Patent US4870608A successfully refreshed from Google Patents.
US4870608A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Method and apparatus for floating point operation
Assignee: TRW Fahrwerksysteme GmbH and Co KG
•
Filing: 1987-11-20
•
Granted: 1989-09-26
•
Expires: 2007-11-20
Abstract
Abstract Method and apparatus for floating point operation for calculating an approximate solution in a given argument of a function. An operation unit for carrying out floating point logical operation and floating point multiplication, a first memory for storing data necessary for operation and data produced in a course of operation, a second memory for storing a microprogram for controlling a process of operation of the operation unit, a micro-sequencer for issuing a control command necessary for the operation unit to carry out the operation, in accordance with the microprogram, a third memory for storing a table of solutions of coefficient functions in the series polynomial approximate equation of the function including a coefficient function consisting of numeric logic operation or multiplication operation, and an address latch for designating an address of the solution of the series expansion corresponding to the given argument of the function in the third memory are provided. When the function is instructed and the corresponding argument are given, the micro-sequencer calculates the address in the table of solution of the coefficient function corresponding to the argument, loads the calculated address to the address latch and reads the solution of the coefficient function from the table and supplies it to the operation unit. The operation unit calculates the solution of the given argument of the function based on the solution of the coefficient function read from the table, in accordance with the control command from the micro-sequencer.
Inventors
- 👤 Masatsugu Kametani
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-30 18:03
CPC Classifications
G; G06; G06F; G06F9/00; G06F9/06; G06F9/22; G06F9/223; G06F1/00; G06F1/02; G06F1/03; G06F1/035; G06F1/0356; G06F7/00; G06F7/38; G06F7/48; G06F7/483; G06F2101/00; G06F2101/04
Claims Analysis
Total Claims: 11
Claim 1
Independent Claim
257 words
A floating point operation apparatus for calculating an approximate solution for an externally given argument of a function, said apparatus comprising: means for performing arithmetic operation, numeric conversion and data transfer; first memory means for storing data necessary for the arithmetic operation and data produced in a course of the arithmetic operation; second memory means for storing a microprogram for controlling a process of the arithmetic operation of said operation means; a micro-sequencer for issuing a control command necessary for said operation means to perform the arithmetic operation in accordance with the microprogram; third memory means for storing a table of solutions of coefficient functions in a series polynomial approximate equation of said function; an address latch for latching address information designating an address of a location storing the solution of the coefficient function corresponding to the given argument of said function in said table of said third memory means; said micro-sequencer including means, responsive to the operation for said function being instructed and the corresponding argument being given, for calculating said address information of the location in said table storing the solution of the coefficient function corresponding to the argument, loading the calculated address information to said address latch, reading out the solution of the coefficient function from said table supplying the solution to said operation means; and said operation means including means for calculating said approximate solution for the given argument of said function based on the solution of the coefficient function read from said table in accordance with the control command of said microsequencer.
Claim 2
Dependent Claim
257 words
A floating point operation apparatus for calculating an approximate solution for an externally given argument of a function, said apparatus comprising: means for performing arithmetic operation, numeric conversion and data transfer; first memory means for storing data necessary for the arithmetic operation and data produced in a course of the arithmetic operation; second memory means for storing a microprogram for controlling a process of the arithmetic operation of said operation means; a micro-sequencer for issuing a control command necessary for said operation means to perform the arithmetic operation in accordance with the microprogram; third memory means for storing a table of solutions of coefficient functions in a series polynomial approximate equation of said function; an address latch for latching address information designating an address of a location storing the solution of the coefficient function corresponding to the given argument of said function in said table of said third memory means; said micro-sequencer including means, responsive to the operation for said function being instructed and the corresponding argument being given, for calculating said address information of the location in said table storing the solution of the coefficient function corresponding to the argument, loading the calculated address information to said address latch, reading out the solution of the coefficient function from said table supplying the solution to said operation means; and said operation means including means for calculating said approximate solution for the given argument of said function based on the solution of the coefficient function read from said table in accordance with the control command of said microsequencer.
