CPU utilization comparison: 32-bit vs. 64-bit

It's all a bit complicated, and I will probably get part of it wrong. The system taskmon CPU use is I assume accurately telling you what the real, total load on the system is. The 64-bit Reaper running bridged 32-bit plugins uses more total CPU than the 32-bit Reaper running native plugins. This is probably due to the additional overhead of the bridge host process. One factor in this overhead may be that anticipative FX processing (parallel lookahead processing) is able to balance loads across cores most efficiently with native FX, though that is really just a guess. The top line in the Reaper performance monitor shows you the total process-native CPU used. For 32-bit Reaper running native plugins, this number matches the total system CPU. For 64-bit Reaper running bridged plugins, this number is just the 64-bit processing, and does not include the bridged plugins. The FX CPU use number is not comparable across screenshots. For native FX, the number is an accurate measure of CPU utilization by FX. For bridged FX, it's not, instead it is measuring something like the utilized percentage of available elapsed time (rather than CPU processing time). That's why that number appears so high with bridged plugins.

OK. Thanks. That helps clear it up a bit. ;)

Does this mean that the CPU meter for bridged plugins may be overestimating the measurement? Or underestimating it? Also, is there room for improvement (performance wise/measurement wise)? Or is this an inherent limitation of the bridging procedure?