Will X11 save us from the ASIC vultures?

[flipme]

I see a wave of coin re-designs coming. Do or die.
I would love to see the ASIC freaks left holding the bag with all the flappy crap inside.

Quote taken from here: http://www.reddit.com/r/Hirocoin/comments/215lkd/changing_landscape_for_scrypt/

The GPU mining landscape is about to change drastically and many have not woken up to this fact. KNc recently took $2m in pre-orders for their 100MH Scrypt ASIC. These may start coming on to the network in Q2/Q3 of this year. There is no doubt that these will be turned on to many of the small Scrypt coins out there to strip them, sell the coins and then leave. This is going to leave many small coins unusable, many have less than 1MH. People do not seem to be aware of the problems this is going to cause and perhaps were not around for the SHA-256 ASICs turning up. This left many small SHA-256 at the time in a state where they had to hard fork to resolve the issues and others just disappeared. It is possible to move hashing algos in those coins.
Some may remember that Terracoin was a target for the SHA-256 ASICs, with the very large hashing power Terracoin was left no alternative but to hard fork its way out of trouble. For all that they tried to dodge the ASICs these devices were used again to run some very devastating attacks on Terracoin forcing them to amend their code again and hard fork. The additional hard fork was an exceptional case as their attempts to introduce smoother difficulty adjust left flaws in their code.
There other hash solutions out there the best known of these being Scrypt-Jane and Adaptive-N which for the most part are the same. They both use a variable N factor which is locked to 210 in Scrypt. This is the memory requirement, 210 works out to 1024 bytes. Vertcoin a very popular Adaptive-N coin currently uses an N factor of 211 which works out to 2048 bytes, every time the N factor goes up the performance of miners drops by half. Vertcoin currently has half the performance of Scrypt and has a maximum N of 232 which works out to 4GB, most graphic cards cannot even do work at that level and with the current hash rate the difficulty of Vertcoin could not go low enough to support that level of N. Scrypt-Jane and Adaptive-N may be ASIC hostile but they are also GPU hostile.
There is an alternative that has been largly over looked which was created by Even Duffield. This is the chap who created DarkCoin which looks to include the DarkSend feature that will allow anonymous transactions. X11 uses 11 well known and high performing hashing solutions chained together to generate the hashes required to generate new blocks. Since it uses 11 different hashes it is complex and unlikely to see a ASIC for it any time soon. The good news for miners is that they can use X11 right now to avoid multipools and start gathering coins that are going to become very important when the Scrypt ASICs hit. Mining X11 gives 3-4 times the hash power of Scrypt, uses less energy and generates less heat. This is the solution that people have been looking for.
There is also Hirocoin that has launched recently and adopted X11 as its hashing solution. Hirocoin does not have the DarkSend feature but has been meticulously coded and fully featured on launch including DNS seed and binaries for different platforms. DarkCoin uses an inverse difficulty reward which means that the higher the difficulty the lower the reward. More miners means less coins, Hirocoin is more conventional in that it has 400 coins a block and is likely to be the first choice for miners.
For all that there is the X11 alternative for GPU mining there does not seem to be any Scrypt coins who are seriously talking about changing their hashing algorithms. They may soon have their hand forced in the same way that Terracoin had no choice but to fork. However many small Scrypt coins do not seem to have an active developer in site, just communities of users who may soon be left without a working coin.

[Infectiphibian]

I'm completely for X11. My electricity reduction has cut my power bill in half, my hardware runs cool without extra fans, and I can use my rigs for more than just mining.

I can see many miners not liking it though because they will have to learn configuring new mining software. Not that it is hard, but reading through some threads, some people sure seem to make it a challenge.

[Amph]

I'm completely for X11. My electricity reduction has cut my power bill in half, my hardware runs cool without extra fans, and I can use my rigs for more than just mining.

I can see many miners not liking it though because they will have to learn configuring new mining software. Not that it is hard, but reading through some threads, some people sure seem to make it a challenge.

what? you just leave the settings like they were with scrypt, lol

it's anti-retarded

[tromp]

Which of the 11 hash functions in X11 do you suppose is ASIC resistant?

If you can make an ASIC for each of the 11 hash functions, then you can trivially make an ASIC to do all of them at the same time...

[Infectiphibian]

I'm completely for X11. My electricity reduction has cut my power bill in half, my hardware runs cool without extra fans, and I can use my rigs for more than just mining.

I can see many miners not liking it though because they will have to learn configuring new mining software. Not that it is hard, but reading through some threads, some people sure seem to make it a challenge.

what? you just leave the settings like they were with scrypt, lol

it's anti-retarded

I'm completely for X11. My electricity reduction has cut my power bill in half, my hardware runs cool without extra fans, and I can use my rigs for more than just mining.

I can see many miners not liking it though because they will have to learn configuring new mining software. Not that it is hard, but reading through some threads, some people sure seem to make it a challenge.

what? you just leave the settings like they were with scrypt, lol

it's anti-retarded

Slightly anti-retarded. One of my rigs ran perfect with a basic batch file, without setting thread concurrency or the amount of shaders. Another rig had to have the thread concurrency set, but not the shaders. The second rig also required a the thread concurrency slightly lower than scrypt. Just a little no-brainer tweeking.

[Infectiphibian]

Which of the 11 hash functions in X11 do you suppose is ASIC resistant?

If you can make an ASIC for each of the 11 hash functions, then you can trivially make an ASIC to do all of them at the same time...

ASIC resistant is not really the right term to use. X11 doesn't have enough users to develop an ASIC and sell it profitably. If X11 catches on, then ASICs will be a few years down the road.

[tromp]

Which of the 11 hash functions in X11 do you suppose is ASIC resistant?

If you can make an ASIC for each of the 11 hash functions, then you can trivially make an ASIC to do all of them at the same time...

X11 is also easily targeted by FPGAs, which will significantly outperform GPUs.

Let's instead consider the resistance of my Cuckoo Cycle proof-of-work (https://github.com/tromp/cuckoo):

An ASIC for something like cuckoo728 (graph size 7x2^{28} ), which requires 7GB for a single instance, could not be self contained. It would be like a vastly simplified CPU with perhaps 1024 cores, that runs 32 instances of cuckoo728 at once, each one running with 32 threads. It would have to be connected to 32x7GB, or nearly 256GB of memory, which would totally dominate the cost. Intel is already planning to introduce cpus with hundreds of cores, so such an ASIC would not offer enough performance advantage to justify its development cost.

Similarly, no current GPU card has enough memory to run cuckoo728, and future ones with twice the memory of the current max (6GB) would only be able to run a single instance (the algorithm doesn't allow a time-memory trade-of). Even that may not be competitive with a CPU, which has much better memory latencies. The vastly superior memory bandwidth of a GPU is mostly wasted on Cuckoo Cycle, which makes completely random accesses. Another problem with GPUs is that they cannot deal well with data dependent code flow, which features prominently in Cuckoo Cycle.

[flipme]

That sounds great! I'll give it a try with Darkcoin and sph-sgminer.
https://bitcointalk.org/index.php?topic=475795.0

It'll take some time to make ASICs for all that, or arrays thereof. Just add another algo and they're gone again.
And if the fabulous 2 mil $ deal of KNC turns out to be for the trash can, I won't see a lot of people risking such developments.

Spread the word.  

[foodies123]

I believe no one single algo will ever by asic resistant, I think that multiple independent algos is a much better way because it gives asics a slice of the pie thus there is incentive to expand to other algos.

[flipme]

Which of the 11 hash functions in X11 do you suppose is ASIC resistant?

If you can make an ASIC for each of the 11 hash functions, then you can trivially make an ASIC to do all of them at the same time...

X11 is also easily targeted by FPGAs, which will significantly outperform GPUs.

Let's instead consider the resistance of my Cuckoo Cycle proof-of-work (https://github.com/tromp/cuckoo):

An ASIC for something like cuckoo728 (graph size 7x2{28} ), which requires 7GB for a single instance, could not be self contained. It would be like a vastly simplified CPU with perhaps 1024 cores, that runs 32 instances of cuckoo728 at once, each one running with 32 threads. It would have to be connected to 32x7GB, or nearly 256GB of memory, which would totally dominate the cost. Intel is already planning to introduce cpus with hundreds of cores, so such an ASIC would not offer enough performance advantage to justify its development cost.

Similarly, no current GPU card has enough memory to run cuckoo728, and future ones with twice the memory of the current max (6GB) would only be able to run a single instance (the algorithm doesn't allow a time-memory trade-of). Even that may not be competitive with a CPU, which has much better memory latencies. The vastly superior memory bandwidth of a GPU is mostly wasted on Cuckoo Cycle, which makes completely random accesses. Another problem with GPUs is that they cannot deal well with data dependent code flow, which features prominently in Cuckoo Cycle.

I could only find harsh words for that ugly idea, but you picked the right name for it.
 
We are the miners. We are mostly on GPUs. Thats the status. The industry standard.
Its up to us to defend that position and fight for our investments, or just give it up again to a few robots.
It's comparable to robot work, killing the workforce in a factory.

