flat assembler
Message board for the users of flat assembler.

Index > Heap > Pentium is RISC?

Author
Thread Post new topic Reply to topic
beppe85



Joined: 23 Oct 2004
Posts: 181
beppe85
Hi, all!

I always have it take for granted that Pentium was a CISC architecture. But last semester my professor said Pentium was a RISC. Unfortunately I missed the opportunity to talk about this thing.

Pentium have a complex instruction set, complex addressing modes, a very weird format for instructions, etc...At least externally, it is CISC. I know that Pentium have a RISC core, so the RISCish instructions perform better than others, but I'm still confused a bit.

What do you know about this?

_________________
"I assemble, therefore I am"

If you got some spare time, visit my blog: http://www.beppe.theblog.com.br/ and sign my guestmap
Post 08 Feb 2005, 21:51
View user's profile Send private message Reply with quote
f0dder



Joined: 19 Feb 2004
Posts: 3170
Location: Denmark
f0dder
The short story is that since pentium, x86 processors have basically had RISC-like cores, with CISC frontends - and the frontends do a lot of fancy things besides "just" translating x86->"internal risc format", like register renaming, instruction re-ordering, ...
Post 09 Feb 2005, 00:27
View user's profile Send private message Visit poster's website Reply with quote
MCD



Joined: 21 Aug 2004
Posts: 604
Location: Germany
MCD
This is why, on "newer" PCs, simple CISC instructions like "lodsd" are even a bit slower than those 2 corresponding instructions:
Code:
mov eax,[esi]
add esi,4
    


What a strange world!

p.s.: loop-instructions are even worse!

_________________
MCD - the inevitable return of the Mad Computer Doggy

-||__/
.|+-~
.|| ||
Post 09 Feb 2005, 14:03
View user's profile Send private message Reply with quote
beppe85



Joined: 23 Oct 2004
Posts: 181
beppe85
So I'm still not have a RISC...I'm so sad...

Smile
Post 09 Feb 2005, 18:10
View user's profile Send private message Reply with quote
f0dder



Joined: 19 Feb 2004
Posts: 3170
Location: Denmark
f0dder
I don't really see why those instructions have to be slower, though - it should be possible for the x86 decoder to choose optimal microcode for those instructions too. But I guess there's hardwired logic, and some logic in microcode, to avoid too much hardwired logic? (that's the impression the AMD manuals give me, "directpath" vs. "vectorpath" instructions.)
Post 09 Feb 2005, 20:33
View user's profile Send private message Visit poster's website Reply with quote
MCD



Joined: 21 Aug 2004
Posts: 604
Location: Germany
MCD
I found out these string instructions vs. other ones and loop instructions vs conditional jumps speed differences with my self coded time stamp counter benchmark. They all appear on Pentium II, Pentium III and AMD Athlon XP. I've posted it somewhere, long time ago in the DOS section, but i don'tm find it right now.
Post 10 Feb 2005, 19:00
View user's profile Send private message Reply with quote
Display posts from previous:
Post new topic Reply to topic

Jump to:  


< Last Thread | Next Thread >
Forum Rules:
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum


Copyright © 1999-2020, Tomasz Grysztar. Also on YouTube, Twitter.

Website powered by rwasa.