Quote from: lunobird on June 08, 2018, 03:13:14 PM

Quote from: MagicSmoker on June 08, 2018, 02:46:55 PM

Quote from: lunobird on June 08, 2018, 12:09:43 PM

...
The leap from gpu too Asic is 100x plus hash rate increase so easier to do 51 attack.

That was true for CryptoNight, but not for Ethash (E3 is 180MH/s at 800W) or Equihash (Z9 is 10kSols/s at 300W). Respectively, that makes the E3 about as powerful as a typical 6x RX 570 GPU rig but drawing much less power and costing less as well, while the Z9 is about 2x more powerful than a 6x GTX 1080ti rig while also drawing much less power and costing less. In other words, not even close to 100x more powerful.

Quote from: lunobird on June 08, 2018, 12:09:43 PM

The leap from Asic to better asic, is 2x at best due to the limitations of chip size nm and power efficiency vs high equipment cost to manufacture a better Asic over previous generations. This is not enough to do a 51 attack on whole network.

I literally just pointed out 2 examples where the 2nd gen ASIC was 5x to 10x faster than the previous gen and yet you still tried to claim only a 2x improvement is possible?

Right, okay. I think I'm done here.

Again another poor argument that failed to get the point and failed to put it all together.

You are correct with Ethash. And its b/c its relying on the limitations of memory hardware tech Moor's law, But its only a matter of time before a 51 percent attack will happen.

yes Z9 wasn't 100x more powerful like that X3, it was just only 10x more powerful. Wait only 10X?!?! You kidding me. 10X is enough to do a 51 percent attack especially if the hardware only cost $2000 compared to a gpu rig will need 10k cash to have 10k sols and generate over 2510 watt usage vs the z9mini 300watts!, and thats exactly what happen with a big enough farm that got early access to bitmain Z9 and ate those gpus hashrate up. Again proving my point that its better to have specialized hardware over consumer grade cards meant for games to secure billions of dollars.

You are wrong about your asic example being 5x-10x faster than previous generations. S7 was rated at 4.86 TH and S9 rated at 12.5 TH, That's pretty much 3x faster, they both use around same watts but at the time the S7 cost $2000 initially. S7 to S9 is an impressive leap but no where near able to do a 51 percent attack on a whole network especially at the starting price point. Thats why you don't hear about a 51 percent attack on bitcoin.

Thanks you for further clarifying my point. The closer the hardware is at peak efficiency with no competitive product the safer your blockchain will be.

Thinking video game cards meant for games, is the way to go in securing a network worth billiions of dollars is just foolish and backward thinking. A 51 percent attack will happen and millions will be lost and ultimately will not last 10 years with this type of thinking.

Doesn't matter who produces whatever efficiency ASIC/GPU. A giant corp/individual can still rent out hashpower and attack a network. The blockchain tech matters, Algo and code matters, decentralization matters.

Going by the logic if only ASICs were prevalent and help secured network here's an example. lets assume we have 1st Gen ASICS $500 now and many can buy this and help with decentralization but later Gen 2 ASICS are $10k 10x faster and 20x efficient and only ones available in market; this is the trend going on when ASIC market is mfg by one single entity. Who can afford those ASICs? Are you suggesting people will be willing to shell out $10k and buy 1 ASIC just for the cause?

The same giant corp who mines with GPU farm can get X number of ASICs and initiate a network attack. They are limited only by power requirements, infrastructure, laws (if applicable), maintenance costs.