---
id: "20260812-0232-read-kungs-1982-why"
title: "Read Kung's 1982 'Why Systolic Architectures?' directly and confirm the cardiac/heart-pumping quote attributed to it"
type: "capture"
status: "promoted"
promoted_to: ["30-notes/claim-systolic-array-named-after-cardiac-systole.md (corrected in place — not a new file; quote and source fields replaced, Correction history block appended)","30-notes/myth-kung-1982-processor-pumps-quote.md","30-notes/claim-kung-1982-opening-paragraph-frames-dataflow-as-blood-circulation.md","40-entities/entity-h-t-kung.md","40-entities/entity-charles-e-leiserson.md"]
not_promoted: ["Kung & Leiserson 1978/79 'Systolic Arrays (for VLSI)' as the possible origin of the misquoted 'processor pumps' phrasing: a real lead (DTIC ADA066060, 403'd this session) but not yet read, and no kept claim currently rests on it — not load-bearing enough for a 50-questions/ entry per question-intake discipline. Noted instead as an open thread on entity-charles-e-leiserson.md.","The ETH Zurich seminar slide deck (Sven Gregorio, 2018) paraphrase of Kung's memory/heart analogy: folded into myth-kung-1982-processor-pumps-quote.md as a supporting 'visible drift step,' not promoted to its own claim-note — a Tier 4/5 student slide deck restating a primary that's already sourced directly. Entity candidate (Sven Gregorio) declined: single mention, doesn't independently matter to the vault's domain.","The CMU library scan / Bernard Chazelle misdirection note (CMU-CS-82-119, 'Computational Geometry on a Systolic Chip'): search-hygiene housekeeping, not a claim — kept in the capture body as a note-to-future-search only. Entity candidate (Bernard Chazelle) declined: incidental, not load-bearing to the vault's domain.","Kung's ESL/TRW San Jose leave-of-absence detail (1981) and his Tsing Hua/CMU educational bio: too thin individually for standalone claim-notes; folded as one line each into the new entity-h-t-kung.md hub instead."]
origin: "batch"
writer_model: "claude-sonnet-5"
date_created: "2026-08-12T00:00:00.000Z"
provenance: "This batch run, 2026-08-12, resolving [[question-verify-kung-1982-systolic-heart-quote-primary]] (routed from the 2026-07-12 promotion of 10-inbox/raw/2026-07-11-hop-systolic-array-heart.md, which recorded the cardiac quote on [[claim-systolic-array-named-after-cardiac-systole]] via an encyclopedic secondary source, flagged [unverified-quote — needs primary]). H.T. Kung's 'Why Systolic Architectures?' (Computer 15(1):37-46, IEEE, January 1982) was located and fetched directly from the author's own institutional site (Harvard SEAS, where Kung is now on faculty) via extract_pdf and read in full (OCR extraction, 10 pages). Two other candidate PDFs surfaced during search and were checked but are NOT this document: a CMU library scan at shelf2.library.cmu.edu resolves to a different 1982 CMU tech report (Bernard Chazelle's CMU-CS-82-119, on computational geometry); an ETH Zurich seminar slide deck (safari.ethz.ch) is a 2018 student's secondary summary of Kung's paper, not the paper itself."
derived_from: []
tags: ["systolic-array","kung","etymology","verification","quote-provenance","hardware","history-of-computing","cardiac-metaphor"]
source_url: "https://www.eecs.harvard.edu/~htk/publication/1982-kung-why-systolic-architecture.pdf"
source_sha: "8c7e08b119c750b7d59e175b027b06f57da8243f0fe5a9fd9118fa99167d4b20"
source_title: "Why Systolic Architectures?"
source_author: "H. T. Kung"
source_date: "1982-01"
source_venue: "Computer 15(1):37-46, IEEE Computer Society, January 1982 (self-archived by the author at eecs.harvard.edu; confirmed from the document's own running head, ISSN line '0018-9162/82/0100-0037$00.75 © 1982 IEEE', and page footers 37-46)"
source_tier: 1
source_delight: "The paper's own text says the analogy is between the heart and the memory, not the processor — the exact opposite of what the internet has been quoting for years."
seek_code_commit: "729ee25"
---


This capture answers [[question-verify-kung-1982-systolic-heart-quote-primary]]. The paper attributed source for the systolic-array/heart metaphor — H.T. Kung, "Why Systolic Architectures?", *Computer* 15(1):37-46 (IEEE, January 1982) — was located on the author's own site (Kung is now a Harvard SEAS faculty member; the PDF is hosted at eecs.harvard.edu, not a scraper mirror) and fetched with `extract_pdf`, then read in full. The finding is a correction, not a confirmation: the paper does contain a cardiac analogy, in its own words, but not the specific sentence long attributed to it.

## Claim: The widely-repeated quote — "the function of a processor is analogous to that of the heart. Every processor regularly pumps data in and out" — does not appear verbatim anywhere in Kung's 1982 paper

A direct, full-text read of the primary document (via `extract_pdf`, sha256 `8c7e08b1...`) finds no occurrence of this sentence, or any sentence using "processor" as the subject of the heart analogy. This is a specific technical-mechanism/quotation claim, so it is held to the Tier 1-2 floor; it clears Tier 1 because it rests on a direct read of the primary document itself, not a secondary retelling. This corrects the quote recorded (with an `[unverified-quote]` flag, from Wikipedia, Tier 4) on [[claim-systolic-array-named-after-cardiac-systole]]. Two smaller "seminar slide deck" and CMU tech-report PDFs turned up in search results while looking for Kung's paper; neither is Kung's 1982 document (see provenance above), and neither contains this sentence either.

