Ways of Implementing a System Call

Re:Ways of Implementing a System Call by Paperweight on Friday January 17, 2020 @03:42AM (#22072955)

I wonder if mainstream CPUs will ever get back to being truly programmable with software instructions, instead of running off and speculating on what you might do next.

[ Parent | on HN ]

Re:Ways of Implementing a System Call by pjc50 on Friday January 17, 2020 @05:41AM (#22073558)

Most people don't want a much slower processor. As soon as you have a pipeline the concepts of "now" and "next" become smeared out in time.

[ Parent | on HN ]

Re:Ways of Implementing a System Call by Gladdyu on Friday January 17, 2020 @04:33AM (#22073231)

Intel tried that with Itanium. It didn't go well.

[ Parent | on HN ]

Re:Ways of Implementing a System Call by wolfgke on Friday January 17, 2020 @05:34AM (#22073529)

Also keep the delay slots of MIPS in the back of your mind, which with the evolution of MIPS became a burden.

[ Parent | on HN ]

Re:Ways of Implementing a System Call by saagarjha on Friday January 17, 2020 @06:02AM (#22073645)

They’re a burden to work with regardless of whether they match how pipelines work.

[ Parent | on HN ]

Re:Ways of Implementing a System Call by jbverschoor on Friday January 17, 2020 @09:56AM (#22075065)

Would also save some watts I guess

[ Parent | on HN ]