Any plans for optimized Cuda 7.5/8.0RC binaries for nVidia GPUs with Compute Capability 5.0 or higher? Seems that many improvements and optimizations have been put in since Cuda 5.0.

Plans yes, though Cuda 6 onwards tend to include improvements not really targeted to the kind of processing we do so much. Probably most of the improvements medium term will come from recent contributions from Petri33. Longer term probably looking at trying to leverage some of the ai targeted features not explored in setiathome code yet (longer term prospect).

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

Hey Jason,
Can you expand on this:

Longer term probably looking at trying to leverage some of the ai targeted features not explored in setiathome code yet (longer term prospect).

Cheers,
Rob :-)

Hey Jason,
Can you expand on this:

Longer term probably looking at trying to leverage some of the ai targeted features not explored in setiathome code yet (longer term prospect).

Cheers,
Rob :-)

Well basically our searches look for certain features (spikes, triplets, gaussians, pulses and others), and for the time being uses mostly classical (brute force) Fourier techniques to find them. As better/larger hardware becomes available, and better libraries, ai techniques [, such as convolutional neural networks, ] become more plausible as efficient alternatives. Whether or not machine learning can overtake what we do now in accuracy and performance is an open question, though it's looking more and more likely as time passes.

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

Probably most of the improvements medium term will come from recent contributions from Petri33. Longer term probably looking at trying to leverage some of the ai targeted features not explored in setiathome code yet (longer term prospect).

Yes, I noticed efficiency of Petri33 crunching with his custom applications, 2-3x faster then stock applications. He is obviously very skilled programmer, so it is very good if he is willing to contribute to whole community.

For long term, I agree, we should always be open minded willing to explore new features and techniques to improve "the quest"

---

Probably most of the improvements medium term will come from recent contributions from Petri33. Longer term probably looking at trying to leverage some of the ai targeted features not explored in setiathome code yet (longer term prospect).

Yes, I noticed efficiency of Petri33 crunching with his custom applications, 2-3x faster then stock applications. He is obviously very skilled programmer, so it is very good if he is willing to contribute to whole community.

For long term, I agree, we should always be open minded willing to explore new features and techniques to improve "the quest"

Hi,

Thank you. I'm a former programmer and nowadays I work as a teacher. I like to keep the programming as a hobby :) And it is a hard hobby. I do not like the accuracy bug in my present code. I may have to do a partial backtrack and reimplement all improvements one by one and keep an eye on the inconclusives. The KWSN bench reports all OK at over 99%.

The AI and neural Network stuff may advance especially in pattern recognition.
The pattern in Seti signal is (like said), pulses (repeating regularly), spikes (one high value), gaussians (rising and falling power over time), autocorrelations (self similar signal) and triplets (three spikes at even distance from central frequency). Things are easy to recognise but the input is about one million samples and the input is preprocessed to compensate a doppler shift and rerun a million times.

P.

All that would require a (still at a moment) huge neural Network. But I like to think that the most radical improvements come from where no one expects them.

---

To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones

All that would require a (still at a moment) huge neural Network. But I like to think that the most radical improvements come from where no one expects them.

LMAO, Elon Musk just tweeted that nVidia donated a DG-1 for working on poopy cars :) I think we could make better use of that :P

Yeah I think we'll come across the validation issues. If you're concentrating on backtracking, while I'm spending the limited time on new tools to test accuracy/precision, I think eventually either or both of us will wring out the quirks.

Fun times, LoL.

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

All that would require a (still at a moment) huge neural Network. But I like to think that the most radical improvements come from where no one expects them.

LMAO, Elon Musk just tweeted that nVidia donated a DG-1 for working on poopy cars :) I think we could make better use of that :P

Yeah I think we'll come across the validation issues. If you're concentrating on backtracking, while I'm spending the limited time on new tools to test accuracy/precision, I think eventually either or both of us will wring out the quirks.

Fun times, LoL.

Ha ha,

There is an 'order now' button at the bottom of the link.

