New Challenges: Resilient ML Systems [Pushing the Frontier] (Re: Prolog has become the Useful Fool [Magic Square of 10 x 10])

From	Mild Shock <janburse@fastmail.fm>
Newsgroups	comp.lang.prolog
Subject	New Challenges: Resilient ML Systems [Pushing the Frontier] (Re: Prolog has become the Useful Fool [Magic Square of 10 x 10])
Date	2025-10-21 00:43 +0200
Message-ID	<10d6dvc$aub2$3@solani.org> (permalink)
References	<10d62oi$anvc$1@solani.org> <10d63pl$aokp$1@solani.org>

Show all headers | View raw

Hi,

On small clusters, a failed GPU process
usually crashes visibly. On large TPU systems,
you can get “silent” or “partial” failures
where a chip is alive but unhealthy —

producing NaNs, stalling on the interconnect,
or returning corrupted data that’s
not immediately flagged.

At hyperscale, training becomes a kind of
“distributed computer architecture problem,”
and engineers do end up inventing their own
“parity-check”–like systems — not for

individual bits (as in DRAM ECC), but for
entire tensors, gradients, and replicas.

See also:

Gemini 2.5: Pushing the Frontier
https://arxiv.org/abs/2507.06261

Bye

Mild Shock schrieb:
> Hi,
> 
> Prolog has become the Useful Fool.
> Over the years it was trained that
> constraints are importanter than
> 
> anything else. Even this talk blends
> out natural language processing:
> 
> Manuel Hermenehgildo - 50 Years of Prolog and Beyond
> https://prologyear.logicprogramming.org/videos/PrologDay_Session_1_talk.mp4
> 
> And if you give them some IBM Roadrunner
> precursor, all they can do, is play
> some stupid constraint games:
> 
> Parallel local search for solving Constraint Problems on the Cell 
> Broadband Engine (Preliminary Results)
> https://www.researchgate.net/publication/220481722
> 
> And mentioning in passing how useless
> blunt constraint solving is as implemented
> by most CLP(X), can't even solve
> 
> magic square of 10 x 10.
> 
> Bye
> 
> Mild Shock schrieb:
>> Hi,
>>
>> With their stupid fixation on Emacs Prolog
>> missed a Billion Dollar Business Model.
>>
>> Now people will start paying 4'000 USD per
>> Year and GPU to Red Hat to have some AI
>> model deployment and training enviroment:
>>
>> DGX Spark Arrives at SpaceX
>> https://www.youtube.com/watch?v=peaIkB0NzS0
>>
>> But these environments are nothing then
>> Python virtual environments. Can Prolog not
>> also push some evaluation graphs to
>>
>> a GPU or NPU? Well it seems with its current
>> philosophy only if there is a spare Key
>> Combination in Emacs for it.
>>
>> Bye
>

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

Thread

Prolog missed a Billion Dollar Business Model [DGX Spark ] Mild Shock <janburse@fastmail.fm> - 2025-10-20 21:32 +0200
  Prolog has become the Useful Fool [Magic Square of 10 x 10] (Was: Prolog missed a Billion Dollar Business Model [DGX Spark ]) Mild Shock <janburse@fastmail.fm> - 2025-10-20 21:50 +0200
    New Challenges: Resilient ML Systems [Pushing the Frontier] (Re: Prolog has become the Useful Fool [Magic Square of 10 x 10]) Mild Shock <janburse@fastmail.fm> - 2025-10-21 00:43 +0200

csiph-web