User:Lkcl

Wikipedia:WikiProject_Computing/Computer_hardware_task_force

This user participates in WikiProject Computing.

intro

Hi, I am Luke Kenneth Casson Leighton, i very rarely touch a shift key, i use my full name deliberately to avoid the american-biased search engine algorithms something along the lines of "oo look it's three words therefore it must be a name let's apply some brain-dead conventions to it".

i recently noticed that wikipedia has information on protocols with which i am familiar, from my work on Samba TNG. in particular, i noticed that a lot of people just... haven't got a clue: the dozens of different services running over a dozen or so totally distinct protocols that make up windows nt networking are all glommed in people's heads into one of three specific protocols: NetBIOS, NETBEUI or SMB aka CIFS.

this i find particularly irritating, because it's a bit like saying "hi i want an Internet Protocol and i want it to be on HTTP" - the sentence makes absolutely no sense, and will irritate the crap out of any experienced TCP/IP networking guru. it's also a bit like... oh, say... blaming DHCP for the problems associated with Spam, because you can obtain a networking address from a DHCP server, and as soon as you do so, lots of crap email arrives.

anyway, here's a list of the pages i've been involved in editing. i'm maintaining the list at the mo for my own evil purposes and i hope you find it useful too.

bug in Wikipedia

actually Wikimedia. https://phabricator.wikimedia.org/T399385 forced to unsubscribe from the bugreport as it got dangerous. PTSD triggered

lkcl page edits

not edits but TBD research

find places for these:

added to History of supercomputing so far, Espasa et al mention TI ASC and STAR100 as having 2x32-wide SWAR PEs like the ILLIAC IV! cool! https://www.cs.cmu.edu/afs/cs/academic/class/15740-f03/public/doc/discussions/uniprocessors/vector/vector-past-present-future-supercomputing98.pdf
really good benchmarking library that highlights limitations of SIMD Processing really effectively https://arxiv.org/abs/2309.02680

Miscellaneous

Role_of_microglia_in_disease - as I have Asperger's and have been abused, tortured, violently assaulted and had brain damage at birth requiring life-support at 2-6 weeks, this is a relevant topic for me. see Periventricular_leukomalacia#Association_with_autism
Open-source_car
Random-access_memory#Memory_wall - keep an eye on this as I need to put it into the SIMT and GPU pages somewhere.
Interleaved_memory - was looking for this
GPGPU has to be a redirect due to it being a term also used for hardware - https://en.wikipedia.org/w/index.php?title=General-purpose_computing_on_graphics_processing_units&redirect=no
- GPGPU hardware
- keep an eye out https://en.m.wikipedia.org/w/index.php?search=insource%3A%2F%5C%5B%5C%5BGeneral-purpose+computing+on+graphics+processing+units%5C%7C%2F&title=Special%3ASearch&profile=advanced&fulltext=1&ns0=1
- related - GPU_cluster - and category Category:GPGPU

Consciousness

I have recently written articles/papers about this to keep an eye on the topic, which as a general rule is one hell of a mess.

disclosure of articles: these are non-peer-reviewed academic-style papers

http://lkcl.net/reports/consciousness_definition/ - yes really, it is 2025 and nobody can nail consciousness's "foot to the floor"
http://lkcl.net/reports/consciousness_turing/ - consequently the Turing test is basically useless (how can you test for something that has no definition???) - but there is light, because MacKenzie Cleeremans Jiménez and myself actually have Definitions of C. yippee.
http://lkcl.net/reports/consciousness_epigenetics/ - accidentally demonstrate that Epigenetics displays the properties of Consciousness. wasn't the original goal of the paper!

pages being kept an eye on / edited:

Consciousness - pretty good, accurately reflects how much of a mess human understanding of the concept is
Turing test - whoa this is a big page, and reflects accurately again how much misunderstanding there is of Turing's work.
Artificial consciousness - really good informative page
Global_workspace_theory - Bernie Barrs is cool.
Sparse distributed memory
- https://github.com/trendmicro/tlsh
Particle_swarm_optimization - surprisingly relevant topic esp. to Epigenetics where mathematics of Convolution: 1/x turns into 1/x^2.