---

To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones

All that would require a (still at a moment) huge neural Network. But I like to think that the most radical improvements come from where no one expects them.

LMAO, Elon Musk just tweeted that nVidia donated a DG-1 for working on poopy cars :) I think we could make better use of that :P

Yeah I think we'll come across the validation issues. If you're concentrating on backtracking, while I'm spending the limited time on new tools to test accuracy/precision, I think eventually either or both of us will wring out the quirks.

Fun times, LoL.

Ha ha,

There is an 'order now' button at the bottom of the link.

ROFL, yeah that'll happen. Full points for optimism nVidia.

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

All that would require a (still at a moment) huge neural Network. But I like to think that the most radical improvements come from where no one expects them.

LMAO, Elon Musk just tweeted that nVidia donated a DG-1 for working on poopy cars :) I think we could make better use of that :P

Yeah I think we'll come across the validation issues. If you're concentrating on backtracking, while I'm spending the limited time on new tools to test accuracy/precision, I think eventually either or both of us will wring out the quirks.

Fun times, LoL.

Ha ha,

There is an 'order now' button at the bottom of the link.

ROFL, yeah that'll happen. Full points for optimism nVidia.

The NVIDIA DGX-1 is available for purchase in select countries and is priced at $129,000*. DGX-1 service and support at additional cost. - See more at: http://www.nvidia.com/object/deep-learning-system.html#sthash.5VgPbxFq.dpuf

---

All that would require a (still at a moment) huge neural Network. But I like to think that the most radical improvements come from where no one expects them.

LMAO, Elon Musk just tweeted that nVidia donated a DG-1 for working on poopy cars :) I think we could make better use of that :P

Yeah I think we'll come across the validation issues. If you're concentrating on backtracking, while I'm spending the limited time on new tools to test accuracy/precision, I think eventually either or both of us will wring out the quirks.

Fun times, LoL.

Ha ha,

There is an 'order now' button at the bottom of the link.

ROFL, yeah that'll happen. Full points for optimism nVidia.

The NVIDIA DGX-1 is available for purchase in select countries and is priced at $129,000*. DGX-1 service and support at additional cost. - See more at: http://www.nvidia.com/object/deep-learning-system.html#sthash.5VgPbxFq.dpuf

Honestly for something that can use it that can justify the cost ? Well worth it. i Think they might do well. It does place a price on the rest of us plebs trying to make things work with toothpicks and chewing gum... just sayin

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

....It does place a price on the rest of us plebs trying to make things work with toothpicks and chewing gum... just sayin'

Amen, brother! Now, grab another pack and start chewin'! I'm heading out now on a toothpick run, need anything else while I'm out? Bailing wire, tin cans perhaps? ;-)

---

....It does place a price on the rest of us plebs trying to make things work with toothpicks and chewing gum... just sayin'

Amen, brother! Now, grab another pack and start chewin'! I'm heading out now on a toothpick run, need anything else while I'm out? Bailing wire, tin cans perhaps? ;-)

Duct/Duck-Tape --> We're going high-tech here!

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

If you can't fix it with duck tape, you didn't use enough!!

....It does place a price on the rest of us plebs trying to make things work with toothpicks and chewing gum... just sayin'

Amen, brother! Now, grab another pack and start chewin'! I'm heading out now on a toothpick run, need anything else while I'm out? Bailing wire, tin cans perhaps? ;-)

Duct/Duck-Tape --> We're going high-tech here!

Better to use high strength velco then you can open it up to remove the bugs, before resealing ;-)

Order for fixing things that are moving and shouldn't:
hammer -> BIGGER HAMMER -> Duct tape -> MORE DUCT TAPE

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

The NVIDIA DGX-1 is available for purchase in select countries and is priced at $129,000*. DGX-1 service and support at additional cost.

From $129,000, and you have to pay extra for support?
Got to love the high end of town.

---

Grant
Darwin NT

I guess it's kind of like someone owning a Hummer H1 or a Ferrari. If you can afford the car, you'd better not care that they get 5 mpg or the cost of fuel.

---