Topic: please create NVIDIA-only Coin (Read 1914 times)
this would not be possible.
I would much rather see a coin that can only be mined on an ARM processor, like the processors in every single smart phone and tablet on the market.
This would give me a reason to buy up a TON of old used smart phones on ebay. lol
This would give me a reason to buy up a TON of old used smart phones on ebay. lol
I could see smart phones/tablet thefts going through the roof with this lol. Dont get me wrong I'd love to get my phone and kindle working away! I always think when they are charging they should be earning!
Nvidia right now is completely dead in the water concerning opencl.
Moreover, there is info that the new AMD GPU family "Hawaii" is vastly improved concerning non-gaming related performance.
If maxwell does not change that then crypto's are off nvidias turf for good.
Nvidia is strategicaly lacking, with no APU project, no x86 licence and stuck to old ideas.
It is either gonna die out or get bought by Intel within the next five years.
Moreover, there is info that the new AMD GPU family "Hawaii" is vastly improved concerning non-gaming related performance.
If maxwell does not change that then crypto's are off nvidias turf for good.
Nvidia is strategicaly lacking, with no APU project, no x86 licence and stuck to old ideas.
It is either gonna die out or get bought by Intel within the next five years.
Actually Nvidia Tesla cards are market leaders and way beyond AMD for computation. The problem is that they are expensive.
I don't think wasting time on Nvidia way is good idea. AMD's GPGPU is far ahead of Nvidia's.
Check out the AMD's HSA (heterogeneous system architecture), when it's shown up in the market. Nvidia will be more then ever like the toys or just trash.
Check out the AMD's HSA (heterogeneous system architecture), when it's shown up in the market. Nvidia will be more then ever like the toys or just trash.
Quote
2) This is a variant of the above that embraces the enemy instead of fighting it. Implement (1) with reals as cl_double which is supported in OpenCL by NVidia cards but not by AMD cards. Then short AMD stock and go long NVDA for additiona gains.
I'm pretty sure this isn't true, cl_double is called in the example source code bundled with AMD's APP SDK
The difference in IEEE 754 implementation between nVidia and AMD GPUs is that nVidia tends to be congruent with the results that x86/x64 processors spit out, while AMD's are not because the FP precision is less accurate.
But as far as I know if you can do it in the FPUs and it's invertible, you can do it in integer logic too. Whether the FPU operations would hold an advantage in that case, I do not know.
If there is some math operation which output is bit-to-bit same across all NVIDIA hardware but differs in least significant bits from x86 and AMD, such operation can be used for NVIDIA-only coin. Simulating those least significant bits would be possible, but expensive. Perhaps IEEE 754 floating point math is deterministic and the same across different hardware, but in fast math mode NVIDIA and AMD could be different.
Question was just if it's possible there would be a coin in which nvidia can have the advantage. I'm sure there are lots of people in similar position that would be happy with this.
Hashcash based coins? noCoins based on some other proof-of-work? maybe
It wouldn't be possible to create something that can only be mined on NVIDIA hardware, but there could be some computational problems where NVIDIA hardware might have an advantage in (at least initially).
So, to answer your question, it *might* be possible.
The problem is, I don't think anyone putting in enough work to design such a coin (based on a different proof-of-work) would care about something like this. It really doesn't do anything to secure/decentralize the network. True, it might work at first since there are not any mining farms based on NVIDIA, but there are plenty of render/compute farms based on high end NVIDIA hardware (Tesla), so it wouldn't be something that would be advantageous to laptop/single-GPU users in the long run. Personally, I think it's often forgotten that crypto-currencies are not designed for miners, and mining just provides an incentive to expend the electricity needed to decentralize the system.
I did not create this thread to hear Nvidia is bad for mining we all know that already.
My laptop for example is from 2010 and I'm using it for video processing and 3d world games, it works pretty good.
Not having money for miningrigs/asics but still like to do some mining with it.
Question was just if it's possible there would be a coin in which nvidia can have the advantage. I'm sure there are lots of people in similar position that would be happy with this.
At least a few replies from people that seem to have technical knowledge and can stay on topic and the amount of replies says enough, it can be succesfull.
My laptop for example is from 2010 and I'm using it for video processing and 3d world games, it works pretty good.
Not having money for miningrigs/asics but still like to do some mining with it.
Question was just if it's possible there would be a coin in which nvidia can have the advantage. I'm sure there are lots of people in similar position that would be happy with this.
At least a few replies from people that seem to have technical knowledge and can stay on topic and the amount of replies says enough, it can be succesfull.
A much more realistical scenario seems to be able to sell your NVIDIA GPU power for bitcoin.
Based on future folding@home and other cloud computing developments, NVidia cards may be an effective way to acquire CureCoin. 
Quote
2) This is a variant of the above that embraces the enemy instead of fighting it. Implement (1) with reals as cl_double which is supported in OpenCL by NVidia cards but not by AMD cards. Then short AMD stock and go long NVDA for additiona gains.
