Roughly how many varieties of module were used for a full 604 system?
I suspect quite a few as other "modular" systems in the transistor era like the later IBM Standard Modular System and DEC Flip-Chips ended up with plethora of specialised modules, but I'd be interested if that growth had already begun in the tube era.
I don't have the information for the 604, but the 605 used about 36 different types of tube modules. IBM couldn't resist making different types of modules; the 650 had sixteen types of cathode followers alone.
That's not true - DOS lacks multitasking but it absolutely has a concept of processes. They're called PSPs (program segment prefixes, a struct containing data about that process) and the OS has a range of syscalls to manage them. Any program can spawn a child program and wait for its completion.
The memory allocation syscalls track which program owns each dynamically allocated block of memory (the block has a header which points to the PSP), and frees the blocks allocated by a program when it exits. The filesystem syscalls deal with file handles opened by processes in just the same way.
The famous "terminate and stay resident" syscall causes a program to exit and return to its parent process while keeping its PSP and dynamically allocated memory blocks intact.
16 bit Windows introduced multitasking but reused DOS's PSP based process model and would allocate a PSP for each Windows task, switching the current DOS PSP when switching between Windows tasks.
From somewhere in the middle of Nvidia’s endless press waffle:
“The RTX Spark superchip features an NVIDIA Blackwell RTX GPU with 6,144 CUDA cores and fifth-generation Tensor Cores with FP4 precision, connected via the NVIDIA NVLink®-C2C chip-to-chip interconnect to a high-performance, 20-core NVIDIA Grace™ CPU.
MediaTek, a market leader in Arm-based system-on-a-chip designs, collaborated with NVIDIA on the custom CPU design, contributing to its best-in-class power efficiency, performance and connectivity.“
Sooooo they binned all their old DGX Spark crap to pawn on poor clueless consumers to try to be an ARM fighter and probably still fail miserably behind Apple Silicon and even likely older Qualcomm metal running Windows.
And Mediatek? Oof. I assume the SOC comes pre compromised out of the box.
> And Mediatek? Oof. I assume the SOC comes pre compromised out of the box.
If by compromised you mean “china bad”, mediatek is taiwanese not chinese. Same home as asus, asrock, tsmc, htc, acer, d link, adata, biostar, insyde, gskill, foxconn, realtek and many others. From the chip to bios in your pc probably.
If you mean quality, they make efficient and more powerful chipsets especially gpu wise compared to qualcomm. Most probably only know it from cheap chinese phones.
MediaTek used to supply a lot of lower-performance MIPS and ARM chips during 2010s. GP probably has that MediaTek == bad experience association. The actual chips weren't unreliable or horribly engineered, just slow, but people don't make a clear distinction there.
no, not calling "china bad", calling out all the SOCs they had some pretty glaring exploit holes on that a bunch of stupid OEMs kept shipping long after the holes were well known. i've also had a pretty bad time with a lot of their wifi chipsets on our networks, but hey.
That is true. But to be fair, the exploits needed physical access to the device and Qualcomm also had their fair share of those issues. I believe actually trust worthy companies who ship both chips did actually fix those issues.
It’s the no name brands. And those are all odm devices which mtk already sold off. Talking about MT65xx era where you didn’t even need to unlock the bootloader for a flash.
Mediatek used to make a lot of low-quality cpus for cheap and short-lifespanned products (eg: those cheap tablets that are essentially useless as soon as they hit the market)
It’s kind of irrelevant that they might be behind apple silicon because they’re targeting the non-apple, windows-using, section of the laptop market with a chip that can ostensibly be used for running AI models locally. Whether there’s much appetite for business users doing that remains to be seen.
The problem isn’t the chip. The problem is Windows and the level of Microsoft support going forward and right now Microsoft is preoccupied/distracted with Copilot.
That is the main problem that Qualcomm having currently with Microsoft.
I suspect Windows 2000 (and NT based stuff in general) is too near to modern Windows to be (officially) released. DOS is long dead, but modern Windows still uses the NT kernel and Win32 and so on, and they probably don't want to give an official peek behind the curtain, even if it's an over a quarter century old version.
