The Game Boy has a Z80-like processor, though. Generally, Z80 instructions take more clock cycles to execute than 6502 instructions, so the clock speeds are not directly comparable.
For the same reason that they also include symbols for many obsolete (dead) languages and writing systems, as well as (per a comment above) a character used by the 1959 IBM 1401 computer (https://github.com/shirriff/groupmark). The need to be able to discuss a certain technology in writing does not disappear just because it is out of general use.
Some of their home computers competed pretty well.
The Atari 8-bit series (400/800) had a significant chunk of market share in the early 80's. Not among the top three, but there were so many incompatible platforms in those days, and Atari 400/800 was one of the more common and well-supported ones for sure. http://arstechnica.com/features/2005/12/total-share/4/
Then, the ST didn't do so hot in the US, but here in Europe it was a well-known platform, nearly as popular as the Amiga. Roughly 75% of sold Atari ST's were sold in Europe.
Most famously, nearly every recording studio you'd care to name would have ST's in the control room well into the 90's and in some cases early 00's, due to their built-in MIDI capabilities.
My first gaming was done on my C64 and my bestie had an Atari 800 which we loved because cartridges meant fast load times, my C64 only had cassette tape. He then got an Atari ST which we also thought was a really good computer. This is in the UK. I was so jealous of that 800 back then, it was an impressive machine.
Edit: hey Ross, if you're out there, it's been 20 years, I still remember the good times!
I'm happy to hear that gaming works well, and apparently my decidedly older CPU (AMD Phenom X4 965) supports the technology as well, which is interesting to learn.
However, my main reason for needing Windows is for music recording and mixing via Cubase and other DAW software. This requires really low-latency access to my USB and PCI sound cards.
Does anyone have experience with running applications like this in a VM? Is it feasible?
I would be thrilled if I could finally move to Linux or FreeBSD entirely and just keep a Windows 7 VM, all sandboxed up with no Internet access, just for the music recording.
The simplest thing to do of course is just try it. You can install Windows in a VM without a product key and try it without limitation for something like 30 days.
That said, "really low latency" is relative. Device assignment will give you excellent I/O performance in terms of sending data to/from the device, particularly for bulk DMA transfers. Where you (likely) still pay a penalty is in terms of interrupt delivery, as on most chips these are still routed through the host even in "passthrough" mode. Even so, you're looking at delays at the outside on the tens of microseconds scale, depending on which microarchitecture you're on and how well tuned the fiddly bits of your hypervisor (presumably KVM for most users) are. Some configurations will be much better behaved here. If your audio hardware goes into polling mode (or can be made to do so) when it is recording you can potentially avoid the interrupts entirely.
(source: I work on Google Compute Engine, in particular on virtualized networking, historically on performance. Virtual interrupt delivery latency is a topic near and dear to my heart, although I don't know a damned thing about audio gear and its tolerances there :)
I personally install KXStudio[0] for guys just like you almost monthly. here's some appropriate[1] pornography[2]. Imagine being able to have a skype call with someone, and be able to record their vocals (provided they have a good mic) apply effects, and then record _and_ pipe the vocals back to the vocalist in realtime. Or make recording snippets of audio off youtube trivial. Or stream live instruments in other rooms over wifi/ethernet using only a raspberry pi and a good microphone. Or stream 8 of those tracks/instruments (on different devices) to one computer for mastering/final mixing. It's audio nerd nirvana, basically.
What you need:
Wine
Wineasio (for Wine support in FL Studio or others)
KXstudio
For lack of a response from someone who has actually tried getting ASIO working on such a setup, I suggest you just try it. Grab a Live CD and a spare external hard drive and just give it a whack. Passing through the sound cards should work. USB might be trickier if you're not just passing through a whole USB hub. But considering how new IOMMU is, I imagine the intersection between people who have tried it and people who use Cubase is incredibly tiny.
I have found that a lot of the audio editing software you can get away with quite low specs. I record my podcasts on a 8 year old laptop running Windows XP. It rarely connects to the internet and I record to a thumb drive that is backed up so there is nothing to lose on it. In this particular case something similar might be the best option. For pod-casting anything over 2 or 3 ms in my headphones throws me off my game.
Virtual CPUs with VTx can be made to run at essentially native speed for most workloads on most fairly modern CPUs. Memory performance can be subpar, although there's public documentation on techniques that can improve that: https://help.ubuntu.com/community/KVM%20-%20Using%20Hugepage...
I/O performance is a whole other story, dependent on many workload and hypervisor specific factors.
Really, the 965 supports it? I have the 970 and was under the impression the Phenom II's in general just don't support it. Can you show me where you found that information?
