From: sfuld@alumni.cmu.edu.invalid
On 9/4/2025 1:43 PM, Scott Lurndal wrote:
> Stephen Fuld writes:
>> On 9/4/2025 12:06 PM, EricP wrote:
>>> Stephen Fuld wrote:
>>>> On 9/4/2025 10:19 AM, EricP wrote:
>>>>> BGB wrote:
>>>>>> On 9/3/2025 9:42 PM, EricP wrote:
>>>>>>> MitchAlsup wrote:
>>>>>>>>
>>>>>>>> However, I also found that STs need an immediate and a
>>>>>>>> displacement, so,
>>>>>>>> Major == 0b'001001 and minor == 0b'011xxx has 4 ST instructions with
>>>>>>>> potential displacement (from D12ds above) and the immediate has
>>>>>>>> the size of the ST. This provides for::
>>>>>>>> std #4607182418800017408,[r3,r2<<3,96]
>>>>>>>
>>>>>>> Compare and Branch can also use two immediates as it
>>>>>>> has reg-reg or reg-imm compares plus displacement.
>>>>>>> And has high enough frequency to be worth considering.
>>>>>>>
>>>>>>
>>>>>> Can be done, yes.
>>>>>> High enough frequency/etc, is where the possible debate lies.
>>>>>>
>>>>>>
>>>>>> Checking stats, it can effect roughly 1.9% of the instructions.
>>>>>> Or, around 11% of branches; most of the rest being unconditional or
>>>>>> comparing against 0 (which can use the Zero Register). Only a
>>>>>> relative minority being compares against non-zero constants.
>>>>>
>>>>> The only instruction usage stats I have are from those VAX papers:
>>>>> A Case Study of VAX-11 Instruction Set Usage For Compiler Execution,
>>>>> 1982
>>>>>
>>>>> That shows about 12% instructions are conditional branch and 9% CMP.
>>>>> That says to me that almost all Bcc are paired with a CMP,
>>>>> and very few use the flags set as a side effect of ALU ops.
>>>>
>>>> OK, but does this tell you how many of the CMPs are to a value of
>>>> zero? I expect these to be a significant enough percentage to skew
>>>> your analysis.
>>>
>>> Looking at
>>> Measurement and Analysis of Instruction Use in VAX 780, 1982
>>>
>>> VAX had a TST instruction which was the same as CMP src,#0.
>>> TST has < 2% usage while CMP 10-12%.
>>
>> Thanks. That's interesting. So perhaps ~15% of all compares are to
>> zero. I would have expected higher.
>
> Kinda depends on the compilers used for the workload. I suspect
> that those workloads were mostly COBOL and FORTRAN and maybe BLISS-32
> or MACRO-32.
>
> Without a heavy C or C++ workload, the need to check for NULL
> pointer is rare.
>
> Unlike Unix, successful returns from VMS system and library calls
> was SS$_NORMAL, which had the value 1 rather than zero, which
> would probably also reduce the uses of TST to check for zero.
>
Good call.
--
- Stephen Fuld
(e-mail address disguised to prevent spam)
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
