Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.arch > #115040

Re: Combining Practicality with Perfection

Show all headers | View raw

On 2/12/2026 11:09 AM, MitchAlsup wrote:
> 
> anton@mips.complang.tuwien.ac.at (Anton Ertl) posted:
> 
>> MitchAlsup <user5857@newsgrouper.org.invalid> writes:
>>> {{One can STILL argue whether
>>> deNormals were a plus or a minus in IEEE}}
>>
>> I am surprised to read that from you, who has always written that
>> denormals can be implemented cheaply and efficiently in hardware.  The
>> additional hardware cost (or the cost of trapping and software
>> emulation) has been the only argument against denormals that I ever
>> encountered.
> 
> It is only after IEEE 754-2008 came with FMAC that deNormals became
> a low cost addition. {And that has been my point--you seem to have
> forgotten the -2008 part or the argument}

And, can note, this is assuming that one actually pays the cost of 
native hardware FMAC.


Well, and the secondary irony that it is mainly cost-added for FMUL, 
whereas FADD almost invariably has the necessary support hardware already.

But:
   FMUL is expensive operation + cheap normalizer (if no denormals);
   FADD is cheap operation with expensive normalizer.

FMAC then is gluing the costs of the two units together, but:
   With roughly the latency of both;
   The need to be significantly wider internally to deal with some cases.

So, FMAC is a single unit that costs more than both units taken 
separately, and with a higher latency.



FMAC does suddenly get a bit cheaper if its scope is limited to 
FP8*FP8+FP16, but this operation is a bit niche.


This one makes a lot of sense for NN's, but haven't gotten my NN tech 
working good enough to make a strong use-case for it.

Where, in terms of algorithmic or behavioral complexity relative to 
computational efficiency, NN's are significantly behind what is possible 
with genetic algorithms or genetic programming.


So, for computational efficiency of the result:
   Hand-written native code, best efficiency;
   Genetic algorithm, moderate efficiency;
   Neural Net, very inefficient.

The merit of NNs could then be if one could make them adaptive in some 
practical way:
   Native code: No adaptation apart from specific algos;
   Genetic algorithms: Only when running the evolver, static otherwise;
   NN's: Could be made adaptable in theory, usually fixed in practice.

And, adaptation process:
   Native: None, maybe manual fiddling by programmer;
   Genetic algo: Initially very slow, gradually converges on answer;
   NNs, via generic algorithm: Slow, but converges toward an answer;
   NNs, via backprop: Rapid adaptation initially, then hits a plateau.

Backprop is seemingly prone to get stuck at a non-optimal solution, and 
then is hard pressed to make any further progress. Seemingly isn't 
really able to "fix" any obvious structural defects once it hits a 
plateau, but can sometimes jump up or down between various nearby 
options (when obvious suboptimal patterns persist).

Some tricks that work with GA-NN's don't really work with backprop, and 
my initial attempts to glue GA handling onto backprop have not been 
effective. Also it seems to need at least FP16 weights for training to 
work effectively (though, one other option being FP8 with a bias 
counter; but this is effectively analogous to using a non-standard 
S.E4.M11 format).


Seemingly, my own efforts are getting stuck at the level of very 
inefficiently solving very mundane issues, nowhere near the success 
being seen by more mainstream efforts.

Nor, as of yet, even anything particularly interesting...



Had started making some progress in other types of areas though, for 
example:
Figured out a practical way to get below 16kbps for audio...

By using 8kHz ADPCM and then using lookup table and reversed LZ search 
trickery to make the audio more LZ compressible (without changing the 
storage format).

Or, basically, ADPCM encoding strategy like:
   Lookup a match for the last 4 bytes;
   Look for the longest backwards match (last N bytes);
   Evaluate if the next byte for pattern is within an error limit;
   Select based on combination of error and length
     Longer matches permit more error than shorter ones.
   Check a pattern table,
     seeing if anything is within an acceptable error limit;
     Use pattern if so.
   Else:
     Figure out best-match for next 6 samples,
     using this to encode next 4 samples (1 byte).

Was able to get around a 20-30% reduction in bitrate, or around 12 kbps 
typical, before loss of audio quality becomes unacceptable (starts 
breaking down in obvious ways).


Did version for 4-bit ADPCM, which can get a roughly similar reduction, 
or around 24 kbps, though trying to push it much lower makes 2-bit ADPCM 
preferable.

