Topic: ZTEX USB-FPGA Modules 1.15x and 1.15y: 215 and 860 MH/s FPGA Boards - page 48. (Read 182443 times)

Received my module today. Had a hard time to start it but that was my own mistake  . All is very well documented on his homepage and almost ready to start mining. ZTEX is the man. If you want a quality product and great customer service go with him. With his newest miner version i get a hashrate well above 200 MH/s  . I'm very happy. Again, thank you for the support !

actually a 9600bps COM port is fairly enough for a 127 mining chain.

Very good!

The FPGA can count by itself. Makes one new data block (44 Bytes) every 11s or 4 Byte/s. Reading out the data costs another 48 Byte/s.

I just wondering about USB bus throughput.

Does miner send every datablock for sha256 calculation to the board? If yes, then one unit will load USB bus significantly. How many units will load USB2.0 to 100%?

Having several board costs more development time, I need more capital for having the stock and production costs are higher.

In other words, if I sell 60 single FPGA boards I earn less than with 15 quad FPGA boards,  but I also need less time for development, less capital for the stock, ....

I guess there is a lot I don't understand about this industry.

Why would 15x 4-FPGA boards cost more than 60x 1-FPGA boards if 1x 4-FPGA board costs less than 4x 1-FPGA board?

You do not consider the volume discount.

I just calculated the price. A 4 FPGA board would cost about 1150 EUR (+/- 5%) , i.e. 287 EUR/LX150. This is 40 EUR cheaper than the 327 EUR of a single FPGA board.

If you purchase 15 4 FPGA Boards the volume price is 990 EUR, altogether 14850 EUR.

If you purchase 60 single FPGA boards the volume price 236 EUR, altogether 14160 EUR.

4 FPGA Board make sense for licensed production customers because in this case the costs are dominated by the part prices.

watching

ztek could you build a 2x FPGA, 3x FPGA or even 4x FPGA board cheaper on a per FPGA basis that your current design?

The reason I ask is that most people looking to buy into FPGA are not casual miners.  They are looking to build hashing clusters.  Hopefully someday very large hashing clusters.  If a 2x FPGA board is cheaper than 2x 1-FPGA board then it makes more economical sense to go with a larger board.  I noticed one of your competitors has a lower price on 2x FPGA board however they aren't getting very good throughput on the FPGA. Still the lower production cost makes me think it would be possible to bring the cost down w/ multiple FPGA boards.

If you are concerned about capital risk I would imagine people would be willing to pay a deposit ($50 per unit?) for the first batch.

Some people might be interested in the lowest total cost because they only want or can afford $500 worth however my guess is most of the potential demand cares less on what the per card price is and more on the per FPGA & per MH price.

Burning 8 5850's and 2 5870's here at moment, But recently found more interesting things to invest in and going away for 2 months to india end of the year so its all on hold for now But FPGA is undoubtedly going take over if AMD go down the path that Nvidia have (Less core more complex algorithms or what not) which is what the 79XX cards will be. 78XX may just be the last good BTC mining card..... I hope not, but extra power isnt needed with current software (normal users) not even pushing last gen's tech (although both 40nm). Hell, the gen before (4xxx) can even still hold its own in most things. Fact is more cores are not needed which is why they are heading into the CGN architecture in the first place. Its just another bit of tech going stale and rotten with no real world 'improvements'. At least not like the good ol' days

Hence why i say we needed to make waves in the research of Gallium Arsenide and graphene, Although i hear manchester uni (where graphene was found out) as been given 50 million boost to try keep it at the heart of graphene development. Unlikly, with IBM (EVIL EVIL COMPANY) getting there dirty hands dirty with it.

Its only the gay patenet system and the old dogs (M$, IBM, Apple etc etc) stopping progression and drip feeding everyone. New companies who have anything just get brought/bullied out so the drip can continue. As i say, tech as been dead for a decade with no major advances really. Especially the last 5 years. It's just sad, or im turning into a misrable ol' bar steward

  i can live with that. Better is always a question of your point of view. But with 20+ cents in power cost it's no question what's better for me. I can shutdown my 5850s anytime and use them to watch movies or whatever. But for others with more effective cards and cheaper power cost it may be a different story. To each his own.

