What If Exercise

in gridcoin •  7 years ago  (edited)

I had hard times waking up this morning. Half asleep, I hurried up to catch the morning metro before it got very crowded. While waiting for the train to come, a question came to my mind: what if one could enhance an existing systems with Gridcoin mining capabilities? You have only one answer, and you need to chose the best one: what kind of systems would you thing about?

My choice would be Tesla vehicles.

In 2017, Tesla delivered 103 thousands cars with autopilot capabilities, 35% more compared to 2016. To reach self driving capabilities, Tesla selected a system using radar, 360 degrees cameras, and sonars, resulting in the need of a huge computing power to process all those data. The computing power is provided by a NVIDIA DRIVE PX 2 AI platform, with a performance of 10 FP16 TFLOPS.

What is all this power used for when the car is sitting idle in a queue on the interstate, or when the car is charging its batteries in a dark garage in California? Nothing! What if the cars will search for a cure for cancer, or for extraterrestrial intelligence?


Image courtesy of Daimler-Benz

Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!
Sort Order:  

Wow, that'd be a good solution. Didn't think that it was that much power in those that could be utilized.

That would be great.
Unfortunately GPUs are very bad at multitasking.
That photo looks familiar.

Depends on the architecture. Actually, AMD Radeon GPUs are very good at multitasking, and that is the reason for their intensive use in the second phase of Bitcoin mining. NVIDIA might not be that strong at multitasking, but they might solve efficiently more complex problems.

Yes, the second picture is commonly used by NVIDIA for promoting their self driving technologies. I also seen it in many articles about AI in autonomous driving.

By multitasking I not meant only performing two tasks at once, but performing normal tasks at speed unaffected by low priority (boinc) work. NVIDIA drivers totally ignore process priority and give out most resources to boinc, because it's apps take what they can.
Radeon drivers at least do not lag desktop, but still reduce game FPS with background boinc work.

Now even TVs are equipped with normal processors, then at night - if there is such an opportunity they can also be used.

Interesting idea

Willing to contribute with your idea?