/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

← → days · ↑ ↓ browse · Enter similar · o open

IBM researchers build supercomputer-on-a-chip

San Francisco - Supercomputers may soon be the same size as a laptop if IBM brings to market research detailed on Thursday, in which pulses of light replace electricity to make data transfer between processor cores on a chip up to one-hundred times faster.

InfoWorld

Context & Ripple Effects

The optical-interconnect research lands in the middle of a deliberate IBM push to make large-scale computing a corporate utility. In October 2007, IBM and Google announced a joint program to teach students massively parallel programming, and by mid-November IBM had launched Blue Cloud-style cloud offerings slated for Q1 2008, alongside 'green' data-center certificates aimed at measuring and monetizing energy efficiency. A chip that moves data between cores with light rather than electricity attacks both sides of that strategy at once: throughput and power.

Why it matters: if pulses of light really transfer data between processor cores up to one hundred times faster than electricity, the machine-room constraint stops being the wire and starts being the chip itself — which is what makes the reported path to a laptop-sized supercomputer plausible rather than rhetorical.

First-order effects

  • IBM gains a candidate differentiator for its server and cloud lines: an on-chip optical fabric that removes the electrical data-transfer bottleneck between cores, directly serving the massively parallel workloads its Q1 2008 cloud push targets.
  • High-performance-computing buyers get a credible near-term prospect of supercomputer-class throughput in a much smaller physical footprint, changing procurement assumptions about floor space and cooling.

Second-order effects

  • Rival chipmakers face pressure to answer with their own optical or otherwise faster interconnect technology, since core-to-core bandwidth becomes the visible performance frontier once IBM demonstrates it commercially.
  • Data-center operators — including the corporate customers of IBM's November 2007 energy-efficiency certificate initiative — gain a lever on power budgets, because replacing electrical signaling addresses the energy spent purely on moving data.

Third-order effects

  • If optical interconnects commercialize, the industry's performance narrative shifts further from raw clock speed toward data movement and parallelism — consistent with the direction the Google-IBM parallel-programming curriculum was already pointing in late 2007.
  • Supercomputing could restructure from room-scale installations assembled from many boxes toward dense, integrated systems, concentrating value in whoever controls the interconnect intellectual property.

The trend: Computing performance is migrating from making transistors switch faster to moving data between them — with light, on the chip, as the emerging medium for dense parallel machines.