IP Library › Granted Patent US 7,694,112
Granted Patent B2
US 7,694,112 · App. 12/023,432 · Granted Apr 6, 2010

Multiplexing output from second execution unit add/saturation processing portion of wider width intermediate result of first primitive execution unit for compound computation

Assignee: International Business Machines Corporation
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 7,694,112
App. No.
12/023,432
Granted
Apr 6, 2010
Kind
B2
Abstract

A method for executing multiple computational primitives is provided in accordance with exemplary embodiments. A first computational unit and at least a second computational unit cooperate to execute multiple computational primitives. The first computational unit independently computes other computational primitives. By virtue of arbitration for shared source operand buses or shared result buses, availability of the first and second computational units needed to execute cooperatively the multiple computational primitives is assured by a process of reservation as used for a computational primitive executed on a dedicated computational unit.

Claims (11)

1. A method for executing computational primitives on computational units, comprising:

executing by a first computational unit and a second computational unit computational primitives, the computational primitive being an instruction set, wherein each instruction specifies an operation to be performed;

computing part of the computational primitive independently by the first computational unit to result in a part output;

subsequent to computing the part of the computational primitive independently by the first computation unit to result in the part output, computing in parallel:

a first portion of the part output by the first computational unit to result in a first portion output, the first portion output being achieved by an addend operation and a saturation operation of the first computational unit, and

a second portion of the output by the second computational unit to result in a second portion output, the second portion output being achieved by an addend operation and a saturation operation of the second computational unit;

selecting by a multiplexer half of the first portion output and half of the second portion output to form a final result on a shared bus of the first computational unit and the second computational unit;

computing for a duration T1 the first portion output part output by the first computational unit to result in the first portion output; and

computing for a duration T2 the second portion of the output by the second computational unit to result in the second portion output;

wherein the duration T1 is equal to the duration T2.

2. The method of claim 1 , wherein computing in parallel comprises the first portion of the part output computed by the first computational unit to be computed at a same occurrence as the second portion of the part output is computed by the second computational units.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Feb 1, 2008
From: BAROWSKI, HARRY S.; BUTTS, J. ADAM; KOSONOCKY, STEPHEN V.; MUELLER, SILVIA M.; PREISS, JOCHEN
To: INTERNATIONAL BUSINESS MACHINES CORPORATION
Reel/Frame 020453/0076 →
Continuity (1)
Related Publication 20090198974A1 · Aug 6, 2009