General statistics
List of Youtube channels
Youtube commenter search
Distinguished comments
About
Michael Lenczewski
Core Dumped
comments
Comments by "Michael Lenczewski" (@kayakMike1000) on "Core Dumped" channel.
Previous
1
Next
...
All
This voice is actually pretty good.
8
Yikes... Isn't .NET running on a virtual machine like Java? If you like this kinda stuff, check out RISC-V or ARM microcontroller assembly. It's really deep down the rabbit hole, for sure, but you gain a real appreciation for how the CPU does its thing.
3
I like his text to speech. It's nice.
2
No, simplest program you can write is blinky.
1
Well... Even in MCUs... We have peripheral interrupts that cause the cpu core to jump into some code to handle the cause of the interrupt. Sometimes its simple action to swap a buffer pointer or reset a timer... But the operation of the peripheral could be abstracted as a simple thread that runs concurrently with all the other stuff on the MCU core. That thread has an implicit loop that "wait for something to happen on this interrupt, then do stuff"...
1
The latest micros have 100s of interrupts, they kinda grew over the years. A few dozen became 100 then 200 or 300. You only really use a few dozen at most...
1
@CoreDumpped I see a future RISC-V dev!!!
1
in the RISC-V ecosystem, there are "HARTs". HARTs are hardware supported threads. A core can be constructed with extra hardware that can make one core operate on two threads, as in hyper threading.
1
The parameters of the living world changed just last week. The universe just did that awful vacuum decay thing that we all dread, but it was such an insignificant change, no one noticed.
1
Never communicate by sharing memory. Share memory by communicating.
1
@ddopson you think THAT is silly? We went to war over which end to start eating a hard boiled egg.
1
Rust sucks. It will die eventually.
1
There are systems that don't do any caching and linked lists are pretty good.
1
Operating systems... I am of the opinion that the OS really just needs to handle IRQs. IRQ service routines pretty much run the core in kernel mode, at least in arm cortex cores...
1
You're describing 1970s technology. Transistors in logic gates are typically voltage controlled. A metal oxide insulation layer mostly stops current through the "gate" or base. The presence of charge on the gate determine the conductivity of the transistor.
1
Well... Drivers don't necessarily NEED to run in kernel mode. A kernel could simply handle any peripheral interrupts and then dispatch a signal to a user mode process. Its just not as snappy and has some overhead, but you get the benefit of drivers not crashing your system. Stability comes at a price here...
1
@Censored_Truth_Addict You're correct.... it's LESS affected because this is an SRAM memory cell that likely uses up to 12 transistors to store that bit. DRAM use one transistor and a capacitor. You might be able to flood the CPU with enough ionizing radiation to flip the protection bit, but.... You're way more likely to flip bits in other registers...
1
@NeovanGoth all transistors are on or off by virtue of a small electric field on a MOSFET gate. If a cosmic ray happens to hit that tiny gate, you get a bit flip... DRAM cells have a much bigger target (100x the area) because of the capacitor AND these bits travel across the board. It's not just cosmic rays that cause bit flips, it could be a random beta decay from radon in your basement...
1
Oooooooh.... CMOS works a bit different. This is TTL...
1
Mainframe... That's typically a computer designed for extremely good transactional processing
1
Previous
1
Next
...
All