Welcome to foldingforum.org, Yeren.
Yeren wrote:My concern is it has been stuck at 99% now for the last 3 hours, saying 56 seconds to completion
The time-per-frame is a direct function of the speed of your computer and the number of free cores it can use. SMP slows down nonlinearly, If V7 finds and allocates 4 CPUs to the job and then finds you're using an appreciable percentage of one or more of them, FAH yields resources to your foreground processing and in the process, slows down quite radically. The software can only project when it will reach 100% accurately if it can assume conditions have not changed -- and they are always changing when you're using the computer for something else. THe fact that you run 14 hours per day also tends to distort the estimates.
That doesn't make what you're doing wrong, the software is just limited in that way.
The server indicates that you have two computers plus one GPU folding. One SMP client is a lot more efficient than the other one. Here are the last WUs returned, one from each machine:
Hi Yeren (team 98860),
Your WU (P8013 R76 C33 G111) was added to the stats database on 2012-07-25 02:07:12 for 378.25 points of credit.
Your WU (P8013 R160 C18 G84) was added to the stats database on 2012-07-25 04:07:36 for 1259.7 points of credit.
Both happen to be from the same project but the bonus is radically different because you finished the second one about 10x as quickly as the first one. (Without a bonus, you'd earn the baseline points which are 180.60 for project 8013.)
On the other hand, Project 7646 took you 15.15 Days to complete the WU which exceeded the timeout (preferred deadline) so you didn't get a bonus.
Hi Yeren (team 98860),
Your WU (P7647 R95 C0 G11) was added to the stats database on 2012-07-23 15:08:31 for 2475 points of credit.
If productivity matters to you, I'd certainly plan on allowing that machine to fold more hours per day or see about balancing the number of free cores with the number requested for SMP. Under the actual pattern of use on that computer, it's possible that SMP:3 will fold faster than SMP:4.