I have been tracking the status S4 WU counters for the last ten days:
1532842, 1484374, 1436825, 1387363, 1336391, 1286103, 1237654, 1186583,
1132288, 1081776
linear projected end of new work 21.6258 days.
Using the figures you have recorded, which show on average a gradual increase in units crunched per day. If this increase continues then, using a simple compoud formula, on those figures, the projected end of S4 will be in 19.45 days from the datetime of your last figure.
Winterknight:
I noticed the increase but the sample is to small to justify anything more than simple linear projection.
You could be right, if this is continuing trend. Which it may well be if people are starting to horde WUs.
Looking at number of S4 units still to be sent and the number returned and validated last week. Will all the S4 units, except for re-sends, have been sent in approx. ten days?
Winterknight:
I noticed the increase but the sample is to small to justify anything more than simple linear projection.
You could be right, if this is continuing trend. Which it may well be if people are starting to horde WUs.
When I first made an observation in this thread, on the 4th, we were doing about 1.36 days work/day, that has now increased to nearly 1.5. So that small increase of ~1.004/day compounded over 22 days makes a significant increase.
Winterknight:
You are correct but the results based on linear regression of the entire sample have a standard deviation of about 2000 while the slope is about 50,000. With such a good fit I did not feel justified in fiddling with the sample.
The validated in last week number is much harder to interpret. It is so far out sync with the other numbers I have been ignoring it. I would definitely be interested in hearing more. I think it difference may be in the distinction between result and WUs.
Mike:
Thanks for pointing out that running out of new S4 WUs is not the same thing as running of new work.
RE: I have been tracking
)
Using the figures you have recorded, which show on average a gradual increase in units crunched per day. If this increase continues then, using a simple compoud formula, on those figures, the projected end of S4 will be in 19.45 days from the datetime of your last figure.
Andy
Did you all catch
)
Did you all catch this?
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Winterknight: I noticed the
)
Winterknight:
I noticed the increase but the sample is to small to justify anything more than simple linear projection.
You could be right, if this is continuing trend. Which it may well be if people are starting to horde WUs.
Looking at number of S4 units
)
Looking at number of S4 units still to be sent and the number returned and validated last week. Will all the S4 units, except for re-sends, have been sent in approx. ten days?
@Mike,
Yes, had seen that. Thanks.
Andy
RE: Winterknight: I noticed
)
When I first made an observation in this thread, on the 4th, we were doing about 1.36 days work/day, that has now increased to nearly 1.5. So that small increase of ~1.004/day compounded over 22 days makes a significant increase.
Andy
Less than 1,000,000 units
)
Less than 1,000,000 units left in S4 run.
Winterknight: You are correct
)
Winterknight:
You are correct but the results based on linear regression of the entire sample have a standard deviation of about 2000 while the slope is about 50,000. With such a good fit I did not feel justified in fiddling with the sample.
The validated in last week number is much harder to interpret. It is so far out sync with the other numbers I have been ignoring it. I would definitely be interested in hearing more. I think it difference may be in the distinction between result and WUs.
Mike:
Thanks for pointing out that running out of new S4 WUs is not the same thing as running of new work.