Topic: DIY FPGA Mining rig for any algorithm with fast ROI - page 53. (Read 99494 times)

Have you assessed their latest videos?

I’d be really interested to see those claims be accurate. I feel like my cryptonight7 is pretty optimal and those numbers are just insane.

Actually from the claims, the Dwarf boards could do 100kh/s at $4000 price-point. If true, ie.

Hes not. He was misleading any coin devs.

"I'm now mining monero at 100Kh/s per fpga."
It sounds great. Which FPGA do you use?

If having two GPU per MB will be the new way with this mining method, lets prepare for a shortage on motherboards
Quite interesting to see where this leads.

@ GPUHoarder

The M2 accelerator units seem pretty interesting as they would bring some lifespan extension to the GPUs.

You mentioned in a post before that a GPU must be put in 16x or 8x slot and M2 accelerator can boost that  card. Would that mean that boards such as the Asrock B250M Pro4 which have    2x  PCI Express x16 and 2x M2 Ultra slots can accommodate two accelartors for two GPUs plugged in the 16x pci-es?

I would guess this reason is more than likely why Xilinx is hesitant to let anyone else in the crypto community have their chips. Bitmain is already "partnered" with Xilinx.

by then Bitmain FPGA multi algo miner will be available mining away at all the high profile coins.

I’ll start a thread when I’m ready to post all the details - currently expecting that to be about 2-3 weeks.

I’ve got everyone that PM’d me in a list for first dibs, and I’ll be replying with early details soon.

My bad. Sorry mis read the post.

Are you going to start a thread on your m.2 accelerators? Ebay listing?

I think you’re right in that Intel is farther along. They’ve already implemented HBM (or rather MCDRAM) in KNL and have the interposer experience.

I should clarify I don’t think HBM is being abandoned, but it is being relegated to higher cost cards and GDDR6 is taking over in the mass production end. I remember buying 100 Fury X’s in 2015 and having the serial numbers be hand written and I think they incremented maybe a few thousand over the year I accumulated mine.

If we count Vega as consumer cards, it seems AMD is keeping HBM for their next gen, the Vega 20.

As far as I'm aware they are at least in some limited fashion shipping the stratix 10 mx series. I'll confirm next week for sure. But I do remember hearing about problems they had in 2017 and how they wouldn't be resolved until 2018 because they had to redesign the stratix 10. I believe what we are now seeing is that redesigned chip. As far as I know, altera had never planned 58G serdes in the stratix 10 -- unless I missed it in the literature of the time.

Read the rest of senseless’ comment - no one actually has 100kh cryptonight on one chip. He was making a commentary on how whole coins are making fork decisions based on comments in this post.

Both Xilinx and Intel are failing to sample or ship HBM chips announced more than six months ago. I don’t think that HBM is coming at mainstream prices in FPGAs anytime soon - even AMD and Nvidia are moving away from it in consumer products.

Are you going to share or just boast?

What about this very critical point of view about Intel Stratix 10 HBM ?
https://www.semiaccurate.com/2017/12/19/intels-claims-fpgas-hbm-dont-hold-water/

I think our definitions of flood the market differ. You're thinking about it from a consumer level -- It will eventually be that way -- But not in the next 12 months. If intel had the same practices as Xilinx, the Xeon 4116 system I just built would have cost me $100,000... Not $6,000... And I would have had to design my own motherboard, because they wouldn't have allowed me to produce their reference design... I don't think it's unreasonable to expect a $500-1000 unit cost on a large Stratix or Ultrascale+ part with HBM2 once production is moving 100K units a month. If Xilinx has their way, that part would cost $5,000-10,000 at "volume". If Intel increases their market share by expanding the market, and making these chips available at the enterprise level for hardware acceleration of databases, web servers, etc.... It's not unreasonable to think they could move 100K units a month at a $500-$1000 cost.. Amazon is currently leasing Xilinx fpgas. Do you think AWS would be interested in including FPGA logic in every single one of their servers? What about tencent? What about Baidu? What about every other provider on the planet? There's a storm on the horizon and it's going to be beautiful to watch.

The value of these devices in the server markets is in the 100s of billions. Far beyond their value in crypto markets, defense, etc, IMO. Xilinx has been very short sighted in their attempt to make profit. The world needs low cost FPGA. Not having it is preventing major advancements in computing.

Not exactly flooding the market with cheap ultrascale+ class products ...

https://www.anandtech.com/show/12773/intel-shows-xeon-scalable-gold-6138p-with-integrated-fpga-shipping-to-vendors

sampling now to special customers. spendy and only an arria 10.

it'll be quite a while before we're seeing affordable integrated stratix10 / x86 products that mom and pop can afford.