Posts by jotheberlock 79 publicly visible posts • joined 4 Nov 2010
Can't speak for US TV, but over here IP over your TV/internet provider's multicast network (not wifi, or not necessarily) is the newer thing over Digital Terrestrial Television (an antenna), but uses the same transport mechanism (MPEG transport streams, which are designed for one-way transmission over lossy mediums, which describes both of those)..
It'll get its lunch eaten by on-demand streaming services like Netflix eventually but in the meantime it's more obsolescent than obsolete, especially as Boomers still watch linear TV.
No it really wasn't - http://www.easy68k.com/paulrsm/doc/dpbm68k1.htm 'Three Arithmetic Units' section.
The fact you can do 32 bit operations in the ISA does not mean they are actually being done all in one go. Microcode is driving the operation through the ALU twice, hence it takes twice as long to do those operations. It's in the section I just linked - 'The MC68000 also operates on 32-bit data. This is usually done by taking two passes of 16-bit data, one for the lower word and one for the upper word. This is reflected in the execution time of many 16- and 32-bit instructions.'
If you literally mean 'system call', no, you can't, not in a Linux userspace program. At the very least you'll have to call exit() ( 'mov rax, 60' then 'syscall' on x86-64 ) or execution will continue past the end of your program and you'll segfault. And if that's the only syscall you're doing your only interaction with the outside world is the exit code.
You could in theory write your own kernel I suppose but there's a few more implications there than 'what if they improve the syscall'.
'the software is so badly written that it will use precisely one thread running on one processor'
Look up Amdahl's Law. If the thing you are trying to do is fundamentally sequential, as is often the case - if you need to compute a before you can compute b before you can compute c - then it doesn't matter how many cores you have available, you are going to be stuck using one core for that thing. Not that there aren't cases where they could help and aren't utilised as much as they could be but more cores/threads are not a panacea. A 96 core processor does not and never will do things 96 times as fast as a single core.
'Research chips' aren't shipping.
Source code definitely doesn't count.
Qualcomm is this year still 'getting involved in' a joint venture to make RISC-V chips -https://www.qualcomm.com/news/releases/2023/08/leading-semiconductor-industry-players-join-forces-to-accelerate - and is in a licence dispute with ARM so heavily incentivised to talk RISC-V up.
I'll believe the hype when I start actually seeing a bunch of shipping, commercial RISC-V products out there. Articles from 2017 and I still can't buy a phone with one in?
'What there has been, is the arrival of competition, in the form of RISC-V, which is now taking a slice of Arm’s microcontroller market,' - citation needed I think. People have TALKED about it a lot which is what you do if you want a better deal from Arm, but how many design wins? How many shipping products with RISC-V in? There's a couple but Arm is still heavily dominant.
It did die, pretty much, a modern desktop x86 CPU is RISC-like internally (and despite being a bit ugly, x86 isn't as CISC as say VAX anyway - importantly, for example, you don't get multiple memory operands in an instruction, or some of the wacky stuff like being able to do *******< some pointer> in hardware, in one instruction, which plays hob with modern caches/memory .
(On the other hand, fixed length 32 bit instruction sets /also/ didn't entirely stick around - hence ARM Thumb and MIPS16 - because once you get to a certain point instruction decode becomes much less of a bottleneck anyway)
Most employees are theoretically replaceable - skilled ones take months to hire, then months to learn your stack, before they can become nearly as productive as the previous guy, even assuming you haven't hired a lemon which absolutely does happen. And they still won't have the years of experience of both how the systems and the company works. Employees can't be swapped in and out like lego bricks and if you think so you are the epitome of a terrible manager.
It does, though, if you're on any form of Unix. Calling procedures are defined by what was originally the Itanium ABI. Windows has its own thing, but either way there are two standard calling procedures, Windows and Unix (including MacOS, it does the same thing as Linux) for x86-64. That's as much as, well, any platform has, VAX included as far as I'm aware.
Not that this excuses quite that big a driver (maybe NVidia Experience is on there too?) but bear in mind graphics cards are much much more complicated than they used to be, most of it in userspace. There's going to be a complete optimising compiler on there for the shaders, after all.
You could always use Scheme, which guarantees tail call optimisation (specifically so you can write that sort of recursive code functional programmers etc find so elegant...)
A C compiler can do it too but you're not /guaranteed/ it'll do it, so not something you'd want to rely on in production.
Canonical is based in the UK, where you are not able to go to an employment tribunal in the first two years of employment except for the most obvious of reasons (e.g..being outright told you're being fired because you're black) - so until you hit the two year mark you can be fired about as easily as in the US although your notice period is more likely to be a month+ than two weeks.
A 'reform' introduced by New Labour at one year and then doubled by the Lib Dem/Tory coalition, so thanks to all three major UK parties there.
For what it's worth it's not really 'the specific C++ compiler' any more and hasn't been for years. Windows has its own ABI, everyone else uses the relevant CPU-specific variant of the ABI originally developed for the Itanium 20 years ago (so, g++ on Linux and MacOS, clang on similar, basically everyone else outside of maybe some ancient commercial Unix?)
Try going to an employment tribunal in your first two years' of employment. You can't. You're out of luck, it's not legally allowed, so an employer can basically fire you for whatever with no comeback.
Blair brought in a 1 year limit for 'flexibility' under New Labour and the Lib Dem/Tory coalition extended it to two years so basically all three national parties have screwed us on this one.
I would assume what this really means is 'Intel will provide ARM drivers and official support'. The only actual issue I can think of between CPU and discrete peripheral would be a 64 bit peripheral where the CPU is 32 bit and can't generate 64 bit reads/writes (or where the peripheral has > 4gigs of mappable memory that the CPU needs to talk to directly of course - GPUs maybe?). Endianness is certainly a pain but not theoretically insuperable.
I mean the whole thing with PostScript is it's a complete programming language, not just a graphics file format - removing the Turing completeness was one of the points of replacing it with PDF. You'd need a full-on PostScript interpreter running in the browser to guarantee being able to render arbitrary PostScript, and that has security implications etc just as with Java applets back in the day.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
