I've never seen a honest comparison involving actual verification of accuracy, not even bit-accuracy. I've seen some very skewed benchmarks made with very ugly code that conflated/convolved FPU performance with memory bandwidth/latency limitations. https://en.wikipedia.org/wiki/False_sharing seems to be in fashion nowadays for obfuscation purposes.

Frequently the comparison don't even use the real floating-point but some extended-precision fixed-point in the inner loops because the original CPU/GPU implementation was just a generic library code versus carefully-optimized special-purpose code for the FPGA. It does make business sense, especially with regards to time-to-market; but I wouldn't call that science, even if published in the ostensibly scientific journal.

Do you recall where you've seen those comparisons?

I suppose the release of Keccak first is purely for showing proof of concept? Because I don’t see it making more $12 per card.

Just wondering why you don't develop a new algo..you seem to have a handle on what is needed..its people like you that are needed to move this forward..

I deal with a lot of complex, large FFTs on CPUs, GPUs, and FPGAs. The “only using single precision” is unfortunately true of every vendor - GPU and FPGA. Marketing wants to use the big number - and frankly so do most real world users now. Modern GPUs are horrible at double precision. It is a sad fate. Your comparison also compares a modern Stratix 10 (10 TFLOPs) to the previous generation Ultrascale (not Ultrascale+) with slower fabric and significantly fewer DSP blocks compared to the VCU1525 (XCVU9P-L2FSGD2104E) everyone has been talking about here.

Compared to even modern weak DP GPUs any normal priced CPU is horrible at double precision. A modern GPU runs circles around the Complex FFTs using double precision vs a CPU. Both become quickly memory bound. The FPGA performance is usually on par or slightly better for the double precision, but the benefits in the rest of the calculation are much better. I think you’ll be hard pressed to build a hashing algorithm that is entirely Floating Point like a synthetic benchmark.

The only place FPGAs really fall down is upfront cost.

I’m still a bit confused by why you think sqrt/reciprocal, and the transidentals are so difficult for FPGA, or that’s they are magically free on GPUs/CPUs. On at least AMD GPUs these are macro-ops that take (100s of clockcycles) EDIT: searching for my reference on this, I see these ops are quarter rate. May have been thinking of division) . On the FPGA you can devote a lot of logic to lowering the latency on these functions, or you can pipeline them nice and long with very high throughput to match what you need for the algorithms in question. You have none of that flexibility on the GPU. What you do have is a tremendous amount of power and overhead in instruction fetching, scheduling, branching, caching, etc. to a limited set of ports to implement the opcodes for each GCN/CUDA core.

well Denarius (DNR) created the Tribus algo. Seems smart to do Tribus first seeing DNR is soon to be the fastest most secure crypto that no ones knows about yet. Mining DNR with FGPA miners should rocket DNR's price.

Cant wait

I believe the thinking is:

faster hash rate = more coins = more profitability -> drives more miners -> more interest => more speculation on the coin

There are some iffy transistions but that seems to be the 'pump dump' penny-crypto way

Really not meaning to offend anyone, this has been a very interesting and entertaining thread, even inspiring all around in a way that leads to substantially more decentralization.  However, as an energy industry professional I must say that a kw/h and a kWh is substantially exactly the same thing.  Not sure what you guys are onto here... a kW is a unit of energy, it is a 1,000 Watts.  Watts are convertible to Joules or Therms or any other unit of energy.  And a kW/h is the number of kiloWatts consumed in an hour, as is a kWh, the number of kiloWatts consumed in an hour.  a kW is a measurement of power, and a kWh is a volumetric measurement of energy.  you can use 100kW in one hour is 100kWh or you can use 50kW for 30 minutes and 150kW for 30 minutes and it will also be 100kWh.

How is the original post confusing or misleading again?  No, the OP hasn't offered "evidence" of his experimentation other than several photos and videos, but a lot of people that seem to know the art well are discussing the possibilities in a meaningful way, which makes the claims relatively speaking, plausible.  And considering how FPGA's have been used for ages to mine similar algorithms, and considering how the FPGA's currently available OTS are dramatically larger and more powerful than those original silicon used to mine BTC... it all makes perfect sense.

Why not just wait until May 30th or whatever and let him release his work to the various people that are willing to try it out and have hardware?  And if none of it materializes, the difference between a kW/h and a kWh is totally irrelevant.  And if it does, that is really neat too.

Just saying... of course seeing is believing, but not really any reason not too here.  If Bittware or anyone else is trying to just unload a bunch of hardware, do they really need to do it in a bitcoin talk forum? I think there is more going on out in the world than what people are making out here.  Healthy skepticism sure... but look at the size of this thread!  People realize this is a really important topic, there is a reason for it.  Centralization of hashing power and ASIC's in general are beginning to threaten the security of crypto software... the very thing it was meant to solve.  Not good...

You can convert Joule to Wh, it is indeed a measurement of amount of energy.
Watt's cannot be converted to Joules directly as much as velocity cannot directly be converted to distance without the input of other parameters to build an equation.


A kWh is not the number of kiloWatts consumed in an hour, same difference as you can't say that 'distance is the is the number of meters travelled in an hour'.
In the analogy with distance, velocity & time; kW/h would be acceleration (how much the energy flow changes per unit of time), kW would be velocity (amount of energy flow), kWh would be distance (how much energy has flown)

You can not 'use' xx kW in one hour, you can use xx kWh in one hour though. kW/h is simply the wrong unit to use. But as an energy industry professional you of course know better

This is not really on-topic though but it is simply confusing to people who have little knowledge about this. It might be that most understand that it's an error but using the wrong units on purpose because everyone understands is not productive. I will not go deeper into this as we might as well be discussing whether the earth is flat or not.


Let's see how it turns out, I do think though that they see what profits NVidia and AMD have gotten and want a piece of the pie. A healthy dose of skepticism is never a bad idea.
Possible profits remain to be shown but I think the claims and motivation behind the project are not unreasonable, time will tell.


Curious to see first results!

I own one VCU1525's, I sent the OP a PM.


I'll help confirm the feasibility of this.