Project Jupiter: The Ambitious Attempt to Evolve the PDP-10
In the competitive landscape of early high-performance computing, Digital Equipment Corporation (DEC) sought to push the boundaries of its hardware. One such effort was Project Jupiter, a mission to develop a next-generation PDP-10. This project ran concurrently with Project Venus, which aimed to create a high-performance VAX implementation that would eventually be released as the VAX 8600.
Both projects were housed side-by-side at DEC's Marlboro facility, sharing a common technological foundation. They both utilized Emitter-Coupled Logic (ECL) gate array technology—a high-speed logic family designed to minimize propagation delay and maximize processing speed.
[ไม่มีภาพประกอบ]The Technical Ambitions of Project Jupiter
Led by Bill McBride, Project Jupiter was envisioned as the savior of the DEC-10 product line. The primary goal was to achieve a massive leap in performance by targeting a 15-nanosecond processor cycle. However, this target proved to be beyond the state of the art for the era's engineering capabilities.
Architectural Hurdles
The transition to ECL technology was not without its challenges. Gordon Bell noted that the PDP-10's architecture was inherently difficult to redesign. Specifically, the system's heavily sequential byte-pointer and indirect-addressing behavior—methods the computer uses to locate data in memory—made it awkward to implement using the new design assumptions required for higher speeds.
Developmental Struggles and Failures
Project Jupiter was not alone in its difficulties; Project Venus also faced severe technical problems and delays. By the 1977–1979 timeframe, Project Venus was stuck in design review with deadlines slipping continuously. At that stage, nothing was operational, and gate-array chips had not yet been sent for fabrication. Gordon Bell later described the Venus/8600 effort as an "absolute mess" that had spiraled out of control.
Project Jupiter suffered from a critical loss of institutional knowledge. As Alan Kotok moved away from the PDP-10, DEC found itself in a position where few remaining staff members possessed the specific expertise required to build a PDP-10. Furthermore, the engineer leading the effort had a background in IBM System/360 systems, whose design techniques did not translate effectively to the unique architecture of the PDP-10.
[ไม่มีภาพประกอบ]Key Facts
- Project Jupiter was the effort to build a next-generation PDP-10.
- Project Venus was the concurrent effort to build the VAX 8600.
- Both projects utilized Emitter-Coupled Logic (ECL) gate array technology.
- The target processor cycle for Project Jupiter was 15 nanoseconds.
- Development took place at DEC's Marlboro facility.
- Project Jupiter was led by Bill McBride.
| Feature | Project Jupiter | Project Venus |
|---|---|---|
| Target System | Next-gen PDP-10 | High-performance VAX |
| Technology Used | ECL Gate Array | ECL Gate Array |
| Final Outcome | Technical failure/unrealized | Released as VAX 8600 |
| Primary Goal | 15ns processor cycle | High-performance implementation |
Frequently Asked Questions
What was the primary goal of Project Jupiter?
The primary goal was to develop a next-generation PDP-10 with a processor cycle of 15 nanoseconds to save the DEC-10 product line.
What technology did Project Jupiter and Project Venus share?
Both projects utilized Emitter-Coupled Logic (ECL) gate array technology to achieve high performance.
Why did Project Jupiter struggle with its design?
It struggled due to the PDP-10's sequential byte-pointer and indirect-addressing behavior, as well as a loss of internal expertise in PDP-10 architecture.
Who led the Project Jupiter effort?
The project was headed by Bill McBride.
How did the experience of the lead engineer affect Project Jupiter?
The lead engineer had previously worked on IBM System/360 systems, and those specific techniques did not transfer well to the PDP-10's distinct architecture.