Topic: 🔴 SyncAi - distributed Artificial Intelligence System - page 10. (Read 19013 times)
Researchers want to teach computers to learn like humans
https://www.sciencedaily.com/releases/2018/03/180305093005.htm
I'm going to build a new AI test rig based on the latest AMD X470 platform.
I have a question on PCIe bandwidth. It seems most, if not all, of these motherboards only provide one x16 slot, then one or 2 x8 slots.
My experience with 1080 ti and compute applications like BOINC, Mediccoin, folding etc. is that x16 makes a big difference.
Do you have any test data for the type of AI applications you are envisioning ?
How will less than x16 bandwidth affect a 1080 ti or higher ?
I have a question on PCIe bandwidth. It seems most, if not all, of these motherboards only provide one x16 slot, then one or 2 x8 slots.
My experience with 1080 ti and compute applications like BOINC, Mediccoin, folding etc. is that x16 makes a big difference.
Do you have any test data for the type of AI applications you are envisioning ?
How will less than x16 bandwidth affect a 1080 ti or higher ?
I do not have 1080, but I think that x16 or x8 does not greatly affect these calculations. test data is a normal photo) The speed is more affected by the speed of the video card and the algorithm used.
I'm going to build a new AI test rig based on the latest AMD X470 platform.
I have a question on PCIe bandwidth. It seems most, if not all, of these motherboards only provide one x16 slot, then one or 2 x8 slots.
My experience with 1080 ti and compute applications like BOINC, Mediccoin, folding etc. is that x16 makes a big difference.
Do you have any test data for the type of AI applications you are envisioning ?
How will less than x16 bandwidth affect a 1080 ti or higher ?
I have a question on PCIe bandwidth. It seems most, if not all, of these motherboards only provide one x16 slot, then one or 2 x8 slots.
My experience with 1080 ti and compute applications like BOINC, Mediccoin, folding etc. is that x16 makes a big difference.
Do you have any test data for the type of AI applications you are envisioning ?
How will less than x16 bandwidth affect a 1080 ti or higher ?
Hi
Sorry, Im very busy on the main job (two new blockchains with coins)
Soon i will show ai-storage and ai-connectir. then the beta of the ai-miner
PS ai-miner is useless without marketplace and clients but yes, - it will be possible to test the miner (I will add first a simple photo recognition)
Sorry, Im very busy on the main job (two new blockchains with coins)
Soon i will show ai-storage and ai-connectir. then the beta of the ai-miner
PS ai-miner is useless without marketplace and clients but yes, - it will be possible to test the miner (I will add first a simple photo recognition)
The ultimate question is if there is anything special with life compared to dead matter.
If not, the AI systems will exceed our capabilities eventually.
However, the human race will have to adapt. We can't remain hairless monkeys forever so we need to speed up evolution artificially. Genetic manipulation can make us much smarter on average. There are quite a lot of people with IQ greater than 130. No reason the average must stay at 100. Bring them all up to 150. They are already attempting this in China at industrial scale. They are sampling DNA from the smartest people on earth to map, understand and use it as a service for future parents wanting the best for their kids.
Then we can do cybernetic integration. Internet is a primitive pre-cursor to the future human-Borg collective where we have cybernetic implants connecting our brains directly to AI sub processors, quantum computers, machines and other human beings. The future super intelligent homo sapiens V2.0 uses the AI as augmentation of himself for things unimaginable to us.
Just imagine the number of scam-coin ICOs we will see on the road to all this
Btw. When can we expect some BETA software to test your distributed AI computing with ?
If not, the AI systems will exceed our capabilities eventually.
However, the human race will have to adapt. We can't remain hairless monkeys forever so we need to speed up evolution artificially. Genetic manipulation can make us much smarter on average. There are quite a lot of people with IQ greater than 130. No reason the average must stay at 100. Bring them all up to 150. They are already attempting this in China at industrial scale. They are sampling DNA from the smartest people on earth to map, understand and use it as a service for future parents wanting the best for their kids.
