News for hackers. Stuff that matters.
Re:ARM Cortex-A72 fetch and branch processing by gpderetta on Monday December 14, 2020 @07:44PM (#25425568)
To be constructive, I do believe there is a future beyond classic OoO, but dynamic scheduling is going to be part of it, although possibly not of instructions but of BBs or even microthreads. The papers you referencing are quire interesting as they validate some ideas I was idly thinking about. I still would call these designs OoO though, just extended.

The problem is that building a CPU is very expensive and high risk, going even a bit beyond the classic OoO can bankrupt you. Intel tried with Itanic and AMD almost ate its lunch; AMD tried with buldozer and shared execution units and almost collapsed; SUN tried with flock of chickens and is no more; Nvidia tried with denver, and well, I guess they had a lot of cash. Both AMD and Apple just made completely orthodox designs and were complete successes.

Evolution is going to be in small steps, unless a new design can prove to be 10x better than the existing ones which is very unlikely.

Re:ARM Cortex-A72 fetch and branch processing by Veedrac on Monday December 14, 2020 @08:05PM (#25425739)
> I do believe there is a future beyond classic OoO, but dynamic scheduling is going to be part of it, although possibly not of instructions but of BBs or even microthreads.

I agree, I find it hard to see how you could abandon dynamic scheduling without unreasonable sacrifices.

My point isn't to say otherwise, just that we see less innovation than we would otherwise were our ecosystem built to handle it. GPUs have benefited hugely from the ability to rapidly change the ISA and a lot of internal details between each generation, and for each competitor to have their own take and extensions (Apple even has its own API, to support its TBDR features). If we had that sort of freedom in CPUs, I'm sure we'd have something a lot better than four variants of practically-the-same-thing by four different companies.

Re:ARM Cortex-A72 fetch and branch processing by brandmeyer on Monday December 14, 2020 @08:41PM (#25426012)
> Nvidia tried with denver, and well, I guess they had a lot of cash

More likely: They recognized the failure and cut the program before losing too much money.

Re:ARM Cortex-A72 fetch and branch processing by gpderetta on Tuesday December 15, 2020 @04:56AM (#25428684)
True, also Denver wasn't bad, just not good enough.
Re:ARM Cortex-A72 fetch and branch processing by my123 on Tuesday December 15, 2020 @05:47AM (#25429014)
Denver was superseded by the Carmel architecture. (which is sold on Tegra Xavier today)

The NVIDIA CPU team is alive and well.

Re:ARM Cortex-A72 fetch and branch processing by brandmeyer on Wednesday December 16, 2020 @05:03PM (#25449409)
There's a big difference between cancelling a project and firing the team. I haven't seen any evidence that Carmel utilizes binary translation.