Re: Seymour remembered

From: Bob Lidral (boblidral_at_yahoo.com)
Date: 10/16/03 

Date: Thu, 16 Oct 2003 05:23:38 GMT

"Jan C. Vorbrüggen" wrote:

> AFAIK, the processor has a "loop buffer" for twelve instructions. If you
> manage to get a loop into this buffer, the processor will run without
> needing to fetch instructions. So the compiler told you when it just
> missed fitting a loop into this buffer, so that you could think about
> somehow simplifying the code such that it would fit.
>
> Jan

I wasn't certain from the article which Fortran compiler they were
describing. They called it the COS Fortran compiler but whether the
COS (Chippewa Operating System?) mentioned was the predecessor to SCOPE
on CDC systems or an OS for subsequent Cray computers is not clear from
the article.

The CDC 6600 (and its derivatives in the 70- and 170-series) had a
60-bit word and a 7-word instruction "stack" analogous to cache
containing the 7 most recently executed instruction (60-bit) words.
Instructions were 15, 30, or 60 bits long so theoretically the stack
could hold between 7 and 28 instructions -- though in practice,
meaningful sequences of instructions contained a mix of each size and
included some padding. CDC also produced a model 6400 with the same
instruction set as the 6600 but without the stack and without multiple
processors.

The SCOPE (NOS/BE) and KRONOS (NOS) Fortran compilers would try very
hard to optimize inner loops to fit into the stack and would indicate
on the listing which loops actually did fit. I don't believe the
compiler printed any hints on how to modify source code though there
were suggestions in some of the reference manuals.

It's also possible the article refers to a compiler for one of the
Cray machines, in which case Jan C. Vorbrüggen's answer is probably
more relevant.

Bob Lidral
l i d r a l at a l u m dot m i t dot e d u

Relevant Pages

  • Re: I want to learn forth but...
    ... words in the loop are executed. ... executing a word copies new instructions on the top of the stack. ... where to fetch the next instruction, or it will execute the loop over ...
    (comp.lang.forth)
  • Re: Implement memcmp function with help from gccs inline asm
    ... >> number of iterations by duplicating the loop body. ... > due to how the C compiler interprets unsigned vs signed integers. ... test ecx, ecx ... My version of this code has 2 fewer instructions in the loop body. ...
    (comp.lang.asm.x86)
  • Re: LOOP - Why so slow?
    ... of complicated ones with multiple side effects, so "loop" wasn't ... loop instructions that branch forward). ... This is no big deal for a compiler. ... BS that Intel made up. ...
    (comp.lang.asm.x86)
  • Re: Measurement Accuracy & ANOVA
    ... the number of instructions that had to be executed. ... and in the compiler -- reduce the worth of those direct measures, ... done in that loop that has any variables that are used later, ... For a practical benchmark, you need something that looks like a ...
    (sci.stat.math)
  • Re: Bison/Flex To ByteCode
    ... I would assume there are 2 stacks my compiler ... How would this instruction get handled and represented on a stack ... integers get pushed to the stack with a special opcode that simply ... -> jump instructions, call instructions, comparisons, conditonal jumps, ...
    (comp.lang.cpp)