Then we can do cybernetic integration. Internet is a primitive pre-cursor to the future human-Borg collective where we have cybernetic implants connecting our brains directly to AI sub processors, quantum computers, machines and other human beings. The future super intelligent homo sapiens V2.0 uses the AI as augmentation of himself for things unimaginable to us.
Just imagine the number of scam-coin ICOs we will see on the road to all this
Btw. When can we expect some BETA software to test your distributed AI computing with ?
Will robots take our children's jobs?
Our kids will have to forget their dream job if it's not a plumber. Artificial intelligence will render half of today's professions obsolete in 20 years, but not all trades are heading for the chop
https://www.independent.co.uk/news/long_reads/artificial-intelligence-take-our-jobs-robot-economy-universal-basic-income-a8110031.html
As example
can be connected to SyncAi
https://www.sciencedaily.com/releases/2018/04/180416142443.htm
Researchers have developed an artificial intelligence and machine learning technique that produces a visible face image from a thermal image of a person's face captured in low-light or nighttime conditions. This development could lead to enhanced real-time biometrics and post-mission forensic analysis for covert nighttime operations.
can be connected to SyncAi
https://www.sciencedaily.com/releases/2018/04/180416142443.htm
Researchers have developed an artificial intelligence and machine learning technique that produces a visible face image from a thermal image of a person's face captured in low-light or nighttime conditions. This development could lead to enhanced real-time biometrics and post-mission forensic analysis for covert nighttime operations.
I think this concept has lots of potential when gpu mining dies off. Who do you see using the service the AI provides? I'm a bit familiar with GPU clusters for deep learning. Would you want to know the set up of the mining rigs on the system so that they can best be utilized?
I think that it will be useful to everyone. What kind of neural models do you train? How much time is training?
I think this concept has lots of potential when gpu mining dies off. Who do you see using the service the AI provides? I'm a bit familiar with GPU clusters for deep learning. Would you want to know the set up of the mining rigs on the system so that they can best be utilized?
Your idea and product super.conduct ICO As soon as possible.assemble hard cap in a minute) the future of AI.you have chosen the right direction.thank you for what you have.
well, I will write a benchmark program for you. run it, and it will calculate how much you can earn on your equipment.
For example there is a task for video processing (recognition of criminal actions) from old low quality 300 000 web cameras in the city but one card 1080 can process only 6 cameras
for comparison:
Vectordash Hardware Requirements:
For each GPU you plan on hosting, your machine should have at least:
- a PCIe 3.0 x8 or x16 slot (no x1 risers)
- 1 CPU Core from one of the following CPUs (or equivalent):
- 50+ GB Storage per GPU
- Internet upload & download speeds of at least 15 Mbps
Any suggestions for mainboards that can provide multiple PCIe 3.0 x8 / x16 GPU slots?
Vectordash Hardware Requirements:
For each GPU you plan on hosting, your machine should have at least:
- a PCIe 3.0 x8 or x16 slot (no x1 risers)
- 1 CPU Core from one of the following CPUs (or equivalent):
- Intel i5 or above
Intel Xeon
AMD Ryzen or AMD EPYC
- 50+ GB Storage per GPU
- Internet upload & download speeds of at least 15 Mbps
Any suggestions for mainboards that can provide multiple PCIe 3.0 x8 / x16 GPU slots?
Is there enough work to go around? Like, for people with 40 and 50 cards, if they put everything on this, is that too much? It seems insane that if I sent over 40 1070s/1070TIs I could make that much ether. I mean, obviously I would do it if that's the case, I just wonder if that's sustainable.
Wait, what??
$40 (0.076 ETH) a day with one 1060??
$40 (0.076 ETH) a day with one 1060??
Yes
Its not classic mining. it's a ether-paid job.
Wait, what??
$40 (0.076 ETH) a day with one 1060??
$40 (0.076 ETH) a day with one 1060??
With what hardware ? What kind of requirements ?
Mainly Nvidia Cuda, start from 1060, but it work also on CPU
With what hardware ? What kind of requirements ?
Is there any kind of profitability calculator ?
theoretically, at full load it is $ 40 per day
Is there any kind of profitability calculator ?
Jump to: