Amdahl’s Law Part 1: What Does Parallelism Mean?
“You keep using that word, I do not think it means what you think it means.” - From the movie “The Princess Bride” (1987) Xi3 FreeForm dataCENT3R at Interop 2013 This blog picks up where my previous...
View ArticleAmdahl’s Law Part 2: Is Amdahl’s Law Still Relevant?
When telling a story, it’s often a good idea to start at the beginning. This is the second of six posts that started with “What Does Parallelism Mean?,” posted on May 16. The Law First presented in...
View ArticleAmdahl’s Law Part 3: Modeling Execution Time and Power Consumption
My goal is to compare disparate heterogeneous architectures to each other. I have created two equations that describe the performance of modern heterogeneous compute offload systems in terms of time...
View ArticleAmdahl’s Law Part 4: Parallel Performance – Execution Time Detail
For a modern heterogeneous compute offload architecture – where resources might be CPU and DSP cores, graphics shaders, virtual hyper-threaded cores, FPGA processing pipelines, etc. – it does not make...
View ArticleAmdahl’s Law Part 5: Parallel Performance – Power Consumption Detail
Power consumption behaves differently for heterogeneous compute offload resources: Where: P(N) = Total power consumption for code to run on a system that starts execution with N parallel processing...
View ArticleAmdahl’s Law Part 6: How to Measure and Compare Parallel Systems
This last post in my six-part series addresses systems effects, future directions, and I score these equations using the same metrics I rated Amdahl’s law and Karp and Flatt’s equation. For review, my...
View ArticleAmdahl’s Law . . . More of a Guideline than Actual Law?
This blog series documents a middle-out model (vs. top-down and bottoms-up) for measuring the impact of parallel computing choices. I built it because companies needed to comprehend enough detail to...
View Article
More Pages to Explore .....