Armenius
Extremely [H]
- Joined
- Jan 28, 2014
- Messages
- 42,384
Yeah, I thought when we saw a 2.2 GHz Core 2 CPU was beating a 3.9 GHz Pentium 4 in single threaded performance 15 years ago that we would have learned clock speed doesn't matter as much.Some weird arguments going on here.
Clock speed doesn't matter, nor does the amount of cache. Bottom line, system throughput / speed is all that matters.. Sometimes the way to the best speed overall is to max the clocks. Sometimes it's to max the cache. It's going to depend on what is running, how memory intensive it is, and how cache-friendly the memory access patterns are.
Obviously one would like to have the best of both. The current state of the art doesn't permit both, at least not at an acceptable price point. I want to have a family car that goes 0-60 in 2.5 seconds and gets 50 MPG, too bad for me.
At least AMD gives us a choice based on use case needs. Always need max cores and clocks? 7950X. Mixed workload that sometimes needs clocks and sometimes cache? 7950X3D. Cache heavy workload? 7800X3D. Clock heavy workload that doesn't need all the cores? 7800X. The only one we're missing is the 7950X3D++ with extra cache on both CCD's, and I don't think that would be a common use case anyway.