I'm pretty sure this isn't true, cl_double is called in the example source code bundled with AMD's APP SDK
The difference in IEEE 754 implementation between nVidia and AMD GPUs is that nVidia tends to be congruent with the results that x86/x64 processors spit out, while AMD's are not because the FP precision is less accurate.
But as far as I know if you can do it in the FPUs and it's invertible, you can do it in integer logic too. Whether the FPU operations would hold an advantage in that case, I do not know.
mostly dumb replies ..no shocker there
why do most of you spout off like cocky condescending know it all's ?
OP:
I have the cuda sdk and compile and make my won cuda based builds so yeah i do know hat i am doing..
and i have been buying and using both before half these blow hard's were even alive.
this isn't a debate on whether AMD/ATI is better kids.
floating point eh ? hmmm
Nvidia is no more garbage than any other brand.. don't be ignorant little fanboys..
You might as well start waiving your pathetic windows 8 pom pom's cheerleaders
I'm an expert ROOOAAAARRRR
lol fuck off idiots
edit:
and Opecl reversed ? duuuuuuuh
works both ways ..i know a i'm a cracker
point again ?
why do most of you spout off like cocky condescending know it all's ?
OP:
I have the cuda sdk and compile and make my won cuda based builds so yeah i do know hat i am doing..
and i have been buying and using both before half these blow hard's were even alive.
this isn't a debate on whether AMD/ATI is better kids.
floating point eh ? hmmm
Nvidia is no more garbage than any other brand.. don't be ignorant little fanboys..
You might as well start waiving your pathetic windows 8 pom pom's cheerleaders
I'm an expert ROOOAAAARRRR
lol fuck off idiots
edit:
and Opecl reversed ? duuuuuuuh
works both ways ..i know a i'm a cracker
point again ?
Hi again, SolidCoin developers!
After a dinner I came up with another two ideas:
1) Instead of evaluating a total recursive function over a commutative ring of integers you could try a simpler thing. Require evaluating a specific value of primitive recursive function over the field of reals that has fractional dimension. Good starting example:
http://en.wikipedia.org/wiki/Logistic_map
just pick the value of r that is in the chaotic region.
Implement reals as long double that is 80-bit (10-bytes) on the Intel/AMD CPUs. Not all the C++ compilers really support long double, but the logistic map function is so trivial that you can include in-line assembly which will be in the order of 10 lines.
2) This is a variant of the above that embraces the enemy instead of fighting it. Implement (1) with reals as cl_double which is supported in OpenCL by NVidia cards but not by AMD cards. Then short AMD stock and go long NVDA for additiona gains.
Again, good luck!
After a dinner I came up with another two ideas:
1) Instead of evaluating a total recursive function over a commutative ring of integers you could try a simpler thing. Require evaluating a specific value of primitive recursive function over the field of reals that has fractional dimension. Good starting example:
http://en.wikipedia.org/wiki/Logistic_map
just pick the value of r that is in the chaotic region.
Implement reals as long double that is 80-bit (10-bytes) on the Intel/AMD CPUs. Not all the C++ compilers really support long double, but the logistic map function is so trivial that you can include in-line assembly which will be in the order of 10 lines.
2) This is a variant of the above that embraces the enemy instead of fighting it. Implement (1) with reals as cl_double which is supported in OpenCL by NVidia cards but not by AMD cards. Then short AMD stock and go long NVDA for additiona gains.
Again, good luck!
see: http://en.wikipedia.org/wiki/Floating-point
You need invertible logic based on integers or bit vectors in order to perform cryptography, because you can't be losing or corrupting data among operations. For instance, the extensive use of the easily invertible logic operation XOR as an additive cipher.
You need invertible logic based on integers or bit vectors in order to perform cryptography, because you can't be losing or corrupting data among operations. For instance, the extensive use of the easily invertible logic operation XOR as an additive cipher.
As far as I know is Nvidia better in floating point calculations then AMD, so the coin should be based on that.
FLOPS often involve rounding (result truncation based on the mantissa of the data going into the operation), which makes it lossy and non-invertible. If you reduce the operations to reversible ones I think you essentially just get integer operations.
As far as I know is Nvidia better in floating point calculations then AMD, so the coin should be based on that.
FLOPS often involve rounding (result truncation based on the mantissa of the data going into the operation), which makes it lossy and non-invertible. If you reduce the operations to reversible ones I think you essentially just get integer operations.
I would much rather see a coin that can only be mined on an ARM processor, like the processors in every single smart phone and tablet on the market.
This would give me a reason to buy up a TON of old used smart phones on ebay. lol
This would give me a reason to buy up a TON of old used smart phones on ebay. lol
As far as I know is Nvidia better in floating point calculations then AMD, so the coin should be based on that.
Cryptography is dependent on integer op throughtput
AMD pulls way more IOPS
I doubt nVidia can compete in anything crypto-related
AMD pulls way more IOPS
I doubt nVidia can compete in anything crypto-related
I'm not entirely sure how that would work? It may launch with a CUDA-only miner, but that can be reverse-engineered to work with OpenCL.
Jump to: