Advanced Micro Devices Corporation

Sun Microsystems, Inc.

Digital Equipment Corporation

Rajesh Khanna

Hamid Partovi

William Ralph Wheeler

Ajay Bhatia

Steven W Butler

Michael Leary

Michael A Case

Jacob Thomas

Dhiraj Mallick

Anil L Pandya

Satish Kumar Raj

A timing-constrained circuit (e.g., a self-timed circuit) of optimal performance is achieved by allowing the delay of the circuit to be changed dynamically as a function of operating conditions (e.g., operating voltages or temperatures). The delay of timing signals in the timing-constrained circuit for a given operating condition may be selected to have the minimum margin for that operating condition among the available delays to maximize performance over the entire dynamic range of operating conditions.

Dynamic voltage scaling for self-timed or racing paths

Tools and techniques used in conjunction with integrated circuit path timing information can selectively reduce the channel length of transistors in cells associated with the most critical paths in an integrated circuit, while keeping the overall integrated circuit design within a specified power budget. Moreover, by targeting pins of cells (and thus their associated transistors) that are used by multiple paths, and/or that offer the greatest potential speed improvement, timing violations along critical paths can be reduced or eliminated with a relatively few number of replacements. Paths within a certain timing violation range are selected for analysis. The pins within those paths are ranked by pin criticality, which can depend on, for example, the number of times a particular pin occurs in any path, the timing enhancement associated with replacing a cell having that pin, and the impact of replacing a cell having that pin would have on the power budget. Transistors within cells (or entire cells) associated with pins are replaced based on the pin criticality until timing improvements are sufficient to remove a path from the range of paths being examined. Successive paths, and ranges of paths are analyzed until the power budget is exceeded, or no more improvements can be made.

Selectively reducing transistor channel length in a semiconductor device

A NOR gate pair includes a first and second NOR gate, each with a plurality of inputs and an output. A first NAND gate has a first input coupled to the output of the first NOR gate, a second input coupled to the output of the second NOR gate through a first input inverter, and an output. A second NAND gate has a first input coupled to the output of the second NOR gate, a second input coupled to the output of the first NOR gate through a second input inverter, and an output. A first output inverter is coupled to the output of the first NAND gate and a second output inverter is coupled to the output of the second NAND gate. This configuration assures that NOR gates used in a one-hot-decode decoder will all have logic-low outputs during a precharge phase.

Dynamic NOR gates for NAND decode

A tag comparator and bank selector for a set-associative cache in a computer system operates in a minimum time so that a cache hit or miss signal is generated early in a memory cycle. The data memory of the cache has two (or more) banks, with a tag store for each bank, and the two banks are accessed separately and in parallel using the index (low order address bits) while the tag translation is in progress. Two bit-by-bit tag compares are performed, one for each tag store, producing two multibit match indications, one bit for each tag bit in each tag store. These two match indications are applied to two separate dynamic NOR gates, and the two outputs applied to a logic circuit to detect a hit and generate a bank-select output. There are four possible outcomes from the compare operation: both banks miss, left bank hits, right bank hits, and both banks hit. The later condition indicates a possible ambiguity, and neither data item should be used, so a miss is signalled. The comparator is in large part self-timed using a flow-through design, as distinguished from being enabled on clock edges. Delay elements in the bank select logic allow the banks to be timed against each other, and current limiters are employed to equalize the timing of miss signals, regardless of the number of match lines switching high (which is data dependent). An address producing 19-of-20 match bits will result in a NOR gate output of about the same timing as an address producing no match bits, even though the former will turn on only one transistor to discharge the precharged output node of the NOR gate, whereas the later will turn on twenty paths for discharge. Although a two-way set associative cache is shown herein as an example embodiment, one of the features of the invention is that higher levels of associativity, e.g., four-way and eight-way, are equally well accommodated.

Fast tag compare and bank select in set associative cache

Growing community of inventors

Fremont, CA, United States of America

Rajesh Khanna