Mike,
You're absolutely right about the difference between ISAs to assembly
language programmers. The latitude can be very wide. Long ago I found
Intel 8086 assembler incredibly painful compared to the other ISAs I'd
used a lot and many of those that came later. The 8086 irregular modes
and limited registers , stinking memory segmentation making branch
instruction selection existential and shams like movs that were slower
than loops of plain moves made writing fast and dense coding into an
arcane art. Later I did compiler back ends for the National NS32k and
that was a thing of beauty while still technically CISC. My first RISC
was the Motorola 88k and the only assembler I did was a rewrite of libm
transcendentals (like trig functions) that had been created by guys that
only wanted to win benchmarks, not to satisfy the military with a F18
simulator that happened to work. But the 88k had a ton of instructions
and had a kitchen sink feel. I worked with folks using MIPS and Intel
860: hard core RISC processors and both with exposed pipelines that
forced the programmer to schedule instructions, turning assembly
language coding into another arcane art. The vast majority of folks
using those chips were thankful for compilers that could do the right
thing and allow them to avoid the complexity of assembler.
In 2011 I fell in love with the TI MSP430 RISC ISA (not caring that it
was yesterday's news, to put it mildly). I spent a few years expecting
to have a ball writing structured assembler "any day now". It never
happened as there was always something more important. Now I feel modern
chips are so efficient and the ubiquity of gcc and other advanced
compilers means IMO the only situation needing assembly code is for a
coffee maker or something like a trampoline that cannot be expressed in
C (e.g. code to get you from one stack context to another) and hey,
that's what asm statements are for, right? :-)
Pete
On 8/16/26 18:25, Mike Lisanke wrote:
Pete, On the surface this appears to be pot-shots at RISC in-general.
What many (including me) didn't know is, RISC
has become a significant alternative to CISC. Why? Because orthogonal
register usage makes a lot of sense And
that Addressing modes solve A lot of problems. I was surprises at how
Complex the Reduced Instruction Set Computing
had become and the degree that ISC was a big topic in the computing
industry. I know that MIPS and other specs tell
us RISC doesn't compete well But that hasn't stopped all the small and
big low-power specialized processors from being
RISC... you posted a longish document so I've only scanned it at this
point. But it'd be interesting to banter about RISC v
CISC at some point.
On Sun, Aug 16, 2026 at 6:09 PM Pete Soper via TriEmbed
<[email protected]> wrote:
Here is a critique of RISCV that makes a number of points: some
objective and some subjective:
https://dmitry.gr/?r=06.%20Thoughts&proj=12.%20RV
<https://dmitry.gr/?r=06.%20Thoughts&proj=12.%20RV>
Here is a rebuttal challenging some of the points above and
filling in major blanks. This is from someone with a great deal of
embedded experience who lives in an economically challenged
location. IMO this blurb makes it much easier to see the far
reaching advantages of RISCV:
https://rvembedded.com/blog_post/12/
Pete
_______________________________________________
Triangle, NC Embedded Interest Group mailing list
To post message: [email protected]
List info:
http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
TriEmbed web site: https://TriEmbed.org <https://TriEmbed.org>
To unsubscribe, click link and send a blank message:
mailto:[email protected]?subject=unsubscribe
Searchable email archive available at
https://www.mail-archive.com/[email protected]/
--
Best regards, Mike
_______________________________________________
Triangle, NC Embedded Interest Group mailing list
To post message: [email protected]
List info: http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
TriEmbed web site: https://TriEmbed.org
To unsubscribe, click link and send a blank message:
mailto:[email protected]?subject=unsubscribe
Searchable email archive available at
https://www.mail-archive.com/[email protected]/
