Dark Silicon
Updated 2026-08-03
INTRODUCTION
English translation pending.
CORE DEFINITION
A constraint in semiconductor design observed as process scaling advanced: transistor density grew faster than the ability to remove heat, so a growing fraction of the chip must stay inactive, or dark, at any given time. The term was popularized in a paper by Esmaeilzadeh and colleagues. The consequence is that area no longer translates directly into usable compute. Designers respond with heterogeneous architectures, specialized accelerators, and dynamic voltage and frequency scaling, accepting that different regions light up at different times.
SCAFFOLDING EFFECT
Reduce cognitive load
- Budget realism: accept that only part of your total capacity is usable at once. - Scheduling shift: move the design goal from adding capability to lighting up the right region at the right time. - Architecture choice: prefer specialized units that run briefly over general units that run constantly.
Anchor fast decisions
Power consumed per unit area rises with clock frequency and voltage, while heat removal is limited by the package and cooling system. If every transistor switched at full speed, power density would exceed what the cooling can dissipate and the chip would fail. Keeping most of the area off keeps power density within budget, so throughput depends on scheduling which regions are active rather than on raw transistor count.
MINIMUM ACTION
In progress 0/1Practice this model in one real situation:
account_treeGenealogyexpand_more
menu_bookReferencesexpand_more
Source support: Explicit
- en.wikipedia.orghttps://en.wikipedia.org/wiki/Dark_siliconverified
PRIVATE NOTES · Only visible to you
SAVED Q&A
ENTRY Q&A · Private saving available
Ask with a clear boundary
thinkingmodels answers from published entry context only.
Your question is sent to thinkingmodels. The answer uses public entry context only.
RELATED MODELS