Do the credit bonuses by sub-project make sense? Most do, BUT some don't.

The following shows the average CPU time by task (as of 17 April 2022), the sub-project and the sub-project bonus:
15:07 --- GFN-18 --- No credit bonus
15:42 --- SR5 --- 10% long job credit bonus.
16:59 --- PPS-Sieve --- No credit bonus
25:58 --- AP27 --- No credit bonus
34:40 --- TRP --- 10% long job credit bonus and a 10% conjecture credit bonus.
51:22 --- GFN-19 --- No credit bonus
58:20 --- 321 --- 10% long job credit bonus.
104:58 --- WW --- No credit bonus
156:11 --- WOO --- 20% long job credit bonus.
158:40 --- CUL --- 20% long job credit bonus.
160:12 --- ESP --- 10% long job credit bonus and a 10% conjecture credit bonus.
172:18 --- GCW --- 10% long job credit bonus.
188:00 --- GFN-20 --- 10% long job credit bonus.
232:24 --- PSP --- 35% long job credit bonus and a 10% conjecture credit bonus.
386:21 --- SOB --- 50% long job credit bonus and a 10% conjecture credit bonus.
550:58 --- GFN-21 --- 20% long job credit bonus.

I think GCW's long job credit bonus should be increased from 10% to 20%, since WOO and CUL
both run with less CPU time, and both WOO and CUL have a 20% long job credit bonus.
156:11 --- WOO --- 20% long job credit bonus.
158:40 --- CUL --- 20% long job credit bonus.
172:18 --- GCW --- 10% long job credit bonus.

I believe the times listed on the preferences pages should not be taken as representing the work load of the sub-project task. Those times are a report of the recent active tasks. If the majority of tasks were to be run on slow computers then the times would be "long" implying lots amount of work done. If the majority of work is done on fast computers then the times would be "short" implying not much work has been done.

From my experience it would seem that shorter tasks are processed on the slower computers as those computers could not handle the larger tasks. And the longer tasks are correspondingly done on the faster computers. Thus there is the perception based on processing time that the small tasks require comparatively more work than longer tasks.

It should also be noted that running with hyper-threading can increase the processing times. HT is typically used by passive users who haven't tried to optimize their system for PG, whereas HT is often turned off on those systems trying to optimize for PG.
____________
Werinbert is not prime... or PRPnet keeps telling me so.
Badge score: 2x3 + 5x4 + 5x5 + 5x7 + 1x9 + 3x10 = 125