Software could be changed easy.
An in-app update method is missing anyway.

[flipme]

The talk about changing algos won't go away anymore.
I'm doing Darkcoin since yesterday and I already enjoy the silence and lack of heat mining X11.
My power bill will be cut in half. The chart for the coin looks amazing.

A lot of the hassle regarding fundamental changes in the software would be taken away, if there were an in-app update method contained in the wallets.

[reRaise]

I see a wave of coin re-designs coming. Do or die.
I would love to see the ASIC freaks left holding the bag with all the flappy crap inside.

Quote taken from here: http://www.reddit.com/r/Hirocoin/comments/215lkd/changing_landscape_for_scrypt/

The GPU mining landscape is about to change drastically and many have not woken up to this fact. KNc recently took $2m in pre-orders for their 100MH Scrypt ASIC. These may start coming on to the network in Q2/Q3 of this year. There is no doubt that these will be turned on to many of the small Scrypt coins out there to strip them, sell the coins and then leave. This is going to leave many small coins unusable, many have less than 1MH. People do not seem to be aware of the problems this is going to cause and perhaps were not around for the SHA-256 ASICs turning up. This left many small SHA-256 at the time in a state where they had to hard fork to resolve the issues and others just disappeared. It is possible to move hashing algos in those coins.
Some may remember that Terracoin was a target for the SHA-256 ASICs, with the very large hashing power Terracoin was left no alternative but to hard fork its way out of trouble. For all that they tried to dodge the ASICs these devices were used again to run some very devastating attacks on Terracoin forcing them to amend their code again and hard fork. The additional hard fork was an exceptional case as their attempts to introduce smoother difficulty adjust left flaws in their code.
There other hash solutions out there the best known of these being Scrypt-Jane and Adaptive-N which for the most part are the same. They both use a variable N factor which is locked to 210 in Scrypt. This is the memory requirement, 210 works out to 1024 bytes. Vertcoin a very popular Adaptive-N coin currently uses an N factor of 211 which works out to 2048 bytes, every time the N factor goes up the performance of miners drops by half. Vertcoin currently has half the performance of Scrypt and has a maximum N of 232 which works out to 4GB, most graphic cards cannot even do work at that level and with the current hash rate the difficulty of Vertcoin could not go low enough to support that level of N. Scrypt-Jane and Adaptive-N may be ASIC hostile but they are also GPU hostile.
There is an alternative that has been largly over looked which was created by Even Duffield. This is the chap who created DarkCoin which looks to include the DarkSend feature that will allow anonymous transactions. X11 uses 11 well known and high performing hashing solutions chained together to generate the hashes required to generate new blocks. Since it uses 11 different hashes it is complex and unlikely to see a ASIC for it any time soon. The good news for miners is that they can use X11 right now to avoid multipools and start gathering coins that are going to become very important when the Scrypt ASICs hit. Mining X11 gives 3-4 times the hash power of Scrypt, uses less energy and generates less heat. This is the solution that people have been looking for.
There is also Hirocoin that has launched recently and adopted X11 as its hashing solution. Hirocoin does not have the DarkSend feature but has been meticulously coded and fully featured on launch including DNS seed and binaries for different platforms. DarkCoin uses an inverse difficulty reward which means that the higher the difficulty the lower the reward. More miners means less coins, Hirocoin is more conventional in that it has 400 coins a block and is likely to be the first choice for miners.
For all that there is the X11 alternative for GPU mining there does not seem to be any Scrypt coins who are seriously talking about changing their hashing algorithms. They may soon have their hand forced in the same way that Terracoin had no choice but to fork. However many small Scrypt coins do not seem to have an active developer in site, just communities of users who may soon be left without a working coin.

I know you want to advertise Darkcoin, but you need to keep it real. More hashing functions doesn't mean more secure. Quark is still the king of multiple hashing imo.

Darkcoin in terms of hashing functions it is a combination of Quark's 6 hashing functions and Qubit-coin's 5 hashing functions, so in total Darkcoin has 11. Five of Quark's hashing functions (Blake, Grøstl, JH, Keccak and Skein) were the finalists of the NIST competition https://en.wikipedia.org/wiki/NIST_hash_function_competition

Quark is using 9 rounds of hashing: while using 6 rounds from Blake, Blue Midnight Wish, Grøstl, JH, Keccak and Skein it adds 3 more rounds of hashing randomly: so the computer doesn't know whether it will be Keccak or Grøstl or Blake. And that's one of the uniqiue beauties of Quark. Unfortunately, Darkcoin or Qubitcoin don't do that: the computer remains certain about which hashing function will be used.

