Pages:
Author

Topic: DIY FPGA Mining rig for any algorithm with fast ROI - page 41. (Read 99494 times)

jr. member
Activity: 117
Merit: 3

No. my implementation of the CN is very raw ... I'm ashamed to spread its code (((

Ah, I was all ready to give you a big pile of merit Cheesy -- I'll give you a little stack instead. Thanks for the share.




Do you design your boards? Would you like to test a few algorithms written by me?

Still waiting to receive my copies of the VCU1525. I'm currently just testing on a VCU118 until the dev boards get in.

If you want to pop into the discord here: https://discord.gg/ZEXCHt there's a big group of developers and i'll have one of the admins tag you.



Hello

Was just wondering if you knew if the bitsteam available now.(http://zetheron.com/Downloads/)

Will it work on the AES-KU040-DB-G board?

Thanks
hero member
Activity: 1118
Merit: 541

No. my implementation of the CN is very raw ... I'm ashamed to spread its code (((

Ah, I was all ready to give you a big pile of merit Cheesy -- I'll give you a little stack instead. Thanks for the share.


Do you design your boards? Would you like to test a few algorithms written by me?

Still waiting to receive my copies of the VCU1525. I'm currently just testing on a VCU118 until the dev boards get in.

If you want to pop into the discord here: https://discord.gg/ZEXCHt there's a big group of developers and i'll have one of the admins tag you.

newbie
Activity: 10
Merit: 3

No. my implementation of the CN is very raw ... I'm ashamed to spread its code (((

Ah, I was all ready to give you a big pile of merit Cheesy -- I'll give you a little stack instead. Thanks for the share.


Do you design your boards? Would you like to test a few algorithms written by me?
hero member
Activity: 1118
Merit: 541

No. my implementation of the CN is very raw ... I'm ashamed to spread its code (((

Ah, I was all ready to give you a big pile of merit Cheesy -- I'll give you a little stack instead. Thanks for the share.
newbie
Activity: 32
Merit: 0

Quote
ETH is absolutely bandwidth limited, achieving incredible efficiencies on many GPUs

If I am incorrect on that assumption then feel free to correct me. In answer to that all I can say is ETH may be Memory Bound on some cards but that doesn't mean it is globally bandwidth limited. If it is then link me to that proof please.

Ethash is entirely memory bandwidth and latency limited. That is why OC-ing the GPU does nothing for performance and why reducing CAS on the VRAM and increasing the memory clock speed is where all the performance improvements lie. The only way you are going to improve on it is with faster memory. And 4GB of QDR2 SRAM, even if it were possible, would be prohibitively expensive.

This is why the only available ethash "ASIC" yields no power efficiency improvement over a 1070. You _might_ be able to improve on it with a custom ASIC with 4GB of eDRAM, but that kind of silicon would also be obscenely expensive to produce, even for the current big players like Bitmain and Baikal.
jr. member
Activity: 168
Merit: 2
Some answers:

- The exact part number for VCU1525 is: DK-U1-VCU1525-A-G
(this is the version with a fan; it also comes in a fanless version which needs some type of external air movement system, not recommended; definitely go for the version with a fan)

- I realize even a single VU9P FPGA card might be out of range (pricewise) for some people.  The Avnet KU040 is $975:
https://www.avnet.com/shop/us/products/avnet-engineering-services/aes-ku040-db-g-3074457345630043740/
This FPGA is much smaller than the VU9P.  It can mine Keccak at 3GH/s, SHA-224 is profitable as well, and I am trying to fit Tribus into it which would run at 600MH/s and be quite profitable, but it is a tight squeeze to fit Tribus in there and I can't promise it yet.

- The smallest/cheapest board is the Nexys Video
https://store.digilentinc.com/nexys-video-artix-7-fpga-trainer-board-for-multimedia-applications/
(also available from Digikey and Avnet)
This board has an even smaller FPGA and can only do Keccak and SHA-224 (Bismuth), with an ROI of around 200 days.  So it isn't the most profitable, but it does have the amazing advantage that it needs no fan only can get by with a passive heat sink. 