Computing

broken down into subsections on theory and practice and software

Computing Theory

Amdahl's law - i prefer Koomey's law :)
Central_processing_unit - TODO, add SIMD harmful citation in data parallelism
General-purpose_computing_on_graphics_processing_units (software)
General-purpose_computing_on_graphics_processing_units (hardware)
History_of_supercomputing and naturally History_of_computing_hardware
Out-of-order_execution - not in a good shape, quite unclear
Scoreboarding - factually incorrect, needs a lot of editing. done most of it.
Reservation station
C-slowing and Pipeline_(computing) interrelationship really not made clear. Instruction_pipelining
Lookup_table - sort-of hardware and sort-of software, LUTs are used in FPGAS and also in instruction sets: you put the Hasse diagram into the instruction opcode: vpermd in Power ISA does ternary (LUT3) by using 8-bit of the opcode.
Stream_processing - absolutely great article, except it is highly SIMD-biased. loved the SQL JOIN example.
Register_allocation
Hardware-based_encryption

Architecture types and categories

Simultaneous_multithreading - need to keep an eye on this one as it fits, i think, into MISD Flynn's Taxonomy.
Transport_triggered_architecture - cool!
Asynchronous_array_of_simple_processors - fascinating
Vector processing - pretty good, but contains novel "tutorials" - very useful ones. there are other mathematical articles around with similar tutorials so is ok
Single_instruction,_multiple_threads - a bit of a mess, not intentionally, there is just not any public information. ILLIAC IV was the first key SIMT machine
Single_instruction,_multiple_data - big mess, clarifying.
Flynn's_taxonomy - found 1972 paper with additional categories not well-known, just realized I messed up, as re-reading the ILLIAC-IV documents it has Predicated SIMD PEs! only 2 bits (2x32 FP). also puzzlingly nobody mentions SWAR. 1966 Paper also mentions SOLOMON and ILLIAC IV
Duncan's_taxonomy
Content-addressable_parallel_processor - stub-ish. needs work
Massively_parallel_processor_array
Barrel_processor

ISAs

Instruction_set_architecture
Comparison_of_instruction_set_architectures
Bit_manipulation_instruction_set - got this page taken out of draft, quite a bit of work to do (mostly done)
Bit_manipulation
Bitwise ternary logic instruction - new page, amazing how LUT3 keeps cropping up
Power_ISA
AVX512 and avx512 has ternlog
SuperH
Hamming_weight - aka "Popcount"
Find first set - CLZ etc etc
Multiply–accumulate_operation
GFNI instruction set - Galois Field instruction set from Intel but has really useful stuff

ISA theory

Permute_instruction - naming gets horribly confused. not a mathematical permutation at all due to duplicates being permitted!
Gather/scatter_(vector_addressing)
Swizzling_(computer_graphics) - a miniature permutation actually not really, it's a combination. I think. all very confusing
Predication_(computer_architecture) - masking is frequently confusing because in array Processors that have SWAR ALUs mask bits can apply to individual PEs, and then totally separately: also on sub-words of the PE's SWAR. moo.
Mask_(computing) - helps explain predication?
Control_flow - alternative to predication
Addressing_mode#Conditional_execution - when predication is applied to load/store
Processor_supplementary_capability

Practice

Transputer - really should have the ALICE network and the PARAM 8000 in the page
TX-2 - world's first SIMD! 1958! even predates the term SIMD. cool! also had 32 modes of predicate masking and an innovative bitmanip instruction set. wow, wow
TI_Advanced_Scientific_Computer
ETA10
IBM_3090
FR-V_(microprocessor) vs Fujitsu_VP2000 and Fujitsu_VP - VLIW DSP inspired by Supercomputer
OpenPOWER_Microwatt
Amiga_Original_Chip_Set#Blitter including in the Amiga blitter! was mem-2-mem
Larrabee and this! aieyaaa rescued Tom forsyth's talk as handmade-seattle is offline! https://vimeo.com/handmadeseattle/download/450406346/4f684510e7 https://vimeo.com/450406346
Libre-SOC - being developed (done, thanks User:Henriok) - watch links Special:WhatLinksHere/Libre-SOC