OSDev.org

Love4Boobies wrote:Sounds like a bottleneck to me and something that would only make sense with SMP, which we're already running away from in favour of NUMA.

Are we running away from SMP in favour of NUMA? I haven't observed such trend at all. Please elaborate.

DavidCooper wrote:I'd better throw my thoughts into the ring, though you probably know them already. No one's going to be interested in protection mechanisms of any kind ten years from now. The Turing Test will be passed within twelve months. Am I going to look silly, I wonder...

You are going to look silly. Back in the 70's we were going to have computers more intelligent than humans within five years.

The AI community seem even more gluttons for punishment than end of the world-ers in having their predictions proved wrong.

OSwhatever wrote:

Love4Boobies wrote:Sounds like a bottleneck to me and something that would only make sense with SMP, which we're already running away from in favour of NUMA.

Are we running away from SMP in favour of NUMA? I haven't observed such trend at all. Please elaborate.

AMD DirectConnect Architecture (All Opterons)
Intel QuickPath Interconnect (All Nehalem & later Xeons and Itaniums)

I'm willing to bet that SPARC, POWER and zArchitecture are all also NUMA.

Owen wrote:

OSwhatever wrote:

Love4Boobies wrote:Sounds like a bottleneck to me and something that would only make sense with SMP, which we're already running away from in favour of NUMA.

Are we running away from SMP in favour of NUMA? I haven't observed such trend at all. Please elaborate.

AMD DirectConnect Architecture (All Opterons)
Intel QuickPath Interconnect (All Nehalem & later Xeons and Itaniums)

I'm willing to bet that SPARC, POWER and zArchitecture are all also NUMA.

Right. The bottleneck isn't noticeable for 2 or 4 cores but it will definitely be noticeable when we have 128 cores using the same memory.

berkus wrote:

Solar wrote:I remember when I quite recently re-visited an Amiga website, where they still argue that PowerPC is the better architecture etc. etc. yadda yadda. I dropped something in the "I told you so" ballpark, and someone flamed me like, "yeah, as if you could foresee that PPC wouldn't be the way to go".

Mmm, wait, what? PPC is still alive and is powering your Xboxen and PS3s all over the world. Selling tens of millions of units worldwide, I don't call this dead.

PPC as a CPU family isn't dead. But the decision of Amiga Inc. to try and build a desktop PC system on the PPC base was a stillborn.

Solar wrote:PPC as a CPU family isn't dead. But the decision of Amiga Inc. to try and build a desktop PC system on the PPC base was a stillborn.

What? http://en.wikipedia.org/wiki/AmigaOne_X1000

A much delayed machine based upon an EOL processor family by PA Semi, who these days are Apple's ARM device design division. I wonder how viable it will prove to be.

(That nobody has bought PWRficient of PA Semi is a big shame IMO. It looks like a nice little processor)

Brynet-Inc wrote:

Solar wrote:PPC as a CPU family isn't dead. But the decision of Amiga Inc. to try and build a desktop PC system on the PPC base was a stillborn.

What? http://en.wikipedia.org/wiki/AmigaOne_X1000

Uh-huh. Yeah. That baby will surely put them back on the big screen.

Listen, I won't start that kind of discussion here. Amiga is a dead fringe market, and I'll bet a month's salary against a box of beer that PPC Amigas will never get into the five-digit user range. (They might get a four-digit "units sold", if every single Amiga user left will buy at least one of the things. Chances of attracting anybody but a hardcore Amiga fan with that kind of hardware is nil.) That doesn't mean I'm happy with it, but it's a fact.

I agree with Solar---very few people besides old Amiga fans will be interested in it. However, it still looks interesting enough for OS hacking---certainly better than Xbox if you're interested in Power(PC) and certainly less expensive than an IBM server.

I wasn't arguing about the popularity of Amiga, I was just noting that a new PowerPC Amiga is in existence and there was some hype about it.. honestly don't care about Amiga's or the PowerPC.

DavidCooper wrote:Am I going to look silly, I wonder...

Yes you are. People have been making that prediction since at least the mid seventies.

June 24:

Casm wrote:

DavidCooper wrote:I'd better throw my thoughts into the ring, though you probably know them already. No one's going to be interested in protection mechanisms of any kind ten years from now. The Turing Test will be passed within twelve months. Am I going to look silly, I wonder...

You are going to look silly. Back in the 70's we were going to have computers more intelligent than humans within five years.

The AI community seem even more gluttons for punishment than end of the world-ers in having their predictions proved wrong.

October 1:

Casm wrote:

DavidCooper wrote:Am I going to look silly, I wonder...

Yes you are. People have been making that prediction since at least the mid seventies.

So wait... you necro'd this thread to post an abridged version of the same thing you did four months ago?

I wonder what people would say if I said I was going to make a kernel in C++ back in the 80s. Languages mature and new generations of programmers turns up which will naturally pick up managed code languages. These languages do help productivity and easier to make stable code.

A few questions with managed code I do have. How can swapping to hard drive work without an MMU? How will an expandable stack work? Obviously an MMU has some advantages.

OSwhatever wrote:I wonder what people would say if I said I was going to make a kernel in C++ back in the 80s. Languages mature and new generations of programmers turns up which will naturally pick up managed code languages. These languages do help productivity and easier to make stable code.

A few questions with managed code I do have. How can swapping to hard drive work without an MMU? How will an expandable stack work? Obviously an MMU has some advantages.

Managed code can only be as stable as the virtual machine it runs on, and since that must itself run directly on the hardware I don't see the point. You are just giving yourself (or somebody else) an extra layer of code to write.

How does having an operating system, which runs an operating system, which runs applications, manage to be an improvement upon an operating system which runs applications directly?

OSwhatever wrote:Languages mature and new generations of programmers turns up which will naturally pick up managed code languages.

Where appropriate. But non-managed languages will be here to stay, if not forever, then for a long while.

Java was 1995. C# was 2001. That's a decade-and-a-half and a decade since, respectively. Eons in the IT world. Yet still a sizeable percentage of all new software projects are done in C, C++ or other unmanaged languages, not even speaking of existing code base.

Tossing the idea of unmanaged code because it's no longer "vogue" or "state-of-the-art" (I challenge the latter) certainly feels a bit hasty here. Hum-homm...