Can someone explain the following?
This host, a new laptop, claims almost twice the amount of credits as granted.
This host of mine claims for the same CPU time the credits that are granted.
Both are using stock apps, no OC and the granted credits are the same.
I do not see this behavior at Seti or Einstein.

---

Can someone explain the following?
This host, a new laptop, claims almost twice the amount of credits as granted.
This host of mine claims for the same CPU time the credits that are granted.
Both are using stock apps, no OC and the granted credits are the same.
I do not see this behavior at Seti or Einstein.

The claimed credit is based on the boinc benchmark which doesn't take things like cache size into account. Fortunately the granted credit is much closer to showing your cpu's throughput. There are threads explaining the difference between claimed and granted. The only value that's important is the credit granted per unit of time though. HTH, Danny

---

Thanks for the answer, but I have to disagree about the benchmark.
BOINC uses the bechmark to estimate how many work can be done, so no WU will timed out. It is a local proces on the host.
To calculate the (claimed) credits FLOP counts are used.

When its depending on the benchmark you would see the same behavior, as noticed here, in other projects. And that is not the case. :D

---

It is true, the benchmarks are used local to the client machine to estimate proper size of work requests. But they are also reported back with your results along with the number of CPU seconds used to complete the task. And this is presented as the credit claim when you view your R@h tasks. A faster machine that scores higher benchmarks will show a higher claim per CPU second.

See the discussion here.

---

Rosetta Moderator: Mod.Sense