very interested. When will you have them ready? And where is it possible to buy those high end fgpa? Thanks!

Other than myself I know of at least two other people who are also working on open-platform FPGA rigs, which means in reality there must be even more than that.  Likely several will be released around the same time.

High end cards are available from digikey.com, avnet.com, hitechglobal.com, bittware.com, xilinx.com.  The lowest end card that can ROI in around 30 days is the $490 Digilent Nexys Video [Xilinx XCA7200T] (available from Digilent.com, Avnet.com, Digikey.com).  However, the Nexys card is limited in which algorithms you can mine, and personally I believe the future of open platform FPGA rigs is in the high end cards which can run almost every altcoin algorithm.  The high end cards cost around $4K to $6K each, which is around the same price of custom mining rigs, with much better ROI's, more flexibility, and none of the 'screw-you-over' attitude of the big mining companies.  With full open access to your own hardware, there are no 'secret' or 'locked' algorithms which are out of your control.

I'm working full time on this project, I might have something publicly available by June or July.

I forked tpruvot's CPU miner, the miner works the same on the command line, with -a specifying the algorithm, and the PC mining software loads the correct bitstream into the FPGA card right before you start mining.  If you want to run profit switching, the software just reconfigures the FPGA in a few seconds and then switches algorithms.

The other tremendous gain is the low power consumption.  A high end FPGA card burns around 150-200W and makes $40 to $270 per day.  Which means you can 'live' off mining revenue without a complicated cooling system, 220V circuits, and all the other headaches of GPU's.





Very interesting indeed, thanks for sharing. I have a few questions, but for folks doing searches, the correct part for the Nexys is XC7A200T

So, on to the questions, and please correct me wherever my assumptions are wrong; not an expert in FPGA by an means.

1. The low-end fpga in your example, seems to only come with 512MB 800MHz DDR3, which also means that it would not be able to mine algos that require 2GB of memory or more, correct?
2. Do you have a few high-end fpga example boards to link to?
3. Do you have a working proof-of-concept, what algo is it running, on which board, and is it stable?
4. Further to #3 above, what are the temperatures like at full blast?

The high end board I am developing on has 64GB of DDR4 and there are 8 boards in the rig (512GB), however another board I am working with has 288MB of QDRII+ RAM which is way faster than DDR4/GDDR for random access.  Still, generally the best ROI is for algorithms that are not that memory intensive and can be done internally on the FPGA without external memory accesses.

I am building hash functions one by one (functions, not algorithms, like X11 = 11 different hash functions), and no, I have not finished enough hash functions to run a full multi-function algorithm yet.  However based on the speed of the individual hash functions it is possible to calculate the hash rate of the full chain well in advance.

Anyway, time to get back to work.



Nice work! I think you may firstly profile the C version algo, then find out the hotspot. After that, you may try to split the algo into CPU part and FPGA part. Lastly, implement the FPGA part with RTL. Right? I am interested in implementating the CN heavry algo with FPGA. Just cannot start the project. :-)
jr. member
Activity: 108
Merit: 1
However, some algorithms like Timetravel8, Timetravel10, X11Evo, X16R, X16S, they change the algorithm every block.  An FPGA can rapidly reconfigure itself to optimize itself each block; an ASIC cannot.  Therefore these algorithms are truly ASIC-proof, in that no ASIC can beat an FPGA (ever) on those coins, unless the ASIC itself is just designed as an FPGA which would be a waste of money since you can just buy ready-made FPGA's.