A slightly higher reduction rate is possible if the baseline sample-rate 
is increased to 16kHz, but still doesn't get as low as when using 8 kHz.



Note that it is possible to just use a pattern table directly to give an 
equivalent of 8 kbps ADPCM (each byte encoding an index into an 8-sample 
table, which is then decoded as 2-bit ADPCM), but the audio quality is 
unacceptably poor (for much of any use-case).


Though, all this was mostly dusting off an experiment from last year, 
and putting it to use in my packaging tool (inside BGBCC).

Mostly it is a case of:
It is "good enough" to at least allow for optional super-compression of 
ADPCM without breaking the existing decoders.


...



>>
>> - anton

Back to comp.arch | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Combining Practicality with Perfection John Savard <quadibloc@invalid.invalid> - 2026-01-31 18:28 +0000
  Re: Combining Practicality with Perfection jgd@cix.co.uk (John Dallman) - 2026-01-31 22:36 +0000
  Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-05 01:57 +0000
    Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-05 06:06 +0000
      Re: Combining Practicality with Perfection Robert Finch <robfi680@gmail.com> - 2026-02-05 08:10 -0500
      Re: Combining Practicality with Perfection scott@slp53.sl.home (Scott Lurndal) - 2026-02-05 14:50 +0000
      Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-05 23:58 +0000
        Re: Combining Practicality with Perfection jgd@cix.co.uk (John Dallman) - 2026-02-06 16:37 +0000
          Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-06 22:32 +0000
            Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-06 22:34 +0000
            Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-07 00:57 +0000
              Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-07 02:45 +0000
          Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-06 22:40 +0000
        Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-09 22:46 +0000
          Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-10 03:49 +0000
            Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-10 19:10 +0000
              Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-11 23:04 +0000
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-11 23:55 +0000
                Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-12 06:04 +0000
                Re: Combining Practicality with Perfection quadibloc@invalid.com (John Savard) - 2026-10-24 00:22 +0000
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-10-24 19:32 +0000
                Re: Combining Practicality with Perfection jgd@cix.co.uk (John Dallman) - 2026-10-25 06:00 +0000
                Re: Combining Practicality with Perfection quadibloc@invalid.com (John Savard) - 2026-10-25 17:38 +0000
                Re: Combining Practicality with Perfection Stephen Fuld <sfuld@alumni.cmu.edu.invalid> - 2026-10-25 22:43 -0700
                Re: Combining Practicality with Perfection jgd@cix.co.uk (John Dallman) - 2026-10-26 09:35 +0000
                Re: Combining Practicality with Perfection John Levine <johnl@taugh.com> - 2026-10-26 19:53 +0000
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-10-26 13:40 +0000
                Re: Combining Practicality with Perfection quadibloc@invalid.com (John Savard) - 2026-10-26 22:05 +0000
                Re: Combining Practicality with Perfection quadibloc@invalid.com (John Savard) - 2026-10-26 22:12 +0000
                Re: Combining Practicality with Perfection Stephen Fuld <sfuld@alumni.cmu.edu.invalid> - 2026-10-26 20:13 -0700
                Re: Combining Practicality with Perfection quadibloc@invalid.com (John Savard) - 2026-10-27 14:30 +0000
                Re: Combining Practicality with Perfection quadibloc@invalid.com (John Savard) - 2026-10-26 22:09 +0000
                Re: Combining Practicality with Perfection EricP <ThatWouldBeTelling@thevillage.com> - 2026-10-26 18:20 -0400
                Re: Combining Practicality with Perfection quadibloc@invalid.com (John Savard) - 2026-10-27 03:00 +0000
                Re: Combining Practicality with Perfection antispam@fricas.org (Waldek Hebisch) - 2026-10-26 22:51 +0000
                Re: Combining Practicality with Perfection John Levine <johnl@taugh.com> - 2026-10-25 18:47 +0000
                Re: Combining Practicality with Perfection David Schultz <david.schultz@earthlink.net> - 2026-02-11 17:57 -0600
                Re: Combining Practicality with Perfection John Levine <johnl@taugh.com> - 2026-02-12 00:04 +0000
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-12 02:04 +0000
                Re: Combining Practicality with Perfection John Levine <johnl@taugh.com> - 2026-02-12 03:31 +0000
                Hex floating point (was: Combining Practicality with Perfection) Thomas Koenig <tkoenig@netcologne.de> - 2026-02-15 16:53 +0000