Thats a rather silly statement imo as its obvious FPGA is better. No offence We just need more people PRODUCING them, then comes lower prices. Seen as silicon shouldn't have even gone below 40nm, never mind the ridiculous crap we have now that wont last. 22nm is an achievement, 14nm, which intel currently as in its possession is absurd. I wouldn't even like to think about all the problems there having. The fact is silicon as been dead for 5-10 years, hence multi threading (which IS NOT THE ANSWER). The fact that theres been a replacement at alot lesser cost, heat, power and production cost that will even work with the modern tools we use for silicon is due to the powers that be and now were at the end of a could-e-sac with nowhere to go. Gallium Arsenide as been around for ages and research as been made (little) which would have stem'd the gap until we get the graphine process sorted. Now, were stuck here with no advance in technology and Intel/TSMC/Global Foundaries/Samsung/IBM etc etc all spending silly money on a technology (silicon based processors) with little or no improvement over the generation before. Most release 2 chips on the same node these days, This is just a scam. Yes, its 'improved' but its NOT NEW! If they plummeted as much money into Gallium Arsenide as they have with silicon just to get it down to the nm we have now, we would have some serious advancement. As it stands, It's been stagnebt for the last decade. Hell, i only got back interested cause of Bitcoins

Now, FPGA as got to be the future for bitcoin minung, Simple fact is its like ARM vs x86(x64), ones made for performance per watt. the other is chuck it all in and dont worry to much about consumption. In the end, although expensive right now, the risk takers will profit from getting involved with FPGA.

Also, AMD is shifting to CGN architecture which will be devistating for BC mining. Less cores which can do more. Good for games/folding @home but devastating for bitcoing mining. I expect all radeon HD8xxx to be based on the new CGN and not VLIM4 or VLIM5 so if you ask me, FPGA is the only way forward for bitcoing mining.

Of cause, this is my own collected knowledge and my opinion but i see where technology is and as been for a long time. There's been no progression really, Your just all been hood winked with PR and Sales tactics. AMD's Bulldozer is the perfect example that multi threading ISNT upgrading the CPU, it just allows software (Incredibly complicated software and HORRID to work with) to take advantage of having multiple cores. Multiple slower and slower cores. BACKWARDS! Thats why your 5 year old PC still works as good as a new one and you have no need to upgrade. I'm afraid that every-thing's getting so dumbed down, people who like to mess and play with there hardware just wont be able to in the future because of all the lock downs made by companies such as apple and intel alike. They even employ the people who do hack them, Clever move in all honesty. Bad for us, the consumers that actually care about technology though All going integrated bull crap

Pulled the trigger on one of your boards. ZTEX seems to have a very good customer service + he speaks my language (in case of problems)

FPGA is the future here in Switzerland.

IMHO an ATX case is not suitable for this purpose. Best way is to build your own case with material from DIY store.

The easiest solution is to purchase a shelf from the DIY market and put everything in it. The FPGA boards can be mounted with standoffs. Of course, this in not very beautiful and wastes space. (But at At least the hash rate per cubic meter should not be much less than with a GPU rig )

Nice.  I likely will need to wait until after the first of the year but I am interesting in pickup up 10 boards.  A single ATX powersupply makes it a lot easier to keep them managed.  You got any good ideas for mounting 10 boards in a standard ATX case?

A description (with a few images) about how a standard ATX PSU can be modified in order to power a cluster of FPGA board has been added to the initial post.

What model 5830? that's pretty impressive. care to share your secrets? phoenix 1.6.4 with phatk2 and overclocked only gets me to about 270.

Just find 2 other people you are willing to trust with $10K+ and split a 50 GH/s run. 

Heh, if you could reach that $/MH on a 15GH/s build (or close),
and mentioned it a few months ago, it could have saved me quite a lot of time  

-rph