The spike downwards is due largely to the following:

1) Transition from short HFCC workunits to normal length
2) The workunits now flowing for FightAIDS@Home are going to average 9.5 hours vs 5.5 hours.

As a result the average work unit length is increasing fast which always causes a short term drop in returned and validated runtime (Sat avg length was 5.4 hrs/result, Sun avg length was 5.5 hrs/result and so far today it is 5.77 hrs/result).

Also, we have concluded our 'stress test' of the new environment. We pushed the new hosting site setup to just under 400k results/day and found a few issues that we were able to easily resolve. Once we dealt with those the system has run fine at that level. I feel very confident that we can reach 500k results/day with our environment configured as it is now - with one exception. More results/day translates into more storage required. We are pushing the limits of the active BOINC directory space at 400k results a day so when we need to push beyond that size we will need to increase that space. Fortunately that is SAN storage so we can increase it in about 1-2 weeks time (faster in an emergency situation). For now though we don't need that more space and so we are going to target workunits to average 7 hours per project. We have increased the deadline up to 10 days. This should be our stable configuration for now.