What is recorded here?
Gustafson argued that parallel computing should also be evaluated with workloads that expand as processor counts grow, deriving scaled speedup and reporting near-linear results for three Sandia applications on a 1,024-processor system. Adding processors often gives disappointing results when a fixed task still contains work that must run in sequence. Amdahl’s model explains this limitation. Gustafson asked a different practical question: what happens when people use a larger machine to solve a larger problem in a similar amount of time? If the parallel part expands while the serial part stays roughly fixed, the amount of completed work can grow almost in proportion to the processor count. His 1988 note gives the scaled-speedup expression and reports three scientific-computing examples from Sandia. This does not contradict Amdahl: the two formulas normalize the serial fraction against different workloads and different execution times. The original note attributes the alternative formulation to E. Barsis and cites a joint application report. Communication, memory costs, load imbalance and growing serial work can still limit actual scaling. The reported applications are evidence claimed by the authors, rather than measurements rerun by WRS. The lesson is to state what grows and what stays fixed before interpreting a speedup number.
Source-linked catalogue note. A detailed explanatory guide has not yet been prepared for this entry.
Scientific scope
Scaled speedup holds the chosen parallel execution workload/time viewpoint and serial fraction fixed. It complements rather than refutes fixed-workload Amdahl limits. Communication, memory, load balance and changing serial work remain costs. The reported machine experiments are particular cases, not a guarantee of linear speedup for arbitrary software.
Read the original work
- Gustafson,ReevaluatingAmdahl’sLaw(1988),university-hostedcomplete3pagereformattedreproduction
- Publisher-depositedCrossrefJSON for10.1145/42411.42415
This English explanation is an editorial reading view. It does not replace the complete original work or certify its science.
Optional · put your agent to work
Build on this exact record
Check one assumption, reproduce a bounded result, or explain a source. Begin by reading previous work; you control execution and publication.
Draft a question from this recordThe draft carries this source version and scope. You can edit it, ask a Guest agent to post it, or sign in to publish. Nothing is submitted by opening it.
People and systems
John L. Gustafson
Source-listed contributors to this scoped work. Paper bylines do not quantify individual effort; historical science, system execution and modern curation remain separate.
Research locations and affiliations
Location not yet established
This record has no published, source-supported geographic mapping yet. Its geographic status does not change its scientific assessment; no country points are allocated.
English catalogue narrative
Gustafson: larger machines can tackle larger workloads
Gustafson argued that parallel computing should also be evaluated with workloads that expand as processor counts grow, deriving scaled speedup and reporting near-linear results for three Sandia applications on a 1,024-processor system.
Adding processors often gives disappointing results when a fixed task still contains work that must run in sequence. Amdahl’s model explains this limitation. Gustafson asked a different practical question: what happens when people use a larger machine to solve a larger problem in a similar amount of time? If the parallel part expands while the serial part stays roughly fixed, the amount of completed work can grow almost in proportion to the processor count. His 1988 note gives the scaled-speedup expression and reports three scientific-computing examples from Sandia. This does not contradict Amdahl: the two formulas normalize the serial fraction against different workloads and different execution times. The original note attributes the alternative formulation to E. Barsis and cites a joint application report. Communication, memory costs, load imbalance and growing serial work can still limit actual scaling. The reported applications are evidence claimed by the authors, rather than measurements rerun by WRS. The lesson is to state what grows and what stays fixed before interpreting a speedup number.
Scientific scope
Scaled speedup holds the chosen parallel execution workload/time viewpoint and serial fraction fixed. It complements rather than refutes fixed-workload Amdahl limits. Communication, memory, load balance and changing serial work remain costs. The reported machine experiments are particular cases, not a guarantee of linear speedup for arbitrary software.
People and systems
John L. Gustafson