HNHacker News
TopNewBestAskShowJobs

bullen

750 karma · joined September 11, 2015

https://multiplayeronlinestandard.com https://tinspin.itch.io
submissionscomments
bullen··on ARM’s Cortex A53: Tiny but Important
I still use original Raspberry 2's (A53) in my server clusters, they are the lowest power devices that can saturate (good) SD cards on random writes/reads and still have performance left to serve calculations.

A72 in Raspberry 4 is really the pinnacle low power CPU performance if you count $ and compatibility (linux <-> GPU). ~3x can saturate symmetric 1Gb/s with advanced computation and still have calculations left.

You can buy neither, we'll see if the 4 ever comes back. The only thing I'm 100% of is that the 5 will have some drawback compared to both the 2 and the 4.

I can concur the peak of A55; the 5W TDP 22nm RK3566 in my RG353M is mind boggling! HL1 at 60 FPS and soon HL 2 at 30 FPS.

But Risc-V is not progressing with linux, mainly GPU drivers and integration is the problem. Historically the Chinese boards never got any attention and unfortunately you cannot depend on attention happening this time either.

Partly because now companies across the board(ers) are hiding kernel configs again. And "board support packages" are slow to be mainlined if ever.

bullen··on Ask HN: Suggestions to host 10TB data with a monthly +100TB bandwidth
I think 2x 1 Gb/s symmetric home fibers + SuperMicro 12x SATA Atom Mini-ITX with Samsung drives can solve this fairly cheaply and durably depending on write intensity.

That said above 80 TB is looking hard for eternity, unless you can provide backup power and endure noise of spinning drives.

bullen··on Building a Handheld PC
What SoC is this using? For the life of me I can't find it in the article?

Edit: ok it's a CM4? on the back of the display...

For others searching for the secret sauce: BigTreeTech Pad 5

bullen··on Server-sent events
Try the connection from other server + client locations.
bullen··on Server-sent events
Websockets use one TCP socket for starters with SE you use separate sockets from ingress and egress which is much simpler to scale from an infrastructural point of view.

I clearly said many-to-many, everything else is trivial to scale because only many-to-many requires strictly shared memory atomic parallelization.

You could write a better server in Java than javascript/c(++) but I haven't found any, and to be honest the extra complication of the binary protocol doesn't even make me curious enough to waste 5 seconds to google it.

SSE over HTTP/1.1 is the final solution. Everything else is a waste of time for eternity.

bullen··on Server-sent events
It wasn't "not supported" they had a bug that they never could fix because it would have broken 50% of the internet.

Be careful what you release because it might not be fixable after release.

bullen··on Server-sent events
Try debugging HTTP/2 on a live machine when you have 10.000 concurrent users and get back to me how well you sleep.

console.log or wireshark

bullen··on Server-sent events
They work EVERYWHERE except through some very rare buggy anti-viruses and ISP implementations.

Real life stats: "759 out of 205.875 customers refunded because their antivirus blocks the pull HTTP stream", so 99.7%!

They even go through the great firewall without bureaucracy.

bullen··on Server-sent events
String parsing is never the bottleneck. You need to parallelize (shared memory) and understand your system and for that text is pretty convenient.

So also avoid HTTPS.

It's trending because it's the final real-time solution for eternity.

bullen··on Server-sent events
You need to multi-thread your server on shared memory to build many-to-many solutions. The simplest way to get that working is Java + NIO with concurrent package, I have done it for you: http://github.com/tinspin/rupy

My web server outperforms everything on the planet for multiplayer.

Websockets are notoriously hard to parallelize if you want shared memory which you need for many-to-many, mostly because JavaScript and C++ have very poor atomic concurrency, or atleast I never found any Websocket server that could do it.

bullen··on Server-sent events
You need to send "keep-alive" "noop" to keep the routers happy, I do it ever 5 seconds.
bullen··on Server-sent events
Just create more subdomains if you really need more sockets.

Or move away from the browser. Browsers are toxic now.

bullen··on New TLDs: Not Bad
No, this is ridicoulus. The browser does not execute a jpeg. It displays it.

The reason malwares worked was insecure automatic execution in IE. You do NOT need https for the browser to type check and securely sandbox downloaded data.

MITM is not the problem, because if I can send a malware from my https server you still have the problem, all https does is waste a ton of electricity for your job (in)security.

If your browser requests a jpeg an my https server returns a exe that the browser saves and the user is foolish enough to execute it, that is not a http problem.

Stop using your monopoly to force waste.

bullen··on New TLDs: Not Bad
When I link to http (because I want to share public information that never will need encryption), your browser rewrites that to https and then the user gets an error.
bullen··on New TLDs: Not Bad
Can you stop redirecting http to https if there is nothing on port 443 please.

