UMA vs. NUMA

Equally mentioned on the last page, yous switch between retention access modes on Threadripper processors. 'Distributes' mode uses 'Uniform Retention Admission' is enabled by default and then this is what we've been testing with so far. However, you lot can switch to 'local' mode for 'Non-Uniform Retentiveness Admission' and this allows each of the Zeppelin dies to prioritize which cores admission certain parts of the arrangement memory. This basically prioritizes the nearest cores to improve overall latency for gaming applications that tend to place a premium on fast retentiveness admission. So let's see how this impacts gaming performance equally well as a few productivity workloads...

Switching to NUMA has improved the Battlefield 1 functioning and now Threadripper is delivering like frame rates to the Ryzen 7 1800X. In fact, the 1950X creeps ahead e'er so slightly and now is that much slower than the Core i7-7820X and Core i9-7900X.

F1 2022 also sees decent gains when switching to NUMA though this time isn't able to lucifer the Ryzen vii 1800X and this means it's as well well downwards on the Intel CPUs.

Performance in Civilization Six was already dandy and no extra functioning was had when switching to NUMA.

Looking at application performance, we encounter a slight performance decline with POVRay, but zero serious.

Blender also saw a slight reject in performance with the 1950X being about 4% slower using NUMA.

Last upwards we accept the Corona test, in which both NUMA and UMA offered the same operation. There's not much to meet hither, so let's motion on to some consumption figures.