Correct me if im wrong but i think that algos like X16R still can be attacked with ASICs, you dont have to implement all of the algos that this algo is cycling through you just have to pick the once that are for example 3-4 times even more faster on ASIC compared to a GPU with a much less electricity consumed then you mine when blocks with the implemented algos are coming.
newbie
Activity: 95
Merit: 0
Hi, guys. Some friends told me that they can update the Baikal Giant N to support the Cryptonight V7. I just have no idea how it is possible... I think the Baikal N is ASIC, not FPGA. Updating the firmware only cannot support new algos. Right? Maybe they have some other way?
copper member
Activity: 166
Merit: 84
Thank you for the amazing work you have been doing with FPGAs. I truly respect you for pushing the boundaries of today's mining knowledge.
So let's say that FPGAs turn out to be a healthy and profitable substitution to GPUs as we hope, so far Cryptonight is the only algo that was able to put ASICs out of the game for a while, this being said, won't ASICs just be developed for the remaining algos, beating FPGAs at what they can do best?
If so this leaves me thinking that FPGAs profitable work could then be allocated to Cryptonight, which is ASIC resistant as we speak, but as you have stated yourselfthis algo implementation is not the best mining use case for FPGAs and at that point once more accessible FPGA mining solutions are released and more FPGA hardware is purchased won't that put us back where things are now?

An expertly programmed FPGA can even compete with an ASIC on the same algorithm, although the ROI is not spectacular, it is still worthwhile to mine.  For example, the VCU1525 can make a decent profit even while mining Myriad-Groestl (Baikal X10 ASIC), so even competing against a dedicated ASIC, the FPGA can still mine with profit, although the ROI will not be maximized (ROI when competing against ASIC's is 1-2 years; ROI when mining non-ASIC coins is 50-150 days).

However, some algorithms like Timetravel8, Timetravel10, X11Evo, X16R, X16S, they change the algorithm every block.  An FPGA can rapidly reconfigure itself to optimize itself each block; an ASIC cannot.  Therefore these algorithms are truly ASIC-proof, in that no ASIC can beat an FPGA (ever) on those coins, unless the ASIC itself is just designed as an FPGA which would be a waste of money since you can just buy ready-made FPGA's.

Once I publish the website for the DIY-FPGA mining, I will strongly push new crypto coins to go with one of those dynamically changing algorithms, since they are the only ASIC proof algorithms and they would guarantee decentralization in the future world of DIY-FPGA mining.

I will publish an X16R bitstream which can do Ravencoin at an incredible rate; however these dynamically changing algorithms are extremely complicated to implement in FPGA's so it will be early 2019 by the time I have that one ready.

jr. member
Activity: 59
Merit: 1
..
So let's say that FPGAs turn out to be a healthy and profitable substitution to GPUs as we hope, so far Cryptonight is the only algo that was able to put ASICs out of the game for a while, this being said, won't ASICs just be developed for the remaining algos, beating FPGAs at what they can do best?
If so this leaves me thinking that FPGAs profitable work could then be allocated to Cryptonight, which is ASIC resistant as we speak, but as you have stated yourselfthis algo implementation is not the best mining use case for FPGAs and at that point once more accessible FPGA mining solutions are released and more FPGA hardware is purchased won't that put us back where things are now?

Let me answer. All the answers to your questions - yes. But you can do a lot before these problems arise.

I think that ASIC miners will appear for all more or less good crypto-currencies that exist now. But until then, the FPGA will bring income of 100-200% for 1-2 years, it is quite real.

There are several ways to keep FPGA solution, inclding:
1. Adaptation to reality every 3...6 month by implementing a new algorithm.
2. Creating FPGA-solver, that is impossible to reproduce in ASIC more effectively, than in FPGA.
3. Creating of altcoin with POW function that would be usable in some real-life application (genome, networking, etc). See Amazon cloud instance F1 with Virtex Ultrascale+, it's price ~1 USD/h..

Speaking of current problems, it is much more simple to implement miners by using existing hardware. Ironically, people who have FPGA (and programming skills) do not trust in cryptocurrency and do not consider such activity. So we have a chance.

newbie
Activity: 12
Merit: 0
Gosh! Looks like the trend is changing like crazy...I guess GPU would gradually back off and FPGA is the future of mining
Pages:
Jump to: