Topic: Algorithmically placed FPGA miner: 255MH/s/chip, supports all known boards - page 32. (Read 119429 times)

I'm not sure whether it is reasonable to use the tricone bitstream on ZTEX FPGA board. Here here my considerations.

1. I made some more exact current measurements with the ZTEX bitstream. At 212 MHz the core current is approximately 6.7 A, i.e. the frequency limit with this bitstream is about 250 MHz on ZTEX FPGA boards (with 8A core voltage regulator).

2. According to eldentyrrel the tricone bitstream is about 13% less efficient. This bitstream  would have to be limited to 217 MH/s on the ZTEX FPGA boards.

3. At 12V input voltage it is probably possible to override the max. current of AOZ1025DI by 0.5A   (this would require some long term tests at 9.5A). But IMHO its not worth it: At 8.5A the tricone bitstream should deliver about 233 MH/s. The price for additional 17 MH/s (21*0.8, 20% goes to eldentyrrel) is  approx. 2.5W more power on the wall and a reduced reliability.

4. I'm concerned about the reliability (due to the 2-year warranty):

4.1. At 8A the power dissipation of the FPGA is about 10W. The thin CGS484  packages have a junction-case thermal resistance of 2.2 K/W. Plus 0.3 K/W for the thermal grease this results in  junction temperature of 70.5°C, if the bottom of the heat sink is 45°C warm. This should be still o.k. but there is not much margin for improper installed heat sinks or so. And  many users had problems with this because the plastic packages are not very flat.

For comparison: The thermal resistance of the thick FGG484 packages which are used for most other LX150 FPGA boards is 3.7 K/W. Plus 0.3 K/w for the thermal grease
leads to a junction temperature of 85°C at 8A,  96°C at 10A, and 106°C at 12A

4.2. There is no on-die temperature sensor. If the heat sink is not installed perfectly the core temperature reaches critical levels and there is no chance to recognize this. The indirect overheat protection by error measurement (as implemented in BTCMiner) does not work if the frequency is limited.

I'm sorry but due to the 2 year warranty and from my experience its seems to be too critical for me to support the eldentyrrel bitstream actively. (Of course, everyone can do this on ones own risk.)

Oh, probably not that much worse than other open source software developers here. What I wanted to point out is that more than 90% of that was paid by board manufacturers who want me to support their boards, not users, and that users paid much more in direct donations than in the profit share model.

So if you want some software developer to be compensated somewhat fairly, you will either have to force the profit share model on users like ET does, or you have to make deals with the hardware manufacturers who are hoping to increase sales by taking advantage of your software. However both of which are unlikely to be competitive if you consider the real time investment.

I can't understand the harsh criticism from people here on him / his concept.
It's actually simply perfect for everyone involved.

Fact is - no other concept is better for him and his "customers"

- Donations? Most people don't donate - fact - and even if some do - those are usually rather small amounts
- Selling the bitstream? (some) people will resell it to recover their "investment" and sooner or later it will be "free"
- Further, he couldn't make a difference in terms of pricing for one or a hundred fpgas in use at that customer (it wouldn't be a fair system)

He created it - it's his property - it's his decision what to do with it... Simple as that.
He did what was fairest to everyone - and puts everyone including himself in a winning position.
You don't need to pay upfront, you get more performance out of your existing hardware - and he gets a 20% cut of that additional performance. Don't like it ? Don't use it !

Same goes for server down times, higher stress on the hardware, higher power consumption... and so on - nobody forces you to use Elden's system

"of course we don't just want to sell it to you outright - we want to be able to license it to you" 

Though, of course, if the commission system is shut off, that means you will have to put your FPGA's back to their original firmware.
The new firmware will not work with any other pool directly.

So I'd take that to mean that no FGPA developer would supply hardware with the new bitstream
If you want it you will have to be able to update your bitstream yourself
... since it depends on another person keeping an internet resource available to work.

seems there's gotta be some sort of reasonably fair compensation method we could all agree on... though i've no idea what it is. upfront pricing is disagreeable to some, this commission idea disagreeable to some... hmm. this just struck me as i type this.

what if the commision system was intentionall shut off say, a year from now (or whatever) after appropriate compensation had been given? ET gets the compensation he feels apropriate, we get another small increase in hash, and no one pays upfront either.

And introduces a centralized point of failure into the equation as well...

This is exactly why, when I first read of the share commission model, that I felt that greed was the sole motivator of this work.  Maybe this was a snap reaction that warranted further deliberation.  Whether ET gets what he wants, much greater or much less I really have no clue.

I do know that I always dislike perpetual relationships where I have no negotiating power and, most importantly, no idea how much I will eventually pay for a product or service.

Well if he does then he is quite literally the first person around here to want to be paid that much for software made for the BTC community in the last year

ckolivas has spend way more time and effort on cgminer and got nothing like that for his efforts
Even I did one of the highest paying software bounties round anywhere here in a long time when I added the API to cgminer at the end of the year but even over the short time I took to implement it, the 'pay' was certainly less than the time I spent on it
It initially only took me 2 days to do the basic implementation ... but a lot more time until everyone was using it ... but I've spent a lot more time and effort on it since for what totals less than 1/2 of that initial bounty

hmm - what was that history about Bill Gates and Basic ...

But the arguments given above about it - well 1 of them is certainly relevant IMO and that is the cost of mining certainly does make the % higher
(that's something thing that many people seem totally oblivious of when they do calculations: the cost of mining)

But that other argument is certainly original:
People using the bitstream MUST trust eldentyrell but he doesn't trust anyone using it
I would imagine it would be difficult in the short term for someone to copy his design and it not be obvious when they released their 'copy'

However in the long term, I'd be very surprised if any bitstream released by anyone was 'the best' for more than a few months.

I was estimating 1000-1500 in small farms which would actually use the ET bitstream. Larger farms are more likely to get a bitfury bitstream

I would actually be surprised if there were only 1500 LX150s around, not including Bitfury. I'd also expect that number to jump pretty quickly in the coming months, unless what BFL is waiting to announce really is as revolutionary as they imply and people believe their specs/timeline.

I still wonder about the accuracy of the 1000-1500 FPGA numbers. I'm not saying anyone is wrong, but do those figures include the various privately licensed production runs that have been done around here? I'm sure Antirack isn't the only one to have done this, and I'm not even sure how many devices his run includes.

It would probably be sufficient to match what he's about to get from the profit share model. Which I'd estimate at about 1000-1500 FPGAs * 5% * a year of mining. Probably like $20k or BTC4k.

Depends if he actually wants to get paid fair market value for his time. I'm an associate Engineer and my company charges out my time at over $100/hr. How long do you think Tyrell spent on this bitstream?

I think a group buy would be the best then release the source so the community can tweak it as well. Except how much would one charge for a bitstream? 500 btc? 1000?

Why

I really don't understand ... I still think he should sell bitstream outright like say if you have 10 FPGAs one price and for 100 FPGAs another price and for just 1 or 2 another price like antivirus licenses etc.

This whole taking a cut and his own servers is junk for miners. DDos his servers before diff change and everybody wins though as most FPGAs are then neutered

Hey Entropy-uc, there hasn't been one report of somebody other than ET using this. You are jumping to conclusions. His solution will be scrutinized and time will tell if it is any good. If you don't like it, don't use it. I am happy to give ET the benefit of the doubt and a big KUDOS too. Everything you wrote is a huge pile of Bulanula.

Yes, I got one person so far, yourself.

So, who are you really?  Looks like you created a special account to come and espouse your particular point of view.

Part 7:

Would miners prefer to invest upfront paying money for disclosure and make benefits to parties who are not miners yet OR would they invest money into hardware ? What is more profitable ?

Even without calculation - I would say that it is more profitable to invest in hardware.

Then - who will compensate ? Community in general is not interested in that, community of FPGA miners is likely to invest in FPGAs and not in bitstream (they would better take bitstream "for free" not counting how difficult it is to get one working). Me or someone else with decent bitstream to slow down development of open-source baseline... hmm... actually too late, as 28-nm will be too soon... Hmm... Developers of these boards ? Many do the job as their hobby - no significant volumes - they will return money for ages...

Please read what I wrote:

Part 6:  Evaluating ET's potential return

By now you probably have noticed that I think this deal is not very good for miners.  But you should always look at both sides of a deal.  ET must be raking it in on these terms right?

Uh No.  It's a lousy deal for him as well I think.

My calculations suggest he would be doing well to make $30k / year off this plan after the reward halving in December.  And for that he is going to have to maintain his server, fend off the DOS attacks, deal with user complaints, and provide technical support to everyone using the system. Whatever he earns is going to be eaten away by whatever downtime compensation he offers. Not to mention his out of pocket costs for hosting.

All of that work is going to detract from what he does best, and what I'd like to see him focus on.  Building faster bitstreams.

So the pertinent question here is: How can the community fairly compensate him?

I would like to see something that both encourages him and others to continue working on improvements, and also encourages collaboration so that others with an interest can build off of the baseline.