Yes, Microsoft had established themselves as supplier of BASIC interpreters to most of the US microcomputer manufacturers. There initial contact with IBM was to provide a version of Microsoft BASIC for the new computer.
MS-DOS doesn’t have memory protection, so another option is that the running program[1] or something like a TSR or driver could have corrupted the headers.
[1] I guess in a modern system a process can still trash its own malloc, but not the kernel’s page allocation data.
You wouldn’t be getting a protection fault when protection isn’t enabled.
I should also note this error is the return code for MS-DOS’s memory management functions (such as int 21h ax 48h / 49h), vaguely similar to malloc() returning NULL. It’s not a fatal error, so how it’s handled depends on the programme. It could bail out, perhaps with a more general “out of memory” error, or try and carry on, or perhaps just start overwriting parts of the interrupt vector table as that’s where segment 0007h would start at…
(Though in the latter case on DOS just blindly assuming a memory allocation worked would be rather unwise so you’d hope just about everything checks the carry flag first).
It would presumably be the August 1991 coup if it were the Soviet Union, as it was one of the factors leading to the USSR dissolving at the end of the year. The Autumn 1993 coup was in the Russian Federation (and the geriatric plotters in the Kremlin kinda won that one). So 1992?
That's what I thought, too, but the top of the article says "For the Conference on HyperNetworking, Oita, Japan" which was in '94. So, I thought maybe "last summer" was internally even off by a couple years? I'm really not sure, but someone around these parts probably knows. Worth mentioning that also in the early 90s people did refer to the Russian Federation as the Soviet Union sometimes as a shortening of "the former Soviet Union".
The estimate of "The Internet" connecting 800,000 computers is probably also pretty surgically date-identifying (at least to isolate 1992 to 1998 given how fast it was growing at the time, though estimation error might cause a little trouble!). For example, https://web.mit.edu/people/mkgray/net/internet-growth-summar... also suggests 1994 (although that estimate was 0.6 million) while 1992 would be more like 200,000 although as per my scare quotes (and that MIT link) "The Internet" was also a somewhat vague term at the time. And by 1998 it was surely over 10 million which makes the @karel-3d quite likely incorrect, although who knows - maybe that's when the EFF first put it up on their web site?
EDIT: I mostly think it matters since observations that might have seemed quite prescient in 1992 (like also-Mormon Orson Scott Card's even more prescient ideas in 1985 Ender's Game with Locke & Demosthenes political chat personas based on 1980s BBS/UUCP network activity) were very much things everyone was saying by 1998.
FWIW, Claude followed a similar trail and honed in on Spring 1992:
Three internal anchors line up tightly on 1992:
- EFF's age. Barlow writes that EFF, "after almost two years of operation," now thinks of itself as building "the united Mind of Humanity." EFF was founded July 10, 1990, so "almost two years" puts the writing at roughly spring/early summer 1992.
- "Last summer's coup in the Soviet Union." The August 1991 Moscow coup. Read naturally, "last summer" was written sometime between fall 1991 and the end of summer 1992.
- Internet size. Barlow says the Internet "connects some 800,000 (mostly UNIX) computers." Per RFC 1296, the host count was 727,000 in January 1992 and reached ~992,000 by July 1992. 800,000 sits squarely in spring 1992.
He also doesn't mention Mosaic, which launched in January 1993 — by 1993 it was almost a reflex in Barlow's writing — which is weak supporting evidence that this predates that release.
MCGA was basically a cut down VGA on the motherboard of the low end 8086 / ISA PS/2s, and both chipsets were introduced at the same time. Next-to-no-one would want a not-quite-VGA that lacks high res colour or EGA modes, so nearly nobody bothered cloning it, only the higher end option. The only known clone is in a couple of Epson models, and like the PS/2s it’s a motherboard integrated chip, not an expansion board.
I suspect quite a few as other "modular" systems in the transistor era like the later IBM Standard Modular System and DEC Flip-Chips ended up with plethora of specialised modules, but I'd be interested if that growth had already begun in the tube era.
reply