BitBOINC

I have four computers, therefore four graphics cards. Oddly enough, they don't receive the same number of credits.

These: https://bitboinc.athena.org.tr/results.php?hostid=204&offset=0&show_names=0&state=4&appid=

and these: https://bitboinc.athena.org.tr/results.php?hostid=254&offset=0&show_names=0&state=4&appid=
receive more than 12,000 credits.

These: https://bitboinc.athena.org.tr/results.php?hostid=255&offset=0&show_names=0&state=4&appid=
and these: https://bitboinc.athena.org.tr/results.php?hostid=253&offset=0&show_names=0&state=4&appid=
but only a tenth of that. Very strange.

I urge the operators to put a stop to this absurdity as soon as possible.

This standard BOINC credit system isn't nonsense. It awards credits based on workload.

---

Why are only two out of four graphics cards receiving one-tenth of the credits of the other two? You can't tell me that's BOINC's fault. The credits are allocated by your server, not BOINC.

I should add that none of the four graphics cards have ever had this problem in any other project—only in yours. So please don't tell me that's normal.

Other projects use their own fixed credit systems rather than the BOINC system. I use the BOINC credit system directly. We initially implemented a fixed-rate credit system; as a result, bullies from other projects attacked us, and the statistics site expelled us.

https://github.com/BOINC/boinc/wiki/CreditNew

---

In reply to Alperen Yavuz's message of 1 Oct 2026:

This standard BOINC credit system isn't nonsense. It awards credits based on workload.

To be fair, a LOT of people think that CreditNew was and is a disaster that leads to unfair outcomes and unnecessary support load. Scoring based on the work actually done, either by fixed credit per task if appropriate, or by the validator "counting operations/scans/whatever", is preferable by far.

I am developing a credit system entirely from scratch so that low-end ARM-based devices won't end up as discarded junk; I intend to integrate this into every application. There will be a calculation delay of approximately 10–40 seconds, but everyone will receive the points they deserve based on the work they perform. I am currently testing this in a closed environment, and it might be ready in about 4–5 days. In the meantime, the BOINC credit system will remain in place—we have already tinkered with the credit system quite a bit. I may apply a retroactive credit adjustment later on.

---

After reboot one of the meanwhile 3 systems with low credits got normal credits. The last two are stuck
So i hope your new creditssystem will be fine and fast in place.

In reply to Sabroe_SMC's message of 1 Oct 2026:

After reboot one of the meanwhile 3 systems with low credits got normal credits. The last two are stuck
So i hope your new creditssystem will be fine and fast in place.

I dont know for sure your situation but if those computers are new BOINC may not run "Run CPU Benchmarks" yet. Once you run it, boinc would adjust computer info and this affects scoring on the projects that use standard credits

---

I crunch for Ukraine

Sadly BOINC doesn't run GPU benchmmark an uses fixed TFlops ratings based on GPU names. How this connects with CreditsNew, I don't know actually.

All my machines are now at the best way. All machines get the higher credit of 10-14K
I hope this delay doesn't affect all crunching machines.

See you at one billion credits - bye!