Hello,

you have to copy the needed libs into your library folder (for me /lib64).
then create a link:
e.g.:
lrwxrwxrwx 1 root root 21 29. Dez 23:46 libcudart.so.12 -> libcudart.so.12.0.107
lrwxrwxrwx 1 root root 21 29. Dez 23:52 libcudart.so.12.0 -> libcudart.so.12.0.107
-rwxr-xr-x 1 root root 675168 29. Dez 23:45 libcudart.so.12.0.107

after this is done mfactc.exe can execute the self test:

Selftest statistics
number of tests 26192
successfull tests 26192

kernel | success | fail
-------------------+---------+-------
UNKNOWN kernel | 0 | 0
71bit_mul24 | 2586 | 0
75bit_mul32 | 2682 | 0
95bit_mul32 | 2867 | 0
barrett76_mul32 | 1096 | 0
barrett77_mul32 | 1114 | 0
barrett79_mul32 | 1153 | 0
barrett87_mul32 | 1066 | 0
barrett88_mul32 | 1069 | 0
barrett92_mul32 | 1084 | 0
75bit_mul32_gs | 2420 | 0
95bit_mul32_gs | 2597 | 0
barrett76_mul32_gs | 1079 | 0
barrett77_mul32_gs | 1096 | 0
barrett79_mul32_gs | 1130 | 0
barrett87_mul32_gs | 1044 | 0
barrett88_mul32_gs | 1047 | 0
barrett92_mul32_gs | 1062 | 0

selftest PASSED!

The app must be static not dynamic which I have got.

Iam still trying to optimize the plan_class for cuda apps. There is a backward compatibility in drivers and cuda toolkit, so the latest change for Compute Capability 3.0-7.5 should only getting work with cuda100 apps. cuda110 will be deprecated soon, cant see a difference. Still observing any errors and sending out new work.

cuda12 for linux is now available and working.

new cuda12 app for windows is up

update:
A lot of windows hosts got download errors, I dont know why, changed the plan_class to send out work only to 4xxx cards.

update2:
second update of the app, an older wrapper was used, observing...

update3:
app is working now, supported for cc6-12 with cuda12

Can you change the mfaktc.ini configuration? GPU Sieve Size is set to 64, but the version supports up to 2047. There is quite a bit of performance to be gained with that change.

Not to mention the godawful coil whine my 4090 produces with this low amount.

rebirther, I noticed during the last week that some of my computers are showing a predicted run-time for TF gpu tasks of 1 second. This causes the machines to download large numbers of tasks. Is there a way to adjust the predicted run-time?

./mfaktc.exe: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.34' not found (required by ./mfaktc.exe)

I have already asked the user to recompile the app with an older Glibc version, most of the hosts have still 2.31, for the time being you will get a mix of cuda120 and cuda111 apps, the problem hosts getting some errors with cuda120 but can run cuda111 app.

./mfaktc.exe: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.34' not found (required by ./mfaktc.exe)

I have already asked the user to recompile the app with an older Glibc version, most of the hosts have still 2.31, for the time being you will get a mix of cuda120 and cuda111 apps, the problem hosts getting some errors with cuda120 but can run cuda111 app.

TF jobs makes my GPU hit power limit & no progress update after an hour, is this normal?

Dear all,

The Trial Factoring application doesn't run well on my GPU (AMD Radeon RX 7900 XTX, Ubuntu 22.04 x86_64). Completion percentage remains stuck at 0,01%, while ETA quickly goes up (several days) before I abort the task.

more stderr.txt 12:31:22 (1840): wrapper (7.16.26016): starting 12:31:22 (1840): wrapper (7.16.26016): starting 12:31:22 (1840): wrapper: running ./mfakto-x64 ()

What am I doing wrong ?

Best regards,

Samuel

Hello,

Sorry for the late reply.

I'm running BOINC headless, so no event log I'm afraid (I'm using boinctui).

I saved the "slot" directory of that failed task to my home directory though. Could there be some useful information in there ?

Best regards,

Samuel

Hello @rebirther,

I already provided you with that information on September, 24th:

more stderr.txt 12:31:22 (1840): wrapper (7.16.26016): starting 12:31:22 (1840): wrapper (7.16.26016): starting 12:31:22 (1840): wrapper: running ./mfakto-x64 ()

There's nothing else in that file, I'm sorry.

Best regards,

Samuel