                Re: Combining Practicality with Perfection anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2026-02-12 08:45 +0000
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-12 17:09 +0000
                Re: Combining Practicality with Perfection Terje Mathisen <terje.mathisen@tmsw.no> - 2026-02-14 18:03 +0100
                Re: Combining Practicality with Perfection BGB <cr88192@gmail.com> - 2026-02-19 02:10 -0600
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-19 17:30 +0000
                Re: Combining Practicality with Perfection Michael S <already5chosen@yahoo.com> - 2026-02-19 20:15 +0200
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-19 19:55 +0000
                Re: Combining Practicality with Perfection BGB <cr88192@gmail.com> - 2026-02-20 05:08 -0600
                Re: Combining Practicality with Perfection Terje Mathisen <terje.mathisen@tmsw.no> - 2026-02-20 15:22 +0100
                Re: Combining Practicality with Perfection BGB <cr88192@gmail.com> - 2026-02-20 15:26 -0600
                Re: Combining Practicality with Perfection Paul Clayton <paaronclayton@gmail.com> - 2026-02-21 20:27 -0500
                Re: Combining Practicality with Perfection Michael S <already5chosen@yahoo.com> - 2026-02-12 10:49 +0200
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-12 17:10 +0000
                Re: Combining Practicality with Perfection Michael S <already5chosen@yahoo.com> - 2026-02-12 23:38 +0200
                Re: Combining Practicality with Perfection Terje Mathisen <terje.mathisen@tmsw.no> - 2026-02-12 14:28 +0100
                Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-12 05:51 +0000
                Re: Combining Practicality with Perfection Michael S <already5chosen@yahoo.com> - 2026-02-12 10:53 +0200
                Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-12 18:40 +0000
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-12 20:02 +0000
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-12 20:33 +0000
                Re: Combining Practicality with Perfection Stephen Fuld <sfuld@alumni.cmu.edu.invalid> - 2026-02-11 19:50 -0800
                Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-12 05:55 +0000
                Re: Combining Practicality with Perfection Stephen Fuld <sfuld@alumni.cmu.edu.invalid> - 2026-02-12 08:40 -0800
                Re: Combining Practicality with Perfection scott@slp53.sl.home (Scott Lurndal) - 2026-02-12 16:51 +0000
                Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-12 18:49 +0000
                Re: Combining Practicality with Perfection scott@slp53.sl.home (Scott Lurndal) - 2026-02-12 15:54 +0000
                Re: Combining Practicality with Perfection Stephen Fuld <sfuld@alumni.cmu.edu.invalid> - 2026-02-12 08:52 -0800
                Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-12 18:37 +0000
                Re: Combining Practicality with Perfection antispam@fricas.org (Waldek Hebisch) - 2026-02-14 16:33 +0000
                Re: floating pain, Combining Practicality with Perfection John Levine <johnl@taugh.com> - 2026-02-14 18:12 +0000
                Re: floating pain, Combining Practicality with Perfection antispam@fricas.org (Waldek Hebisch) - 2026-02-17 01:16 +0000
                Re: floating pain, Combining Practicality with Perfection John Levine <johnl@taugh.com> - 2026-02-17 01:24 +0000
                Re: floating pain, Combining Practicality with Perfection antispam@fricas.org (Waldek Hebisch) - 2026-02-17 16:21 +0000
                Re: floating pain, Combining Practicality with Perfection John Levine <johnl@taugh.com> - 2026-02-17 19:20 +0000
                Re: floating pain, Combining Practicality with Perfection antispam@fricas.org (Waldek Hebisch) - 2026-02-17 19:52 +0000
                Re: floating pain, Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-17 19:09 +0000
                Re: Combining Practicality with Perfection jgd@cix.co.uk (John Dallman) - 2026-02-15 14:37 +0000
                Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-17 18:57 +0000
                Re: Combining Practicality with Perfection antispam@fricas.org (Waldek Hebisch) - 2026-02-17 20:43 +0000
                Re: Combining Practicality with Perfection MitchAlsup <user5857@newsgrouper.org.invalid> - 2026-02-18 00:50 +0000
                Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-18 08:52 +0000
                Re: Combining Practicality with Perfection Robert Finch <robfi680@gmail.com> - 2026-02-18 08:40 -0500
                Re: Combining Practicality with Perfection quadi <quadibloc@ca.invalid> - 2026-02-20 08:14 +0000
                Re: Combining Practicality with Perfection Thomas Koenig <tkoenig@netcologne.de> - 2026-02-19 18:49 +0000

csiph-web