Electrical computers: arithmetic processing and calculating – Electrical digital calculating computer – Particular function performed
Reexamination Certificate
1998-12-30
2001-06-05
Shin, Christopher B. (Department: 2182)
Electrical computers: arithmetic processing and calculating
Electrical digital calculating computer
Particular function performed
C708S525000
Reexamination Certificate
active
06243731
ABSTRACT:
TECHNICAL FIELD OF THE INVENTION
The present invention relates in general to processors and, more particularly, to an apparatus and method for extending register dynamic range on a processor.
BACKGROUND OF THE INVENTION
Data operations on processors may be performed with the use of registers, data buses and memory arrays. Most processors may have one or more registers called accumulators. An accumulator register may be where the arithmetic and logic operations are performed for the processor. Data may be taken from memory in the processor and stored into the accumulator register. Operations may be performed with the contents of the accumulator, such as ADD and SUBTRACT commands. After these operations, the results may be stored in the accumulator. The contents of the memory register remain unchanged.
In many controlled algorithms, such as state estimators and Kalman filters, accumulative results in the accumulator register may exceed the range of the register, resulting in register overflow. For example, a 32-bit accumulator register may have a result that exceeds the most positive or the most negative value of the 32-bit register. To accommodate these overflows, the register may be extended beyond its current bit size to extend its dynamic range. However, instructions and code written for the original bit size of the accumulator register would not be compatible with this new dynamic range. The same numerical operations are not maintained, and the effects on the carry and overflow flags for the register are different in the extended case. Floating point operations may be implemented instead of physically extending the range of the register, but these operations also require more instructions and software code. Further, the above solutions all require more silicon and higher costs to implement on a processor.
SUMMARY OF THE INVENTION
From the foregoing, it may be appreciated that a need has arisen for a method for extending register dynamic range on a processor without requiring more silicon or increasing costs. In accordance with the present invention, an apparatus and method for extending register dynamic range on a processor is provided that substantially eliminates or reduces the disadvantages and problems associated with conventional register operations.
An apparatus for extending register dynamic range on a processor is disclosed. The apparatus has a register for performing a set of processor operations and a counter with a predetermined value. During the set of operations, the counter is incremented when positive overflow occurs on the register and decremented when negative overflow occurs on the register. Upon completion of the set of operations, a logic unit saturates the register with a positive value when the value is greater than zero, and with a negative value when the value is less than zero. If the value is zero, the register is not saturated.
In another embodiment, a method for extending register dynamic range in accordance with the present invention comprises five steps. The first step comprises initializing a predetermined value in a counter on the processor. The second step comprises performing a set of processor operations on a register. The third step comprises incrementing the counter when positive overflow occurs on the register during the set of operations. The fourth step comprises decrementing the counter when negative overflow occurs on the register during the set of operations. The fifth step comprises, when the set of operations is complete, saturating the register with a positive value if the value is greater than zero, and saturating the register with a negative value when the value is less than zero, and setting the value to zero in the counter.
A technical advantage of the present invention is that a method for extending register dynamic range on a processor is provided. Another technical advantage is that the detectable range of a register is extended, while maintaining compatibility with existing devices. Another technical advantage is that the detectable range of a register is extended with minimal use of silicon and reduced software code.
REFERENCES:
patent: 4825400 (1989-04-01), Simoncic et al.
patent: 5408427 (1995-04-01), Einaj et al.
patent: 5450553 (1995-09-01), Kitagaki et al.
patent: 5550767 (1996-08-01), Taborn et al.
patent: 5553015 (1996-09-01), Elliott et al.
patent: 5675339 (1997-10-01), Andren et al.
patent: 5684728 (1997-11-01), Okayama et al.
patent: 5710549 (1998-01-01), Horst et al.
patent: 6037947 (2000-03-01), Nelson et al.
patent: 6148309 (2000-11-01), Azagury et al.
Brady III Wade James
Shin Christopher B.
Swayze, Jr. W. Daniel
Telecky , Jr. Frederick J.
Texas Instruments Incorporated
LandOfFree
Apparatus and method for extending register dynamic range does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for extending register dynamic range, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for extending register dynamic range will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2541637