- Some people have asked me about future boards.  There is the upcoming VCU1526 (VU13P FPGA), XUPP3R-VU13P, as well as the VU35P/VU37P FPGA's with high bandwidth memory (most useful for ethereum).  I do have some plans on supporting upcoming boards, it does require changes to the software, but for the near term the only four boards I am supporting are the VCU1525, XUPP3R-VU9P, KU040 and Nexys Video.  The reason I am focusing on those boards is they have the best ROI.  The absolute biggest FPGA available (VU13P) used in some of the 'upcoming' boards is so much more expensive that the ROI is actually worse.  It's kind of like the Titan V GPU; faster than a 1080 Ti, but the extra price isn't worth it.


In interested in this part: The smallest/cheapest board is the Nexys Video
When @whitefire990 you will release the bitstreams and support for this board? Seems it is affordable and easy to grab, I think the market will also be huge for this.


newbie
Activity: 10
Merit: 3

No. my implementation of the CN is very raw ... I'm ashamed to spread its code (((
newbie
Activity: 31
Merit: 0
What is/does that code show/do?

Best regards
Paul.
hero member
Activity: 1118
Merit: 541
newbie
Activity: 10
Merit: 3
member
Activity: 154
Merit: 37

Dude it's true?

It can't be true! Due to memory bandwidth limitation. There is a lot of many FPGA-mining manipulations last time, IMHO it's only a marketing strategy of some dev-kits distribution corps.

I searched and sourced in forums. Cards support only 2 algorithms. Only crpytonote v1 and lyra2

That’s not the case at all, those are two we can speak to. Dozens of developers are working on more algorithms, and I expect most of WhatToMine.com to be covered around shipment.

only if i could buy one of those FPGA's to mine keccak... but i think i will wait until this tech goes more "mainstream" and the fpga's with their software are more accessible

Care to clarify? All the support for Keccak is out now.
legendary
Activity: 3836
Merit: 4969
Doomed to see the future and unable to prevent it
Correct and these are not being designed and manufactured in secret, they can just as easily be forked away. And anyone that thinks because just their array can change on the fly does not understand the basic restrictions inherent in the design.

1. The supply chains to build these things are not infinite, and some lead time components stretch 42 weeks or more now.

2. You can not “fork away” from what’s coming in the FPGA world. There is exactly one piece - memory bandwidth - left, and it will be solved with in the year. It is already in the works.

OK, so your point 1 is making my point from above that they are not being made in secret and you point 2 is just wrong. Bandwidth is but one component and GPU's are not even Memory Bandwidth limited yet and when they are there is a thing called a bus that can be increased. Your arguments on memory can all be applied to any processing unit so your running on circular logic with that argument when it comes to a advantage one process has over another. Sure Programmable dies are not eproms anymore but what differentiates them from hard coded substrate is what makes it impossible for them to ever be on a even level. I am not arguing that they have not made great strides but you are in a fantasy land here if you think they can ever compete evenly.

This is just a foolish statement, it's about curtailing future centralization such as what we see with bitcoin. Lose the hyperbole. We tend to play the long game and are not the fools in this scene that can be manipulated like your off the cuff suggestion. You might be able to catch some redditt fools on that but not here. This has been a ongoing discussion since day one and in no way shape or form came as a surprise, as a matter of fact I would bet it was staved off for years just do to the discussions about what our stated reaction would be.

"some" reddit fools? You should check the subreddits again. Also, Those fools are what give your coin value. You might not want to call them fools.

Zcoin forked away from a single [marketing] comment. My lyra2z 8,8,8 design isn't even complete yet.


My coin? I do not control any coin, I just want to free the world from the yoke of tyranny through the monetary system. ATM monero is the best case scenario for that, if that changes then I will change with it. BTW that value was in existence long before those fools showed up, the only reason they are there is because its one big moderated echo chamber.

So Zcash forked away from a single marketing comment? I wouldn't know or care about that. I have no interest in any trusted setup, if I did I wouldn't be here I'd just stay under the banking yoke. BTW I find that hard to believe but stranger things have happened. My guess is you don't have all the information.

Your sole reason for being here may be monetary gain but don't automatically assume the same is for everyone else.

So just looking at FPGA mining. Intresting but what a price $4k-$5 per card. Any other cards out there cheaper and do the job.
I see intel offering some FPGA https://www.intel.com/content/www/us/en/fpga/devices.html

This is a niche industry that has been around for far longer than crypto currency and has it's place in many applications. IIRC inititially it was used as a testbed for cpu logic before the costly manufacturing process. It's alot more fun to find a bug in the logic before you roll out a few million cpu's. Just ask Intel.


Your comments are a circle around things that aren’t true - ETH is absolutely bandwidth limited, achieving incredible efficiencies on many GPUs. No one said FPGAs were being produced by the millions in secret.

“Bus can be increased” - FPGAs have the largest “bus” of any general purpose reconfigurable computational device bar none. What are you referring to? FPGA is also far from a niche industry, enjoying a market cap as big as the big GPU manufactures. If you want to argue give me real numbers in technical terms. Your statements are largely incoherent.


OK, lets step back a bit as this entire conversation is becoming incoherent and your reply to my reply to dll70dll is taken out of context. When I said niche to dll70dll I was specifically referring to just the act of mining. As he stated

So just looking at FPGA mining.

so I thereby answered in that context. And you twisting that around to a different context and then argued against that, which we all know is a logic fallacy as well as trollish behavior of which ironically I was accused of.

Now as to your comment

Quote
Bus can be increased” - FPGAs have the largest “bus” of any general purpose reconfigurable computational device bar none.

I don't see why you are even mentioning that as I never said they did not, all I have ever said is to increase bandwidth you just increase bus width. This is a given and I have never said otherwise anywhere EVER.

I am assuming you were trying to attack this statement I made

Quote
Bandwidth is but one component and GPU's are not even Memory Bandwidth limited yet and when they are there is a thing called a bus that can be increased.

with your previous statement and you are supporting that argument with this statement.

Quote
ETH is absolutely bandwidth limited, achieving incredible efficiencies on many GPUs

If I am incorrect on that assumption then feel free to correct me. In answer to that all I can say is ETH may be Memory Bound on some cards but that doesn't mean it is globally bandwidth limited. If it is then link me to that proof please.

Now I think this all comes down to whether we agree on this.

Quote
2. You can not “fork away” from what’s coming in the FPGA world. There is exactly one piece - memory bandwidth - left, and it will be solved with in the year. It is already in the works.

I believe the bolded is incorrect, that is not the only aspect of a FPGA that can be exploited. But I will not argue with you about that point as I am keeping my thoughts to myself on that. But I am sure you can find hundreds of people that will argue that point with you.

In closing I think you guys believe I am against FPGA mining as I am against ASICs and that is why I am in this thread butt hat is a misconception on your part. And I believe thats why I was attacked and insults directed at the project I support. I am not in any way shape or form against mining on FPGAs as I do not see them in the same category as many of you do. I do not believe there will ever be anything to fear from FPGA centralization and that is my greatest concern when it comes to the POW security model.

Also there are probably errors in here as I got tired of sorting this wall of text and I'm out of time ATM so I will just post as is and await your reply which I hope is cleaned up a little as I've tried to clean this one up.
member
Activity: 154
Merit: 37

Dude it's true?

It can't be true! Due to memory bandwidth limitation. There is a lot of many FPGA-mining manipulations last time, IMHO it's only a marketing strategy of some dev-kits distribution corps.

I searched and sourced in forums. Cards support only 2 algorithms. Only crpytonote v1 and lyra2

That’s not the case at all, those are two we can speak to. Dozens of developers are working on more algorithms, and I expect most of WhatToMine.com to be covered around shipment.
legendary
Activity: 3654
Merit: 8909
https://bpip.org
You sure about that? With ProgPoW having to make FPGA's replicate a GPU, it will erase benefits of specialization, and as such will make a GPU more cost-efficient as they have economies of scale at their side.

