[RC5] core questions
blitz at macronet.net
Wed Mar 13 15:52:38 EST 2002
Then was the AMD core optimized for the available instruction set? Im sorry
i missed the discussion on it previously.
Checking, I see its using the Core #5 RG Risc Rotate -II
By the way, I dropped a line to Jasonp about his core, and he is quite
informative, he also said I'm only the second person who's contacted about
his work since he wrote it. (1998)
I thanked him for his work on the project..an attaboy every now and then to
our volunteers, specially the indispensable coders is a good idea IMHO.
He's at jasonp at boo.net
At 20:59 3/12/02 -0500, you wrote:
>I think you are confusing MMX with 3dNow!. My AMD K6 166, 200 and 233
>chips were all MMX. They licensed it from Intel I believe. Then they
>came out with 3DNow which added extensions to make the chip better at 3D
>video and audio than a Pentium chip.
>According to AMD:
>Q: How are MMXT and 3DNow!T technology different?
>A: MMXT technology was developed to improve the integer-intensive
>operations used in 3D rendering at the back end of the graphics
>pipeline. However, by the time MMX technology reached the market, 3D
>graphics accelerator performance had outpaced most of what MMX could do.
>While MMX technology can enhance integer-intensive applications, today's
>state-of-the-art graphics accelerators provide an optimal performance
>boost for 3D rendering. 3DNow!T technology complements and enhances the
>performance of these accelerator products by speeding up the
>floating-point-intensive, front-end stages of the graphics pipeline.
So...if properly coded, they might enhance "floating-point-intensive" RC5
calcs too??? (Im asking, cause I'm not a coder)
>From: owner-rc5 at lists.distributed.net
>[mailto:owner-rc5 at lists.distributed.net] On Behalf Of blitz
>Sent: Tuesday, March 12, 2002 7:03 PM
>To: rc5 at lists.distributed.net
>Subject: Re: [RC5] core questions
>Interesting...I'm "assuming" AMD did add the necessary MMX sort of
>instruction set by the time things got to say 233 mhz or so....I recall
>they called it something different than MMX, but it was the same thing.
>box I'm on rite now, is a 233 amd, and I know it supports the mmx stuff.
>So despite mmx being a dud for video, it makes Dnet cores run
>I've always liked the AMD chips because of their risc architecture,
>something I got a bit familiar with when I had a SGI Indy. (wish I still
>At 14:45 3/12/02 -0500, you wrote:
> >I don't know that the turning of the crunchometere had anything to do
> >with speed. But MMX instructions do make a big difference. Also,
> >certain x86 chips didn't have instructions in hardware to do things
> >that we needed to do. Early AMDs didn't have a rotate instruction that
> >we needed to do, but Intel chips did, they were much faster than AMD
> >chips. MMX is just integer math handler.
> >On Mon, 11 Mar 2002, blitz wrote:
> > > I just did a processor changeout in a linux box on my lan. I had an
> > > old cyrix 166 mhz clone chip and replaced it with a pentium P5mmx at
> > > the same speed. (running Mandrake Linux 8.0 w/64mb memory) I see the
> > > core changed, and the speed of processing went way up in comparison
> > > to the old chip. The core running now is a "jasonp" I believe..
> > >
> > > Questions: Is there actually THAT much improvement in handling the
> > > RC5
> > task
> > > I should be seeing the little indicator turning that much faster?
> > >
> > > and, does the mmx functions actually contribute something to the
> > processing
> > > speed? I always thought they had something to do with video
> > > processing,
> > not
> > > raw processing power.
> > >
> > > In any case, I want to thank "jasonp" for a core thats visibly more
> > > powerful than the old cyrix one, amazing what just a little tweaking
> > can do huh?
> > >
> > > --
> > > To unsubscribe, send 'unsubscribe rc5' to
> > > majordomo at lists.distributed.net rc5-digest subscribers replace rc5
> > > with rc5-digest
> > >
> > >
> >To unsubscribe, send 'unsubscribe rc5' to
> >majordomo at lists.distributed.net rc5-digest subscribers replace rc5 with
>To unsubscribe, send 'unsubscribe rc5' to
>majordomo at lists.distributed.net rc5-digest subscribers replace rc5 with
>To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
>rc5-digest subscribers replace rc5 with rc5-digest
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5