US4334284A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Multiplier decoding using parallel MQ register
Assignee: Unisys Corp
•
Filing: 1979-12-31
•
Granted: 1982-06-08
•
Expires: 1999-12-31
Abstract
Abstract There is described a floating point processor architecture which permits multiple bit shifting over strings of binary 1's and strings of binary O's in a single machine cycle. During a multiply operation, an MQ register (arranged in parallel) which stored the multiplier, shifts the multiplier out for decoding at a rate comparable to the rate at which the partial product is shifted. This is made possible by using a parallel MQ register so that two bits may be shifted per clock cycle. This architecture permits extremely fast multiplication by using a multiple bit shift architecture while minimizing hardware requirements.
Inventors
- 👤 Wilson T. C. Wong
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-25 10:55
CPC Classifications
G; G06; G06F; G06F7/00; G06F7/38; G06F7/48; G06F7/52; G06F7/523; G06F7/533; G06F7/5332
Claims Analysis
Total Claims: 18
Claim 1
Independent Claim
83 words
A floating point processor (FPP) architecture which permits multiple bit shifting, comprising, first register means for storing a multiplier signal, decoder means connected to receive signals from said first register means, arithmetic logic circuit means connected to receive signals from and supply signals to said decoder means, second and third register means connected to supply signals to and receive signals from said arithmetic logic circuit means, and source means for supplying signals to and receiving signals from said first and second register means.
Claim 2
Dependent Claim
83 words
A floating point processor (FPP) architecture which permits multiple bit shifting, comprising, first register means for storing a multiplier signal, decoder means connected to receive signals from said first register means, arithmetic logic circuit means connected to receive signals from and supply signals to said decoder means, second and third register means connected to supply signals to and receive signals from said arithmetic logic circuit means, and source means for supplying signals to and receiving signals from said first and second register means.
Claim 3
Dependent Claim
25 words
The FPP architecture recited in claim 1 wherein, said first register means is arranged in parallel thereby to produce at least two output signals concurrently.
Claim 4
Dependent Claim
29 words
The FPP architecture recited in claim 2 including, latch means connected between said first register means and said decoder means to permit serial output by said first register means.
Claim 5
Dependent Claim
61 words
The FPP architecture recited in claim 3 wherein, said decoder means receives at least two signal bits from said latch means and at least one signal bit from said first register means in order to decode a series condition represented by the signal bits presented to said decoder means and to control the arithmetic logic circuit means as a result thereof.
Claim 6
Dependent Claim
27 words
The FPP architecture recited in claim 1 wherein, said decoder means includes PROM means the contents of which determines the operation of said arithmetic logic circuit means.
Claim 7
Dependent Claim
30 words
The FPP architecture recited in claim 5 wherein, said decoder means includes prealign decoder means for aligning the floating point processor number as a function of the exponent portion thereof.
Claim 8
Dependent Claim
29 words
The FPP architecture recited in claim 5 wherein, the contents of said PROM means is selected by an address which includes the signals supplied by said first register means.
Claim 9
Dependent Claim
34 words
The FPP architecture recited in claim 5 wherein, the contents of said PROM means incorporates a unique code wherein certain code arrangements can be eliminated thereby to reduce the size of said PROM means.
Claim 10
Dependent Claim
21 words
The FPP architecture recited in claim 1 including, shift means connected to control the operation of said arithmetic logic circuit means.
Claim 11
Dependent Claim
23 words
The FPP architecture recited in claim 9 wherein, said shift means receives signals from said decoder means and from said third register means.
Claim 12
Dependent Claim
27 words
The FPP architecture recited in claim 9 wherein, said decoder means includes shift control means to control the left or right shift operation of the shift means.
Claim 13
Dependent Claim
27 words
The FPP architecture recited in claim 9 wherein, said shift means includes a combinational shifter connected from said third register means to said arithmetic logic circuit means.
Claim 14
Dependent Claim
22 words
The FPP architecture recited in claim 1 wherein, said second and third register means store the multiplicand and partial product signals, respectively.
Claim 15
Dependent Claim
27 words
The FPP architecture recited in claim 1 wherein, said source means supplies signals representative of a floating point number only when an FPP load operation is selected.
Claim 16
Dependent Claim
62 words
The FPP architecture recited in claim 1 wherein, said decoder means operates upon the signals received from said first register means to produce code signals which control the operation of said arithmetic logic circuit means in terms of signal shifts and additions or subtractions to be performed by said arithmetic logic circuit on the signals received from said second and third registers.
Claim 17
Dependent Claim
76 words
The FPP architecture recited in claim 1 wherein, said multiplier signal comprises a plurality of binary ones and zeroes, said decoder means examines said multiplier signal and produces output signals as a function of the pattern of said binary ones and zeroes in said multiplier signal, and said arithmetic logic circuit means selectively adds or subtracts the multiplicand signal stored in said third register means in response to the output signal produced by said decoder means.
Claim 18
Dependent Claim
40 words
The FPP architecture recited in claim 16 wherein, said decoder means detects a string of binary ones, or a string of binary zeroes, or an isolated binary zero, or an isolated binary one and produces an output signal representative thereof.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1985-06-07 | 1985-12-07 | 1986-06-08 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1989-06-07 | 1989-12-07 | 1990-06-08 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 1993-06-07 | 1993-12-07 | 1994-06-08 | $7700.00 | Window Closed / Lapsed |