Branch Transition Rate (BTR)

Concept

Branch Transition Rate (BTR) is a control-flow characteristic metric used in the Genesys workload generation framework to quantify the probability that a branch instruction changes its direction between successive executions. BTR is one of the user-controllable knobs in Genesys and is highly correlated with branch misprediction rate, making it a key lever for shaping the control-flow behavior of synthetic benchmarks.

First seen 6/13/2026

Last seen 6/13/2026

Evidence 2 chunks

Wiki v1

WIKI

Overview

Branch Transition Rate (BTR) is metric 6 in the Genesys workload generation framework and belongs to the set of Control-flow Characteristics metrics. It is represented as a single user-controllable knob (1 value) and captures how frequently branch instructions switch their taken/not-taken outcome across executions ^[1].

Definition and Intuition

READ FULL ARTICLE →

NEIGHBORHOOD

No graph connections found for this entity yet. It may appear in future ingestion runs.

explore full graph →

RELATIONSHIPS

1 connections

Genesys ← uses 100% 2e

Genesys uses branch transition rate as a core metric for controlling control-flow behavior.

LINKED ENTITIES

2 links

Genesys USES Extracted graph relationship

Branch Misprediction Rate CORRELATED_WITH The evidence states explicitly that an application's branch misprediction rate is highly correlated with the transition characteristics (BTR) of its branch instructions, and Genesys can model individual branch transition probabilities to hit a target misprediction rate.

CITATIONS

4 sources

4 citations — click to collapse

[1] BTR is metric 6 in the Genesys framework and is one of the control-flow characteristics metrics represented as a single knob. Genesys: Automatically Generating Representative

[2] An application's branch misprediction rate is highly correlated with the transition characteristics of its branch instructions; higher transition probability makes a branch harder to predict. Genesys: Automatically Generating Representative

[3] Genesys models BTR by selecting a set of control instructions with high transition probabilities (frequent switching) and assigning the remaining branches very low transition probabilities, and can also tune individual branch transition probabilities to hit a target branch misprediction rate. Genesys: Automatically Generating Representative

[4] BTR is enforced in code generation by using an X86 test operation to set the condition codes that affect the outcome of the conditional branch at the end of each basic block, with the test operand controlled to achieve the target BTR metric. Genesys: Automatically Generating Representative