← Back to Patent Explorer
US5701507A Expired / Lapsed US ⏱️ 0.0 years remaining

Architecture of a chip having multiple processors and multiple memories

Assignee: Texas Instruments Inc
•
Filing: 1995-06-07
•
Granted: 1997-12-23
•
Expires: 2015-06-07
📄 Original PDF 🌐 Google Patents 🔄 Sync Now

Abstract

Abstract A method of manufacturing integrated circuits uses an architecture having multiple processors and multiple memories, such that there is at least first and second groups of processors and memories. The first group has at least a first processor and at least a first memory. The second group has at least a second processor and at least a second memory. Regardless of where the architecture is sliced, the integrated circuits have a majority of the same address and data pin-outs.

Inventors

  • 👤 Walt C. Bonneau, Jr.
  • 👤 Karl Guttag
  • 👤 Robert Gove

Bibliographic Details

Application Number —
Jurisdiction US
Data Source google_patents
Last Synchronized 2026-11-09 02:27
CPC Classifications
G; G06; G06F; G06F13/00; G06F13/14; G06F13/16; G06F13/1605; G06F13/1652; G06F13/1663; H; H10; H10W; H10W90/00; H10D; H10D89/00; Y; Y10; Y10T; Y10T29/00; Y10T29/49; Y10T29/49002; Y10T29/49117; Y10T29/49121

Claims Analysis

Total Claims: 10
Claim 1
Independent Claim 297 words
A method of manufacturing integrated circuits using semiconductor chips, comprising the steps of: a. making an architecture having multiple instances of a modular unit including a processor, a memory and a crossbar link disposed therebetween, said crossbar links of said modular units connected together providing direct communication between any processor and any memory of a predetermined maximum number of said multiple modular units, and having input/output pads for connecting said architecture to external circuits; b. grouping said modular units into at least first and second groups, including in each group at least one of said modular units; c. selecting a first desired number of modular units for a first integrated circuit; d. slicing said architecture between any two groups, to give said selected number of modular units; e. repositioning said input/output pads; f. terminating said connection between crossbar links at said slicing between said two groups; g. constructing an integrated circuit having said selected first desired number of modular units, said constructing step including 1. disposing a first subset of said selected first desired number of modular units on a first semiconductor chip and a second subset of said selected first desired number of modular units on a second semiconductor chip, and 2. bonding said first and second semiconductor chips together by connection of said crossbar links of a predetermined modular unit disposed on said first semiconductor chip and a predetermined modular unit disposed on said semiconductor second chip; and h. repeating steps d, e, f and g for a second desired number of modular units; wherein said first desired number of modular units is different from said second desired number of modular units, and wherein said integrated circuits have a majority of the same address and data pin-outs, regardless of said number of modular units chosen.
Claim 2
Independent Claim 297 words
A method of manufacturing integrated circuits using semiconductor chips, comprising the steps of: a. making an architecture having multiple instances of a modular unit including a processor, a memory and a crossbar link disposed therebetween, said crossbar links of said modular units connected together providing direct communication between any processor and any memory of a predetermined maximum number of said multiple modular units, and having input/output pads for connecting said architecture to external circuits; b. grouping said modular units into at least first and second groups, including in each group at least one of said modular units; c. selecting a first desired number of modular units for a first integrated circuit; d. slicing said architecture between any two groups, to give said selected number of modular units; e. repositioning said input/output pads; f. terminating said connection between crossbar links at said slicing between said two groups; g. constructing an integrated circuit having said selected first desired number of modular units, said constructing step including 1. disposing a first subset of said selected first desired number of modular units on a first semiconductor chip and a second subset of said selected first desired number of modular units on a second semiconductor chip, and 2. bonding said first and second semiconductor chips together by connection of said crossbar links of a predetermined modular unit disposed on said first semiconductor chip and a predetermined modular unit disposed on said semiconductor second chip; and h. repeating steps d, e, f and g for a second desired number of modular units; wherein said first desired number of modular units is different from said second desired number of modular units, and wherein said integrated circuits have a majority of the same address and data pin-outs, regardless of said number of modular units chosen.
Claim 3
Dependent Claim 21 words
The method of claim 1 including the further step of arranging placement of said groups so that said groups are symmetrical.
Claim 4
Dependent Claim 42 words
The method of claim 2 wherein said symmetry is arranged so that if a first part of said architecture is flipped over onto a second part of said architecture, said processors are facing each other, and said memories are facing each other.
Claim 5
Dependent Claim 37 words
The method of claim 2 wherein said symmetry is arranged so that if a first part of said architecture is flipped over onto a second part of said architecture, said groups are mirror images of each other.
Claim 6
Dependent Claim 26 words
The method of claim 1 wherein said step of making said architecture further includes making a nonmodular unit to be included in every integrated circuit manufactured.
Claim 7
Dependent Claim 86 words
The method of claim 5 wherein said nonmodular unit includes a master processor, a master memory and a master crossbar link disposed therebetween and connected to said crossbar link of an adjacent modular unit, said master crossbar link and said crossbar links of said modular units together providing direct communication between said master processor and said master memory and any memory of a predetermined number of said multiple modular units, and further providing direct communication between said processor of any modular unit and said master memory.
Claim 8
Dependent Claim 48 words
The method of claim 6 wherein said nonmodular unit includes a transfer/memory controller connected to said master crossbar link and to at least one external device, said transfer processor providing communication between said master processor, said processors of each modular unit and said at least one external device.
Claim 9
Dependent Claim 13 words
The method of claim 5 wherein said nonmodular unit includes a frame controller.
Claim 10
Dependent Claim 21 words
The method of claim 5 wherein said nonmodular unit includes a master clock distributed to all portions of said integrated circuit.

References Cited (Backward Citations)

US4447881A Cited
US4398248A Cited
US4968977A Cited
US4978633A Cited
US5226125A Cited
US5369596A Cited

USPTO Annuity & Maintenance Windows

Annuity Stage Window Opens Payment Deadline Surcharge Period Until Estimated Fee Status
3.5 Years (Year 4) 2000-12-22 2001-06-23 2001-12-23 $2000.00 Window Closed / Lapsed
7.5 Years (Year 8) 2004-12-22 2005-06-23 2005-12-23 $3760.00 Window Closed / Lapsed
11.5 Years (Year 12) 2008-12-22 2009-06-23 2009-12-23 $7700.00 Window Closed / Lapsed

Assigned Categories & Tags

No categories assigned to this patent yet.

Assign Category Tag