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

Groups > comp.lang.prolog > #15636

Parallel π-WAM: 1.7 Giga Lips on a CPU (Was: Dogelog Player: 11.4 Giga Lips with a Budget Laptop)

Show all headers | View raw

Hi,

We recently implemented a parallel π-WAM
on a GPU backend and could demonstrate an
estimated 11.4 Giga Lips. In this post we
report a further experiment, this time
presenting a parallel π-WAM on a CPU backend,
that can lift specialized Prolog, currently
to 1.7 Giga Lips performance.

Having an excess number of threads is a
bad idea. What if we do context switching
on our own? With this approach we could
bring down the execution time of 128 Hack
VMs by 33%. We estimate for the test which
had 11.4 GLips on the GPU, that we reach
1.7 GLips on the CPU.

Bye

See also:

Parallel π-WAM: 1.7 Giga Lips on a CPU
https://medium.com/2989/8a984e75af44

Mild Shock schrieb:
> Hi,
> 
> At the end of 2025 we acquired a couple of AI
> Laptops , that were still cheap, since RAM prices
> had not yet rocketed. The intend was to tap into
> the Copilot+ certified hardware, and shave off
> some of the TOPS to do Prolog inferencing. Amazingly
> our π-WAM can churn 11.4 GIGA LIPS.
> 
> GPUs have evolved form lock-step to independent
> thread scheduling. This made it possible to port
> the Hack VM variant, that forms the basis for
> our π-WAM, to WebGPU computer shaders. Using
> NUM_SHADERS = 4096 we could produce 11.4 Giga Lips
> on a Ryzen AI 7 350 w/ Radeon 860M.
> 
> Bye
> 
> See also:
> 
> Dogelog Player: 11.4 Giga Lips with a Budget Laptop
> https://medium.com/2989/899b0d5c027b
> 
> Mild Shock schrieb:
>> Dear All,
>>
>> We are happy to announce a new edition of
>> the Dogelog Player:
>>
>> - Unicode 17.0:
>> We lifted the Dogelog Player for Java build
>> to JDK 26 and regenerated our Unicode database
>> for category and number value for the JavaScript
>> and Python build. The supported version is now
>> 17.0 and we managed somehow to reduce the footprint
>> from 16848 words to 15000 words.
>>
>> - Emulating π-WAM:
>> The new library(edge/brainfog) permits the
>> execution of certain Prolog goals in a π-WAM
>> backend. The concept of a π-WAM embraces a
>> fusion of a processs (π) calculus and a Warren
>> Abstract Machine (WAM). Optimized for speed the
>> WAM is very primitive and currently only supports
>> the ‘$SEQ’/2 control construct, plus rudimentary
>> 32-bit integer arithmetic and between/3.
>>
>> - Executing π-WAM:
>> Both emulate/1 and execute/1 compile into an
>> identical instruction stream. In the virtual
>> machine the instructions are 32-bit inspired by
>> combinations of the 16-bit A and C instructions
>> from the original Hack. The resulting π-WAM
>> currently doesn’t support a stack, a trail
>> or choice points but can nevertheless
>> do backtracking.
>>
>> Have Fun!
>>
>> Jan Burse, July 07, 2026, https://www.herbrand.ai/
> 

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

Thread

ANN: Dogelog Player 2.2.3 (Introducing π-WAM) Mild Shock <janburse@fastmail.fm> - 2026-07-07 19:26 +0200
  Dogelog Player: 11.4 Giga Lips with a Budget Laptop (Re: ANN: Dogelog Player 2.2.3 (Introducing π-WAM)) Mild Shock <janburse@fastmail.fm> - 2026-07-08 18:39 +0200
    Parallel π-WAM: 1.7 Giga Lips on a CPU (Was: Dogelog Player: 11.4 Giga Lips with a Budget Laptop) Mild Shock <janburse@fastmail.fm> - 2026-07-15 18:08 +0200
      Parallel π-WAM: JavaScript Workers as CPU Backend (Was: Parallel π-WAM: 1.7 Giga Lips on a CPU) Mild Shock <janburse@fastmail.fm> - 2026-07-20 18:50 +0200
        Parallel π-WAM: An Interleaved Synchronous Emulator (Re: Parallel π-WAM: JavaScript Workers as CPU Backend) Mild Shock <janburse@fastmail.fm> - 2026-07-25 03:51 +0200

csiph-web