SU1297040A1 - Device for calculating function values - Google Patents

Abstract

The invention relates to computing and can be applied as a functional expander in a composition with a large computer or as an autonomous specialized high-speed calculator. The chain of the invention is to reduce hardware costs by placing reference values in the middle of the approximation interval. The value coming from the outputs of the high-order register of argument 1 is fed to the inputs of blocks 2 of the memory of reference values and numbers of intervals 3, Values from the outputs of memory 3 and the low-order register outputs of argument 4 determine the value of the correction function stored in memory block 5, which is subtracted or added to the reference value in summator-subtractor 6, the control input This element is NOT associated with the high-order output of the lower-order bit of argument 4. 1 Il. with (L

Description

eleven

The invention relates to computing and can be used in high-speed specialized computers.

The chain of the invention is to reduce hardware costs by placing the function reference values in the middle of the approximation interval and, consequently, reducing the size of the function increment code and the interval number code for calculating the function.

The drawing shows the block diagram of the device.

The device for calculating the function contains a register of 1 high-order bits of the argument, a block of 2 memory of reference values, a block of 3 memory of slot numbers, a register of 4 lower-order bits of the argument, a block of 5 memory of a correction function, a subtractor 6, an item of NOT 7.

The device operates as follows.

The function is calculated by the formula

) |

J

“F A (

.2 - 2 t rni

± cf (qi, x).

de F A (.

, t "-k

cp (q, o

m

) - the reference value of the function,

the code of the corrective function depends on the subband number q and the low-order register x,

q is the subband number.

 (k is the bit size) of the interval of the randomness of the function increment code and the interval number code, the interval number memory block, the correction function memory block, the totalizer subtractor and the NOT element are added to it, the output of the high order argument register is connected with the address inputs of the memory block of the interval numbers, the outputs of which are connected to the address inputs of the first

 range of argument change

split into 2

senior register

fishing The value of the elementary function in 5 groups of the memory block of the corrective

the middle of the interval is taken as the support function, the address inputs of the second group of which are connected to the outputs of the bits of the register of the lower bits of the argument, the output of the most significant bit of which is connected through the element NOT to the control input of the adder-read value

r «„ .хк2 +2 F А (J

The required value of the function is calculated by adding or correcting the correction code if (q, x) to the reference value, where the control points are divided into subbands taking into account the rate of change of the first derivative of the function.

l, the first and second informational inputs of which are connected to the outputs of the memory block of the reference 55 values and the memory block of the correction function, respectively, and the output of the totalizer-calculator is the output of the device.

At the output of blocks 2 it is processed

And 3 memories you 5

0

respectively supporting

D) 1 value of the function F А (, - - It code

subrange numbers. At the output of memory block 5, an additional correction function code is inserted. If, in the high order of the de register 4 lower bits, the unit (i.e., the code x:), the signal from the high order output of the register of the lower 4 bits of the argument converts the subtractor 6 into the subtraction mode, and from the function code at the reference point the code of the correction function, if in the high order of the register of the lower 4 bits zero, the code of the correction function read from the output of memory block 5 is added to the value of the function at the reference point.

Claims (1)

Claims An apparatus for computing a function comprising a high-order argument register, a lower-order argument register, and a reference value memory block, wherein the outputs of the high-order argument register are connected to the address inputs of the reference memory block. values, characterized in that, in order to reduce hardware costs by placing the function reference values in the middle of the approximation interval, reducing the bit of the function increment code and the interval number code, the interval number memory block, the correction function memory block are added to it, adder-subtractor and NOT element, with the output of the register of the higher bits of the argument connected to the address inputs of the memory block of the number of intervals, the outputs of which are connected to the address inputs of the first corrective memory block group l, the first and second informational inputs of which are connected to the outputs of the memory block of the reference 55 values and the memory block of the correction function, respectively, and the output of the totalizer-calculator is the output of the device.