Also zip does not execute.

bullen··on Implement DNS in a Weekend
I'll dig up my old comment about how easy DNS is to work with using DNS4J:

  Message query = new Message(data);
  Header header = query.getHeader();
  Record question = query.getQuestion();
  Message response = new Message(query.getHeader().getID());
  response.getHeader().setFlag(Flags.QR);
  response.addRecord(question, Section.QUESTION);
  Name name = question.getName();
  int type = question.getType();
  int dclass = question.getDClass();
  String host = name.toString(true).toLowerCase();
  ...
  response.addRecord(new ARecord(name, dclass, 300, "someIP"), Section.ANSWER);
  ...
  response.getHeader().setFlag(Flags.AA);
  return response.toWire(512);
bullen··on Freakwan: A MicroPython driver for the SX1276 LoRa chip
http://radiomesh.org
bullen··on Freakwan: A MicroPython driver for the SX1276 LoRa chip
In EU yes.
bullen··on Is sequential IO dead in the era of the NVMe drive?
Seems SSDs are reaching the end game with ~8TB on a 2.5" platter.

Now you just need to nail the peak of $/GB, I'm looking at both 870 EVO / QVO, they have 4TB versions for $300.

For "write once" or append only data these could work well, think user registry or sequential log.

But you need to offload the active data on SLC drives or take the risk of the wearing driver firmware in the exact devices you buy/patch.

The OS needs to be on SLC anyhow.

bullen··on Freakwan: A MicroPython driver for the SX1276 LoRa chip
I thought the chirps where actually not spanning more than a channel band worth of frequency?

Also you cannot control the frequencies according to the spec.

You need lower frequencies for range and control over those frequencies to not saturate the ether.

The bandwidth will be low for these radiomesh networks, but if you build a robust enough routing and get enough channels (300 with 5KHz separation) you can scale.

I'm going to use the upper part of the open 169MHz in Europe.

bullen··on Ed Sheeran demonstrates how copyright is destroying musicians
“when Alexander (The Great) saw the breadth of his domain he wept, for there were no more worlds to conquer." - Hans Gruber / Hollywood

OR

When for more worlds the Macedonian cried,

He wist not Thetis in her lap did hide

Another yet, a world reserved for you

To make more great than that he did subdue …

  —“A Panegyric to My Lord Protector,” lines 73–6
bullen··on SpaceTraders: A multiplayer game built on a free web API
So far it does not.

I have not pushed it though because I'm making my own 3D engine to go with the backend: http://talk.binarytask.com/task?id=5959519327505901449

That said, I think it's the only way forward.

Both the tech and the license needs openness.

Unfortunately open-source means free forever, while "source available" got the backseat and allows money to be involved.

Not explicitly but indirectly through "you can just download and compile it yourself" for "free".

This is also an incentive to make build systems opaque, another trend I have noticed.

Eventually compiling something will take weeks and months to setup. XD

bullen··on SpaceTraders: A multiplayer game built on a free web API
Backronym ftw.

Have been switching between system and standard for the S, standard is kinda pretentious as it needs adoption, but you gotta stay positive.

bullen··on SpaceTraders: A multiplayer game built on a free web API
While this seems focused on trading, kinda like EVE (spreadsheet MMO), I think physical action is the real attraction of 3D graphics/audio:

So I made a very scalable MMO protocol: https://github.com/tinspin/fuse

It's event based so analog movement should be compressed, which means headshot FPS are not the target audience. But it works fine with action games that take scalability into concern.

I'm curious how hard people find it to adopt. I'm guessing I need to deliver a good and open game on top of it to see adoption.

bullen··on RedisRaft
Worst is 0.7 seconds for save and 0.1 for load it seems.

Average is 4.8 millisec and 0.5 millisec respectively.

But those are just the JVM doing it's thing.

The numbers I like are these: 190.2 158 331

200ms global save average, 158ms min and 331ms max from europe to east+west US and Asia. Without fault, so consistent, much of that can be attributed to AWS improving so much over the years.

As to why the load/save are slower it's because the complete global roundtrip stats I only have for registers which are rare now so they don't hit the GC I'm guessing.

The thing I'm most proud of is async-to-async meaning the system will saturate all cores (without io-wait) on all machines without problems... it just keeps solving the problem at 100% efficiency, no memory leaks and 5 years uptime without any crashes.

Slowing down is the worst case, and if that is a problem just upgrade instance type, no uptime then though.

bullen··on RedisRaft
I made my own distributed JSON over HTTP database back in 2016.

It has been running in a intercontinental production environment with 100% read uptime since 2017.

It's 2000 lines of code: http://root.rupy.se (this test env. has 3 nodes: fem, six and sju)

bullen··on Lifestyle design from first principles
This is a good way to learn nothing of value.

You need to think different and to use your body as a petridish is the fastest way to understanding the world.

You're going to die anyway.

The real hard part is getting people to accept your discoveries. Most problems need to be presented as solutions and by that time it's often too late.

bullen··on How to set your domain as your handle
"encryption with no preshared secrets" that does not exist, never have and never will.
bullen··on Mojo: The usability of Python with the performance of C
Python is hardly usable for big projects.

The closest I get to this is hot-deploying a C dynamic lib compiled with debug on a core C++ engine compiled with release.

That way I get stacktraces from "the moving code" I can read remotely without losing performance (move the heavy parts into the core engine).

The engine needs to be extremely stable for this to work though.

bullen··on I want to talk about WebGPU
This article is so full of obvious false information, the most important is: "OpenGL ES 2.0 was the same as OpenGL 3.3, somehow."

VAO is the final GL feature.

There is no point to use anything other than OpenGL (ES) 3+ for eternity.

Copying the wheel over and over is the trap.

← PreviousPage 12 of 34Next →