Quote from: micryon on March 05, 2014, 09:35:49 AM

Quote from: pastet89 on March 05, 2014, 09:19:11 AM

Quote from: madbit1000 on March 04, 2014, 08:36:55 PM

Quote from: bigjme on March 04, 2014, 08:34:44 PM

Dark coin uses keccak along side other similar algorithms

Is that Asic resistant. if it is i would suggest we use that.

Listen, people. NOTHING is ASIC resiatant in the long term. When LTC was created, using a scrypt algo, it was meant to be ASIC resistant. However, technologies go further with high speeds and 2-3 years later we do have ASICs for scrypt.
Understand that whatever algo to be used, soon or later the asics for it will come out. One may take longer, but soon or late, ASICs (or even new type of hardware) will be created for anything. Yes, these algos are a short term solution. However, in the long term this is an everlasting war with two parties.

Ultimately if there's a market for it.. ASICs will be superior in whatever app they intend to service... That's what their names imply... Application Specific .. integrated circuits....

However, the idea here is to make it not worth their while to design, develop and fabricate the ASIC.. asic development isn't like writing software where you get 13 guys, and sell a company to FB for a billion.. it takes real capital to fund R&D, buy wafers and fabricate.. lot of risk.. risk wafers, imperfections.. then debug, ECOs and respins... it's a big upfront capital investment.

The nfactor code isn't just a set algorithm.. well it is, and it isn't... it changes 1 parameter on a scheduled basis by posix time to be "adaptive"... okay its pretty lame adaption change just 1 parameter... but hey it does increase the memory b/w dramatically.. 1024 -> 2048.. and on power of 2s... which *might* make it infeasible for ASICs to continually tape out to adjust for it.. and if ASICs should choose to also become adaptive.. well.. they just become generalized processors anyhow, which would yield little to no advantage against GPUs, since they are no longer really application specific... if you support 100+ applications.. well then you're a application processor..

Sure... if at some point nfactor=X has a 100B market cap due to coins.. and its just worth people's time and energy to keep developing and releasing asics that do just that... then yeah, this premise does fall apart.. But the lead time and development time is long for integrated circuits. Idea here is to not make it worth their while... It may fail.. but in my estimation it ultimately it all depends on where crypto market cap in total are going...

Thinking about this.. there's probably room for a more innovative solution to this which not only changes nfactor.. but changes the algorithm based on blockHeight... to align with the direction of where GPU/CPU computing is going.. like for blockHeight > 10000 or something.. swtiches to prime sieving.. (primecoin)... predicting in that time GPUs would catch CPUs in that race in a few years.. and leave ASICs far behind... but anyways who knows, the future is the future.. anything like this is hard to predict. If you could you would be making millions speculating the semiconductor market

this is going to be proven wrong real soon, professional diy chip fab is coming
once i get my manufacturing capabilities online, its actually a project i plan on tackling.
also gpu are stuck to their gaming purpose, an asic can be designed to be adaptive proof with more memory then you would ever find on a gpu, heck you could build an asic that can also interface with pcie and make a hybrid. if your coin doesn't make it to the asic point, its never gona be worth much.