Vato
GTX1650super
(and maybe GT1030 if it works)

biodoc
RTX5080 / 16 GB
Driver Version: 580.178.04
Linux

GTX 1070 results:

https://boinc.eternityathome.org/eternity/workunit.php?wuid=3248611

I'll give it a go on my 2080 Super.

RFGuy_KCCO

3C-714
RTX 4060

Thanks mmonnin, and thanks again for the log you sent: it showed a real bug on our side, not a limitation of your 3070 Ti. The GPU sat idle while the last deep endgames ran on the CPU, so the task stayed just above the point where it hands its last tickets over to CPU tasks. Version 2.70 fixes this with an earlier handoff when the GPU stays idle, and your host has been returning valid results with it since.

On credit: the first GPU tasks had a provisional rate. GPU credit is now calibrated on measured work, on the same basis as CPU credit (about 1.34 credits per ticket). It is paid per ticket, so every card earns in proportion to the work it actually does.

Werinbert

Host 30 -- RTX 5060
Host 31 -- RTX 4060 ti

Thank you all for joining the beta, the response has been fantastic!

Already enabled and returning valid GPU results: Hans Sveen (RTX 4080 and 3080 Ti), Dargo75, Freewill (and thanks for the power figures: 165 W on the Titan V is a nice result), Bill F (GTX 1660 SUPER), FritzB, KetamiNO, biodoc (RTX 5080, Linux), 10esseeTony (including the GTX 1070, which now works with version 2.69 for older cards).

Enabled, first tasks on the way: polaxis, RFGuy_KCCO (RTX 2080 SUPER, Linux), 3C-714 (RTX 4060).

Vato: your hosts are enabled, but they have not picked up GPU work yet. Please make sure "Use NVIDIA GPU" is ticked in your Eternity project preferences, then allow new tasks and click Update.

As a reminder: please run only one GPU task per card during the beta, and you do not need to install CUDA or reset anything. Keep the reports coming!

Hi!
During the night(in Europe!) this unit https://boinc.eternityathome.org/eternity/workunit.php?wuid=3309085 stalled at about 99 % done and time
remaining didn't decrease, aborted this morning to get new work! This is on Pc Id 166!

On the 3080 Ti card pc Id 165 got one wu (https://boinc.eternityathome.org/eternity/workunit.php?wuid=3324409) that didn't validate!

On both of them credits decreased/dived for every utits done during the afternoon and night , but seems to have gone back to previous "normal"

Hans S.

PS:

Just received some explanation for the problem by the news released 😉👍

@Vato: your three PCs are enabled, but none of them has contacted the project since 12 October, so the server can't send them anything yet. Please start them again, allow new tasks and click Update in BOINC. Nothing else is needed.

I'll give it a try.
Username taurec
3060ti - Driver Version: 580.178.04
Linux

In reply to benjamin's message of 21 Oct 2026:

Already enabled and returning valid GPU results: FritzB, KetamiNO, biodoc (RTX 5080, Linux)

Just some infos:
My RTX 5080 (host 1912) is running on Win 11 :)
The GTX 1650 Ti (host 1988) is a laptop GPU.

Host 1989 shows 2x RTX 5070 Ti. In fact, there is one 5070 Ti and one 4070 Ti installed. BM is reporting it correctly but as in other projects two 5070 Ti are shown.

| CUDA: NVIDIA GPU 0: NVIDIA GeForce RTX 4070 Ti (driver version 616.92, CUDA version 13.4, compute capability 8.9, 11.99GB, 11.99GB available, 40090 GFLOPS peak)
| CUDA: NVIDIA GPU 1: NVIDIA GeForce RTX 5070 Ti (driver version 616.92, CUDA version 13.4, compute capability 12.0, 15.89GB, 15.89GB available, 44746 GFLOPS peak)
| OpenCL: NVIDIA GPU 0: NVIDIA GeForce RTX 4070 Ti (driver version 616.92, device version OpenCL 3.0 CUDA, 11.99GB, 11.99GB available, 40090 GFLOPS peak)
| OpenCL: NVIDIA GPU 1: NVIDIA GeForce RTX 5070 Ti (driver version 616.92, device version OpenCL 3.0 CUDA, 15.89GB, 15.89GB available, 44746 GFLOPS peak)

I'm running out of Betas. Do you detach any hosts, if they are no longer needed for testing?

i have enabled my 3 hosts and none of them download any tasks so far

Quite a few users have Nvidia Tesla V100 GPU because they are relatively cheap, but pretty good (12 nm, with HBM memory) at crunching on BOINC.

If you approve it:

BorgProcessor
Nvidia Tesla V100, 16 GB HBM memory
Windows 10
Driver 31.0.15.2889 (NVIDIA 528.89) (Win 10) - its like a workstation driver package

Credit doesn't seem to be consistent with either the run time or CPU time.

1,305.20 1,301.77 69.67
3,274.70 3,274.70 2,234.76
8,460.17 8,460.17 50,198.99
15,306.00 47,977.11 87,825.46
9,289.00 11,495.39 87,313.66

I'll give them a go

Lazarus-uk

GTX 1660 Ti (driver version 560.94)

Window 10 Pro x64

In reply to mmonnin's message of 21 Oct 2026:

Credit doesn't seem to be consistent with either the run time or CPU time.

bejnamin explained it here:

On credit: the first GPU tasks had a provisional rate. GPU credit is now calibrated on measured work, on the same basis as CPU credit (about 1.34 credits per ticket). It is paid per ticket, so every card earns in proportion to the work it actually does.

Then GPUs are not worth running the small tasks. Power wise a CPU will be better.

Also why do the beta CPU tasks have a 2 hour deadline? It interrupts anything else that is running

In reply to mmonnin's message of 21 Oct 2026:

Then GPUs are not worth running the small tasks. Power wise a CPU will be better.

Also why do the beta CPU tasks have a 2 hour deadline? It interrupts anything else that is running

You're right that the small ones are poor value, and FritzB is right about the per-ticket rate (thanks FritzB!). The small tasks you got were test and recovery tasks: a few hundred tickets, plus several minutes of fixed start-up, so per hour they pay badly. They are not the normal size. Your two recent full-size tasks (65,650 tickets) each earned 87,825 credits in about 3 h 10 on the 3070 Ti, roughly 27,600 credits per hour. For comparison, the same 32-thread machine on CPU work would earn around 9,600 per hour on our CPU basis.

About the short deadline: the CPU tasks that finish the last tickets of a GPU batch are kept short on purpose, so a batch doesn't stay open for days. You're right that it pushes other work aside; new ones now have a 4-hour deadline instead of 2.

In reply to BorgProcessor's message of 21 Oct 2026:

Quite a few users have Nvidia Tesla V100 GPU because they are relatively cheap, but pretty good (12 nm, with HBM memory) at crunching on BOINC.

If you approve it:

BorgProcessor
Nvidia Tesla V100, 16 GB HBM memory
Windows 10
Driver 31.0.15.2889 (NVIDIA 528.89) (Win 10) - its like a workstation driver package

Thanks for offering your V100! Your 528.89 driver meets our Windows minimum of 528.33, so no driver update is required before the first beta test. We cannot see the V100 in your BOINC hardware inventory yet. Could you connect/update that PC and send us its BOINC host ID? We can then check and enable the correct machine.