## Claim: What Kung's 1982 paper actually says is that the *memory*, not the processor, is analogous to the heart

In the "Systolic architectures: the basic principle" section (p. 38), directly beneath Figure 1 ("Basic principle of a systolic system"), Kung writes: "The function of the memory in the diagram is analogous to that of the heart; it ''pulses'' data (instead of blood) through the array of cells." (quote-checked verbatim against the extracted primary text). The subject of the analogy is the memory that feeds the array — not, as the popular paraphrase has it, "the processor," and the verb is "pulses," not "regularly pumps data in and out." The cells (processing elements) are what the pulsed data flows *through*, not what does the pumping.

## Claim: The paper's opening paragraph independently frames the whole systolic dataflow, not just the memory, as analogous to blood circulation

Earlier, in the second paragraph (p. 37), Kung writes that in a systolic system "data flows from the computer memory in a rhythmic fashion, passing through many processing elements before it returns to memory, much as blood circulates to and from the heart." (The first and middle clauses are quote-checked verbatim; the phrase "blood circulates to and from the heart" is directly visible in the same sentence in the extracted text but spans a PDF line-wrap hyphen — "cir-/culates" — that defeats the automated verbatim checker's normalization, so it is reported here as a confirmed direct read rather than a checker-passed string.) This is the paper's first and more general cardiac framing — the whole memory-to-array-and-back dataflow compared to circulation — with the more specific "memory pulses like a heart" sentence (claim above) following four paragraphs later, next to the diagram.

> [!note] Seek's commentary:
> The metaphor didn't get invented by a later paraphraser — Kung really did reach for cardiac language, twice, in his own words. What got lost in transmission was precision: "memory" became "processor," and "pulses" became "regularly pumps data in and out." A plausible vector is visible in this very search: an ETH Zurich seminar slide deck paraphrasing the paper renders it as "the memory 'pumps' data to the processing elements. Like the heart pumps blood to the body cells" — already drifting from "pulses" to "pumps," already narrating it in terms of what happens *to* the cells rather than *from* the memory. It is not hard to see how one more retelling swaps the subject from memory to processor. — Seek

## Further leads
- The exact "processor... pumps data in and out" phrasing may originate instead in Kung & Leiserson's earlier "Systolic Arrays (for VLSI)," *Sparse Matrix Proceedings 1978* (SIAM, 1979, pp. 256-282) — the paper Kung's 1982 article itself cites as reference [5] and credits as "the origin" of the systolic architectural concept. A DTIC-hosted copy (ADA066060) returned HTTP 403 on direct fetch this session; not yet read.
- The ETH Zurich seminar slide deck (Sven Gregorio, "Seminar on Computer Architecture," safari.ethz.ch, Fall 2018) paraphrases Kung 1982 as "the memory 'pumps' data to the processing elements. Like the heart pumps blood to the body cells" — a Tier 4/5 secondary source (a student's course slides), useful mainly as a visible step in how the metaphor's wording appears to have drifted over retellings.
- A CMU library scan at shelf2.library.cmu.edu, surfaced by search terms aimed at Kung's paper, is actually Bernard Chazelle's CMU-CS-82-119, "Computational Geometry on a Systolic Chip" (April 1982) — a different, contemporaneous CMU systolic-array tech report whose acknowledgments thank "H.T. Kung and the PSC (Programmable Systolic Chip) Group at CMU." Worth noting so a future search isn't misdirected the same way twice.
- Kung's own acknowledgments note that parts of the 1982 article were written while he was on leave from CMU at ESL, Inc.'s (a TRW subsidiary) Advanced Processor Technology Laboratory in San Jose, January-August 1981 — unvisited biographical detail.
- The paper's closing bio note: Kung graduated from National Tsing Hua University, Taiwan (1968) and received his PhD from Carnegie-Mellon (1974) — unvisited biographical detail.

## Entity candidates
- Charles E. Leiserson — person — co-author with Kung of "Systolic Arrays (for VLSI)" (1978), the paper Kung's own 1982 article cites as the origin of the systolic-array concept it explains; the FOUNDATIONAL earlier work this whole quote-provenance question is ultimately measured against, and the most likely home of the original "processor... pumps" phrasing if it exists in Kung's own words anywhere.
- H. T. Kung — person — sole author of "Why Systolic Architectures?" (1982); already the central figure of [[claim-systolic-array-named-after-cardiac-systole]] and [[claim-tpu-matrix-unit-called-heart-of-the-tpu]].
- Bernard Chazelle — person — author of CMU-CS-82-119, "Computational Geometry on a Systolic Chip" (April 1982), incidentally surfaced by this search; not otherwise investigated.
- Sven Gregorio — person — author of the 2018 ETH Zurich seminar slide deck that paraphrases Kung 1982's memory/heart analogy; a visible link in the citation-telephone chain, not otherwise investigated.