First we need to see progpow in action, i.e. adopted by coins. My guess is that we will continue to have different kinds of POW - ASIC-, FPGA-, or GPU-friendly - and the debate will take years to settle, if at all. There are still CPU/HDD algos out there refusing to die.
member
Activity: 72
Merit: 11
There are 3 threads, where fpga devs are selling(or incite to buy) this devices today.

Most of those algos were FPGA and ASIC resistant in past(like scrypt, x11,x13, quark, etc). But coin devs trying to make fpga-resistant and asic resistant algos. They will increase memory size and bandwidth requirements and make new resistant(today resistant, not tomorrow) algos, like mtp, equihash(with different parameters), phi2, etc.

So, we wait for the next chip and launch then Smiley -- Once the stratix and virtex devices with HBM2 hit the market, there's nothing that the coin devs will be able to do. Literally, nothing.


You sure about that? With ProgPoW having to make FPGA's replicate a GPU, it will erase benefits of specialization, and as such will make a GPU more cost-efficient as they have economies of scale at their side.

https://medium.com/@OhGodAGirl/the-problem-with-proof-of-work-da9f0512dad9
newbie
Activity: 32
Merit: 0
According to https://www.youtube.com/watch?v=x9DJ18y8LcI&t=2s it could be 1.9Gh/s Ethash at 150w which is incredible

If the video isn't an outright fake, I'm going to hazard a guess on what is going on here. It is almost certainly doing a solo running benchmark run with 0 generation DAG, which fits into the card's UltraRAM.
You'll never achieve that kind of a number any other way. Even if you put in QDR2 SRAM modules in those DIMM slots it won't help you because IIRC you can only get 288MB worth of those in - far less than the 3GB+ you need for current DAG size for ethash based coins.

It show 0.9999 GB DAG generated in the console, is that the size you are referring to?

I missed that, but now having taken a closer look:
1) DAG used is < 1 GB
ETH has a DAG of 3GB at this time.
ETC has a DAG bigger than ETH.

2) It apears to be using ethminer in OpenCL mode.

There seem to be a lot of red flags in those few seconds of video.
newbie
Activity: 32
Merit: 0

Dude it's true?

It can't be true! Due to memory bandwidth limitation. There is a lot of many FPGA-mining manipulations last time, IMHO it's only a marketing strategy of some dev-kits distribution corps.

I searched and sourced in forums. Cards support only 2 algorithms. Only crpytonote v1 and lyra2

And keccak. At least that is the publicly available ones.
There is nothing to say that many others don't exist, used by people who are mining with them under the radar.
full member
Activity: 729
Merit: 114
What is the eth performance of this card? what is the estimated price?

Best Regards.

According to https://www.youtube.com/watch?v=x9DJ18y8LcI&t=2s it could be 1.9Gh/s Ethash at 150w which is incredible


Dude it's true?
I doubt it... There is probably some trick used here

no way that's true.
newbie
Activity: 57
Merit: 0

Dude it's true?

It can't be true! Due to memory bandwidth limitation. There is a lot of many FPGA-mining manipulations last time, IMHO it's only a marketing strategy of some dev-kits distribution corps.

I searched and sourced in forums. Cards support only 2 algorithms. Only crpytonote v1 and lyra2
member
Activity: 531
Merit: 29
According to https://www.youtube.com/watch?v=x9DJ18y8LcI&t=2s it could be 1.9Gh/s Ethash at 150w which is incredible

If the video isn't an outright fake, I'm going to hazard a guess on what is going on here. It is almost certainly doing a solo running benchmark run with 0 generation DAG, which fits into the card's UltraRAM.
You'll never achieve that kind of a number any other way. Even if you put in QDR2 SRAM modules in those DIMM slots it won't help you because IIRC you can only get 288MB worth of those in - far less than the 3GB+ you need for current DAG size for ethash based coins.

It show 0.9999 GB DAG generated in the console, is that the size you are referring to?
Pages:
Jump to: