HN user

srcmap

588 karma

www.srcmap.org

https://www.twitter.com/srcmapy

Posts3
Comments342
View on HN

What about this: Use SQLite.

Make it one SQLite file per user. All the read/write are parallelized per user. Backup db file would be smaller, easier. Each service node can handle N users, less for heavy follow users. Should be easier to scale horizontally the with user count by adding system nodes and Storage, networks, web handling process.

Remember need to scrape off bugs from windshield after each trip thru I5 ~10 years ago. No need to do that anymore.....

On the other hand, in my back yard, I grow a lot veggi, strawberry with worm casting from kitchen scraps. I do still lot of insects. A lot of birds nested around my house. Produce 10-15 young birds. I found a very cool looking giant tomato worm a size of my biggest finger. It ate half of (7 ft height ) tomato plant in 2 days.

The Strawberry, cucumbers, other veggies grow with worm casting taste SO much better than from anything store .

Yap,

Someone likely would just docker/vm to partition that 80 cores into micro-service handle httpd proxy and app, db backend.

Those app/VM can easily be converted to app running pi modules.

20 pi modules likely have much better DDR, SSD, Network bandwidth. Probably scale from 2 pi to 200 pi as easily as typical vm setup - and it comes with GPUs for free for those need them.

I have been skipping GYM because of Covid.

But I do go running in the park and follow a few HIIT youtube workout. There are a lot of those in youtube. Just find the one with intensity and length you like and follow them and advance to next level in your own place.

My first compute was ~1984 Timex_Sinclair_1000 $99 and with 2K RAM. https://en.wikipedia.org/wiki/Timex_Sinclair_1000

It definitively the gateway drug that hook me into computer.

Bought $6 book on how to program Sinclair computer and copy the Basic Code from the book into that computer and record that program into audio cassette tape and output to analog TV.

Now I working daily with systems that have 128 cores, 256 threads, 256-768MB of L3 cache memory, 1TB DDR+ 8xGPUs each with 32GB of HBM.

I still think I got more "High/Rush" from the sinclair 1000 37 years ago than the latest 128C/256T CPU. Now is a job, back than it was an adventure, exploration to whole new world!

One gdb trick I used:

   Use simple gdb script,  break, print vals, cont, 
   This allows one to debug and app's internal states without rebuilding with printf enable. 
   Combine the above with automated regression test system, one to debug very hard problems that might take hours/days to repro.

From government prospective, it is probably cheaper than unemployment, continue the lock down and more stimulus checks

Curious if it will triggers different kind of "Cobra Effect" -

https://freakonomics.com/podcast/the-cobra-effect-a-new-frea...

    MEHROTRA: So the “cobra effect” refers to a scheme in colonial India where the British governor, or whoever, the person in charge in Delhi, wanted to rid Delhi of cobras. Apparently in his opinion there were too many cobras in Delhi. So he had the bounty placed on cobras. And he expected this would solve the problem. But the population in Delhi, at least some of it, responded by farming cobras. And all of a sudden the administration was getting too many cobra skins. And they decided the scheme wasn’t as smart as initially it appeared and they rescinded the scheme. But by then the cobra farmers had this little population of cobras to deal with. And what do you do if there’s no market? You just release them. And so this significantly, by a few orders of magnitude, worsened the cobra menace in Delhi.
eBPF on Windows 5 years ago

From the architecture diagram, it looks like just network ebpf only.

Does anyone know if it can profile file/disk io type activities?