| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Community
Forum: Hardware Chat Room
Thread: What subproject to use for CPU comparison? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 2
|
|
| Author |
|
|
mdxi
Advanced Cruncher Joined: Dec 6, 2017 Post Count: 109 Status: Offline Project Badges: 
|
I want to use a WCG subproject as a real-world comparison between various CPUs. My plan is to let the machines involved start from zero, fetch WUs from a given subproject, and run for 72 hours. At the end of that time, I want to do some basic analysis on the job_log files to get some ideas about relative performance.
----------------------------------------What subproject would be good for this? Are there any which are known to be computationally heavy, maybe with AVX usage or similar? I found a lot of posts talking about MIP being a computationally heavy project, but I also found a lot of posts talking about it having high L3 cache usage, leading to dramatic slowdowns when many MIP WUs are run concurrently. Was this ever confirmed, and if so, is it still true? If you don't think any WCG subprojects are a good answer to my question, what would you suggest using instead? Thanks!  |
||
|
|
QuantumEthos
Senior Cruncher Joined: Jul 2, 2011 Post Count: 336 Status: Offline Project Badges: 
|
Here is a set of photos to compare to AMD FX8320E
----------------------------------------The FightAIDS@Home - Phase 2 project seems to want to run a lot, however it is x86 The Microbiome Immunity Project or Mapping Cancer Markers may be a good one or Indeed as you say. Mapping cancer markers is a steady 5+ hour slog - definitely is about maths My impression is that most of these compare genetics though Integer compare sets, However when you compare drugs to RNA & cell compounds in viruses & germs .. that the most likely case where AVX may be useful is found, The most probable reason that fighting aids uses an x86 task is that it will be using AVX & Float mostly. Not sure about AVX and DNA as the Enhanced MMX feature sets of AVX are about Bit-swapping and other additive, subtraction, multiplication/Division of data sets in optimised Cache space. Zikka Is comparing drug compounds to materials found in the Zikka Virus.. Logically AVX is advantageous here also. RS [Edit 3 times, last edit by QuantumEthos at May 2, 2019 5:20:36 AM] |
|||
|
||||
|
|