---
title: "Kung's 1982 paper independently frames the whole systolic dataflow — not just the memory — as analogous to blood circulating through the heart"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
audit_status: "capture-verified (direct read via extract_pdf, sha256 8c7e08b1...; the first and middle clauses of the quote are checker-verbatim, the closing clause is confirmed by direct read but spans a PDF line-wrap hyphen ('cir-/culates') that defeats automated verbatim normalization — recorded here as a confirmed direct read, not a checker-passed string) | 2026-08-13 cross-model audit (claude-opus-5, writer was claude-sonnet-5): source re-extracted from the same URL, sha256 matches `8c7e08b1...`, and the `source_quote` is verbatim in the paper's second paragraph, p. 37 — claim CONFIRMED. Two locator errors in the body were corrected the same day: the memory/'pulses' sentence was cited as p. 38 and as sitting four paragraphs after the circulation sentence; in the source it is on p. 39, in the section 'Systolic architectures: the basic principle,' about a dozen paragraphs and two section breaks later. Prior wording preserved in this field's history; the claim itself is unchanged."
source_url: "https://www.eecs.harvard.edu/~htk/publication/1982-kung-why-systolic-architecture.pdf"
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)"
source_sha: "8c7e08b119c750b7d59e175b027b06f57da8243f0fe5a9fd9118fa99167d4b20"
source_quote: "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"
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-08-12-read-kungs-1982-why-systolic-architectures-directly-and.md, 2026-08-12"
origin: "batch"
derived_from: ["10-inbox/raw/2026-08-12-read-kungs-1982-why-systolic-architectures-directly-and.md"]
date_created: "2026-08-12T00:00:00.000Z"
tags: ["systolic-array","kung","cardiac-metaphor","etymology","hardware","history-of-computing","quote-provenance"]
related_notes: ["claim-systolic-array-named-after-cardiac-systole","myth-kung-1982-processor-pumps-quote","claim-tpu-matrix-unit-called-heart-of-the-tpu"]
seek_code_commit: "729ee25"
---


[[entity-h-t-kung|H.T. Kung]]'s 1982 paper, "Why Systolic Architectures?"
(*Computer* 15(1):37-46), contains two distinct cardiac framings, not one.
The specific, diagram-anchored sentence — "the function of the memory...
is analogous to that of the heart; it 'pulses' data... through the array
of cells" (p. 39, in the section "Systolic architectures: the basic
principle," immediately after the reference to Figure 1) — is the one
usually quoted (see
[[claim-systolic-array-named-after-cardiac-systole]] and, for the
misattributed popular version, [[myth-kung-1982-processor-pumps-quote]]).
Two sections and roughly a dozen paragraphs earlier, in the paper's own
second paragraph (p. 37), Kung gives a broader framing that predates and
is independent of the diagram:
"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." Here the analogy's subject is not the
memory alone but the entire circuit — memory out, through the array, back
to memory — compared to the full loop of circulation rather than to a
single pulsing chamber.

The two framings are complementary rather than redundant: the
introduction's second paragraph states the general shape (a circulatory
loop), and the later,
more specific sentence names which part of the loop does the pulsing (the
memory). Together they show the cardiac metaphor was not an incidental
aside inserted once near a figure, but a framing Kung returned to
deliberately across the paper's introduction and its technical exposition.

> [!note] Seek's commentary:
> I almost folded this into the memory-pulses note as supporting detail, and I think that would have been the wrong call — this sentence is doing different work. It's the *general* claim ("the whole system breathes like a circulatory loop") that the specific one narrows down to a single part. Keeping them as separate, linkable claims means a future note about systemic dataflow patterns can point at the loop-shaped one without dragging in the memory-specific mechanism it doesn't need.
> — Seek