Another advantage of Quark might be, Darkcoin's block generation time is 2.5minutes (150seconds), while Quark is 30 seconds. Which means that Quarks algorithm with an element of randomness (unpredictability) will have to be cracked in 30 seconds to create a double-spend fork, while for Darkcoin this window of opportunity for the attacker is 5 timeslonger: 150 seconds with no element of randomness. If you take all these factors into account, Quark is still the most secure - it's not only about the number of hashing functions."

[tacotime]

X11 is a great unix Windows manager.

The ASIC resistance of the hash function, not so much.  Play around with an FPGA and you'll see you're looking at about the same amount of resistance as SHA256d.  You increase the number of circuits required about 10-fold, but "X11" is also more than ten fold slower than SHA256d on GPUs.

Probably the most ASIC resistant thing you can do is maintain a low market cap.

[flipme]

It would sure give an advantage on the timeline of events.
Better than just sitting scared and doing nothing.

[El Dude]

X11 is a great unix Windows manager.

The ASIC resistance of the hash function, not so much.  Play around with an FPGA and you'll see you're looking at about the same amount of resistance as SHA256d.  You increase the number of circuits required about 10-fold, but "X11" is also more than ten fold slower than SHA256d on GPUs.

Probably the most ASIC resistant thing you can do is maintain a low market cap.

exactly every good coin gets ASICs one way or another.

[AlterEgoCoin]

X11 is a great unix Windows manager.

The ASIC resistance of the hash function, not so much.  Play around with an FPGA and you'll see you're looking at about the same amount of resistance as SHA256d.  You increase the number of circuits required about 10-fold, but "X11" is also more than ten fold slower than SHA256d on GPUs.

Probably the most ASIC resistant thing you can do is maintain a low market cap.

We at the AlterEgoCoin[AEC] development team think that we have found a cure for the survival of GPU mining after FPGA`s starts hitting the market in high numbers.
To have multiple algos rotating in a pre determined cycle will give GPU`s full mining capability and if someone wants to point an ASIC or a FPGA they will only be able to participate in 1 or 2 out of X blocks making it not worthwhile to mine for them, they would also need to code their own miner.
If ASIC`s or FPGA`s would become a problem for the coin we would simply hardfork, update the miner and add more algos to the cycle.
Check us out at:
https://bitcointalk.org/index.php?topic=528925.msg5889698#msg5889698

[tacotime]

X11 is a great unix Windows manager.

The ASIC resistance of the hash function, not so much.  Play around with an FPGA and you'll see you're looking at about the same amount of resistance as SHA256d.  You increase the number of circuits required about 10-fold, but "X11" is also more than ten fold slower than SHA256d on GPUs.

Probably the most ASIC resistant thing you can do is maintain a low market cap.

We at the AlterEgoCoin[AEC] development team think that we have found a cure for the survival of GPU mining after FPGA`s starts hitting the market in high numbers.
To have multiple algos rotating in a pre determined cycle will give GPU`s full mining capability and if someone wants to point an ASIC or a FPGA they will only be able to participate in 1 or 2 out of X blocks making it not worthwhile to mine for them, they would also need to code their own miner.
If ASIC`s or FPGA`s would become a problem for the coin we would simply hardfork, update the miner and add more algos to the cycle.
Check us out at:
https://bitcointalk.org/index.php?topic=528925.msg5889698#msg5889698

That was my solution from more than a year ago, and it doesn't work.  All you need in that case is an ASIC with an instruction cache, which is not circuit expensive.

VertCoin has the right idea (also the bluntest idea), which is to just swap in different memory hard algorithms by a hardfork on a regular basis and not announce what they are ahead of time.

[flipme]

Scrypt-N needs more power, more memory.
I live in Spain, with 40 C outside in the summer. Scrypt-N is no real option.
I was all for it, but I think X11 is better.
And something needs to be done NOW, beside talking and writing white papers.

[IloveAnonCoin]

X11 and SHA3 ( Keccak ) are not ASIC resistant at all, according to NIST (National Institute of Standards and Technology), if you want to become SHA3 candidate, you need to be able to create by ASIC. AND every algorithms in X11 used to be SHA3 candidate until Keccak win the competition and become SHA3.


Here is the paper : http://csrc.nist.gov/groups/ST/hash/sha-3/Round2/Aug2010/documents/papers/SCHAUMONT_SHA3.pdf

[flipme]

Looks more and more like it.
Almost every new coin is X11.

Darkcoin rules.
Scrypt PoW going down the drain.
LTC is tanking.