Claim 3
Dependent Claim
21 words
A floating point operation apparatus according to claim 1 wherein said third memory means stores said table in floating point format.
Claim 4
Dependent Claim
68 words
A floating point operation apparatus according to claim 2 wherein said microsequencer includes means for accessing said third memory means independently from and parallelly to the operation of said operation means and the access to said first memory means, whereby addition, subtraction, multiplication and division and data movement are carried out in parallel with the readout of the solution of the coefficient function from said third memory means.
Claim 5
Dependent Claim
66 words
A floating point operation apparatus according to claim 1 wherein the series polynomial approximate equation of said function comprises a plurality of different coefficient functions, said third memory means includes a plurality of independent memories and at least some of said different coefficient functions are stored in respective independent memories so that said microsequencer can parallelly and independently access said different coefficient functions in said memories.
Claim 6
Dependent Claim
123 words
A floating point operation apparatus according to claim 1 wherein the function is given by f(X), the series polynomial approximate equation is represented by an n-order approximate equation ##EQU6## (m and n are integers) having a coefficient function a n (X') represented as a function of only a neighborhood value X' of the argument X, said coefficient function a n (X') being a coefficient of each term of said n-order approximate equation, said third memory means including means for storing a n (X') as the coefficient corresponding to the value X', and said operation means includes means for calculating the n-order approximate value by using the value a n (X') read from said table to calculate the approximate solution of the function f(X).
Claim 7
Dependent Claim
46 words
A floating point operation apparatus according to claim 5 wherein said micro-sequencer has independent control lines to said operation means and said third memory means so that said micro-sequencer can parallelly control the operation of said operation means and the access to said third memory means.
Claim 8
Dependent Claim
57 words
A floating point operation apparatus according to claim 5 wherein the plurality of different coefficient functions a n (X') which can be operated in parallel are stored in separate and independent memories included in said third memory means, and said micro-sequencer has independent control lines to said separate memories to individually control the access to said memories.
Claim 9
Independent Claim
115 words
A method for floating point operation for calculating an approximate solution for a given argument of a function in accordance with a microprogram, said method comprising the steps of: storing in a memory solutions of coefficient functions of a series polynomial approximate equation, said coefficient functions including coefficient functions having proper solutions corresponding to respective arguments to be given, said memory storing solutions corresponding to all of the arguments to be given; reading from said memory a solution of the coefficient function corresponding to the argument of said function; and carrying out a floating point operation of the series polynomial approximate equation based on the solution read from said memory to calculate said approximate solution.
Claim 10
Independent Claim
115 words
A method for floating point operation for calculating an approximate solution for a given argument of a function in accordance with a microprogram, said method comprising the steps of: storing in a memory solutions of coefficient functions of a series polynomial approximate equation, said coefficient functions including coefficient functions having proper solutions corresponding to respective arguments to be given, said memory storing solutions corresponding to all of the arguments to be given; reading from said memory a solution of the coefficient function corresponding to the argument of said function; and carrying out a floating point operation of the series polynomial approximate equation based on the solution read from said memory to calculate said approximate solution.
Claim 11
Dependent Claim
123 words
A method for floating point operation according to claim 8 wherein the function is given by f(X), the series polynomial approximate equation is represented by an n-order approximate equation ##EQU7## (n and m are integers) having coefficient functions a n (X') each represented as a function of only a neighborhood value X' of the argument X, the coefficient function comprises a plurality of different coefficient functions, said step of storing in said memory including storing some of the different coefficient functions into independent memories, respectively, in a form of coefficient function values corresponding to the neighborhood value X' of the argument X, and said step of reading the coefficient function is carried out parallelly to the memories storing the different coefficient function values.
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) | 1992-09-25 | 1993-03-27 | 1993-09-26 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1996-09-25 | 1997-03-27 | 1997-09-26 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2000-09-25 | 2001-03-27 | 2001-09-26 | $7700.00 | Window Closed / Lapsed |