3 ms·
Seems like an obvious question but is there conventional wisdom on whether compilation/transpilation heavy workloads are more suited to cache size or to higher
by cronin101 3y ago
Seems like an obvious question but is there conventional wisdom on whether compilation/transpilation heavy workloads are more suited to cache size or to higher clocks? Is this a “it depends” situation? Wondering what to pick for my next workhorse.
- mastax 3y agoI thought I remembered benchmarks from when the 5800X3D came out showing it was good at code compilation but that is, at the very least, not always true. https://www.phoronix.com/review/amd-ryzen9-7950x3d-linux/13 https://www.phoronix.com/review/amd-ryzen9-7950x3d-linux/13
- the8472 3y agoAssuming the power is reported correctly it looks like an efficiency win: > While the build times were similar between the 7950X and 7950X3D, with the latter the CPU power consumption was significantly lower.
- mastax 3y agoI'm assuming the vast majority of that benefit is just from the X3D chips having a lower stock TDP. You could achieve the same efficiency by just reducing the 7950X TDP to 120W in BIOS or Ryzen Master.
- re-thc 3y agoYes, but this would reduce performance. The cache makes up for it whilst being able to run at a lower TDP.
- ilyt 3y agoMaybe but at that point you paid more for less
- boyter 3y agoNot really. Chips are mostly factory overclocked these days. They dump in all that power to get the coveted #1 spot in the review charts. Turn on eco mode and you get 99% of the performance at a much nicer power limit. Just pretend it’s 2013 and you got a chip with some nice overclocking potential you don’t really want to use because you don’t like heat or the noise it brings.
- sosodev 3y agoIt’s kinda random. Every workload is going to have different requirements depending on how they’re implemented and the context of what the program is actually doing.