OT: Yet Another Unhappy Customer for Vista

Is it? Preemptive is pretty easy to do, and frees the applications from being responsible for giving up time, or from hanging the system.

John

Reply to
John Larkin
Loading thread data ...

Only until you trollfeeders let him go.

Thanks! Rich

Reply to
Rich Grise

We gotta have a little fun along the way too. I doubt he is going to go away just because he gets ignored today, tomorrow, or even the next day. You have seen the type before. Like I said, Phil has tripped my filter in within the last 24hrs and I have been ignoring his posts for months.

Jim

Reply to
James Beck

No, it isn't. Multi-threading was an OS thing. It was around long before Intel decided to do it at the instruction level, which is what hyperthreading is about.

Reply to
JackShephard

Multiple pipelines have been around for years.

Hell, the PI had like 5.

Reply to
JackShephard

The "cell" processor is massively superscalar.

Look it up. It is what is in the PS3 x 4.

Reply to
JackShephard

You're so branch predictable, it is pathetic. :-]

Reply to
JackShephard

That depends ENTIRELY on which MOBO it gets put on, under which chipset, and with which memory.

Reply to
JackShephard

The Cell CPU and architecture is very likely in our PC future.

Reply to
JackShephard

for

Get him the source code for "MAME32". Then he can emulate a huge array of old Upright gaming platform processors and game environments.

Awesome program. Great step up from what you have him tinkering with now.

Reply to
JackShephard

formatting link

Reply to
JackShephard

The chip was recommended to me, though I have no immediate application. The guy recommending was doing robotics. He found the multiple processors perfectly suited to synchronizing movement in multiple axes.

I like their notion of interrupts: just dedicate a processor!

8 x 20MIPS processors on a chip is pretty cool.

Cheers, James Arthur

Reply to
James Arthur

Many years ago there was a microprocessor that did it. There were two sets of registers and only one ALU. Every other instruction was from one task and the other every from the other. The design didn't need any pipelining because it always had a whole instruction cycle for things to get settled in before they were needed again. It didn't catch on for the simple reason that it was too weird to use.

The later x86 processors all have multiple ALU sections in them. They have special hardware for address calculation. This means that unlike the 8086, the main integer ALU may not be fully busy. Very few program keep the floating point section busy. As a result there has been some hardware in the processor that will be idle a good fraction of the time. It makes sense to try to put this stuff to good use.

Reply to
MooseFET

Only 256 processes, whimp! :)

Reply to
MooseFET

MassivelyWrong, once again. The motherboard and chipset have NOTHING to do with HT performance. Grow up, Dimmy. You're way out of your league, as minor as it may be.

--
  Keith
Reply to
krw

No, Dimmy, it's not. It's massively parallel. The processors have a rather mundane number of E-units.

I know what it is (and isn't), Dimbulb.

--
  Keith
Reply to
krw

PKB! You are simply, well, simple!

--
  Keith
Reply to
krw

As usual, you haven't a clue what you're talking about.

--
  Keith
Reply to
krw

As usual, you don't have a clue what you're talking about. Sheesh, grow up, Dimbulb!

--
  Keith
Reply to
krw

for

"Superscalar" IS a paralleling schema, idiot!

You have zero.

You very likely do not know anything about it.

Reply to
JackShephard

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.