Shop OBEX P1 Docs P2 Docs Learn Events
Prop2 versus Prop1 LMM speed — Parallax Forums

Prop2 versus Prop1 LMM speed

RaymanRayman Posts: 14,665
edited 2013-04-13 08:26 in Propeller 1
Is it too early to tell the speed increase going to P2 from P1?

I'm just playing around, but I think I see ~4X increase per clock, consistent with single clock instructions on P2 and 4 clock instructions on P1.

Do the new instructions give any speed increase?

Comments

  • David BetzDavid Betz Posts: 14,516
    edited 2013-04-12 06:45
    Rayman wrote: »
    Is it too early to tell the speed increase going to P2 from P1?

    I'm just playing around, but I think I see ~4X increase per clock, consistent with single clock instructions on P2 and 4 clock instructions on P1.

    Do the new instructions give any speed increase?
    Certainly rdlongc helps a lot since it only hits the hub one out of every four instructions. This is even better in CMM mode. Right now we're working on a relatively straight-forward port of the P1 compiler to P2. There may be other advantages to be gained from the new P2 features (like the CLUT) but they are likely to require significant compiler backend work and we don't have the time or resources to tackle that on the first pass.
  • RaymanRayman Posts: 14,665
    edited 2013-04-12 07:27
    Thanks. That's kinda what I figured. Sounds like it may get faster in the future then.
  • KyeKye Posts: 2,200
    edited 2013-04-13 07:49
    Should be able to hit above 160 MHz on the real chip. So, roughly a 10 X increase from the prop 1.
  • jazzedjazzed Posts: 11,803
    edited 2013-04-13 08:26
    Kye wrote: »
    Should be able to hit above 160 MHz on the real chip. So, roughly a 10 X increase from the prop 1.

    I seriously doubt that existing tools, except possibly SPIN, will ever achieve 10X performance increase without significant rewrite.
Sign In or Register to comment.