This is the error text from the rx9070xt....

Stderr output
<core_client_version>7.24.1</core_client_version>
<![CDATA[
<stderr_txt>
2026-09-07 11:18:28 (2820): wrapper: running prpll.exe (-d 0)
2026-09-07 11:18:28 (2820): wrapper: created child process 13852
20260907 11:18:28 PRPLL v8.0-2-g8fb01aa starting
20260907 11:18:28 config: -d 0
20260907 11:18:28 device 0, OpenCL 3679.0 (PAL,LC), AMD Radeon RX 9070 XT, unique id ''
20260907 11:18:28 83824649 No FFTs found in tune.txt that can handle 83824649. Consider tuning with -tune
20260907 11:18:28 83824649 config:
20260907 11:18:28 83824649 Using CARRY64
20260907 11:18:28 83824649 FFT: 2M 4:1K:4:256:101 (39.97 bpw)
20260907 11:18:28 83824649 LL loaded @ 0 : 0000000000000004
20260907 11:18:29 83824649 In file included from C:\Users\renix\AppData\Local\Temp\comgr-e3340e\input\CompileSource:1:
In file included from C:\Users\renix\AppData\Local\Temp\comgr-e3340e\include\fftp.cl:5:
In file included from C:\Users\renix\AppData\Local\Temp\comgr-e3340e\include\fftwidth.cl:18:
C:\Users\renix\AppData\Local\Temp\comgr-e3340e\include\fftbase.cl:1136:8: error: assigning to '__private F2' (aka '__private float2') from incompatible type 'ulong2' (vector of 2 'ulong' values)
1136 | base = U2(fma(a, -w.y, w.x), fma(a, w.x, -w.y));
| ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 error generated.
Error: Failed to compile source (from CL or HIP source to LLVM IR).

20260907 11:18:29 83824649 Compiling 'fftp.cl' error COMPILE_PROGRAM_FAILURE (-15) (args -cl-finite-math-only -cl-std=CL2.0 -DEXP=83824649ull -DWIDTH=1024u -DSMALL_HEIGHT=256u -DMIDDLE=4u -DCARRY_LEN=8u -DNW=4u -DNH=4u -DAMDGPU=1 -DCARRY64=1 -DFFT_VARIANT=101u -DMAXBPW=4750u -DWEIGHT_STEP=0.020511601490304197 -DIWEIGHT_STEP=-0.020099332002056696 -DTAILT=U2(-7.52981578e-05f,0.0122715384f) -DTRIG_SCALE=9 -DTRIG_SIN={6.6579027251980952e-07,3.7209369054580932e-23,-4.9188217704570848e-20,1.0901995091303198e-33,-1.1506191102407305e-47,7.0839240359575376e-62,-2.8545227803597818e-76,8.0307778151820938e-91,} -DTRIG_COS={1,-2.2163834349100114e-13,8.1872592175725296e-27,-1.209740380477418e-40,9.5758876440886212e-55,-4.7164085661095887e-69,1.5838111582820031e-83,-3.830342691138796e-98,} -DTAILTGF31=U2(509684486u,293249438u) -DTAILTGF61=U2(249938719029223731ull,1245372627562045535ull) -DFFT_TYPE=4 -DWordSize=8u -DDISTGF31=819200u -DDISTWTRIGGF31=2560u -DDISTMTRIGGF31=768u -DDISTHTRIGGF31=131776u -DDISTGF61=1638400ull -DDISTWTRIGGF61=3072ull -DDISTMTRIGGF61=1792ull -DDISTHTRIGGF61=263040ull -DFRAC_BPW_HI=4169156607u -DFRAC_BPW_LO=4294967295u -DFFT_FP64=0 -DFFT_FP32=1 -DNTT_GF31=1 -DNTT_GF61=1 )
20260907 11:18:29 83824649 Can't compile fftp.cl
20260907 11:18:29 Exception "Can't compile fftp.cl"
20260907 11:18:29 Bye
2026-09-07 11:18:29 (2820): prpll.exe exited; CPU time 0.015625
2026-09-07 11:18:29 (2820): called boinc_finish(0)

</stderr_txt>
<message>
upload failure: <file_xfer_error>
<file_name>LL-dc_83.79-83.83M_wu_388_12_0</file_name>
<error_code>-240 (stat() failed)</error_code>
</file_xfer_error>
</message>
]]>

So just going to crunch llr with nvidia only. :-)

Adrenalin 26.9.1 for the 9070xt. The TF units crunch fine for both thr 9070 and 7900gre. But fail for both gpu's on llr. One of the earlier llr units failed with the previous adrenalin release (gre).

It is possible extend deadline of 52971783?

Task was near finish, but it restarted with lower numbers after PC shutdown for few hours. I don't know why, I have several restarts before, all was OK.

20260912 01:55:16 83796667 74260000 c435ff51069a07c1 7702 ETA 20:24
20260912 01:57:50 83796667 74280000 36b35ddb8025d9ee 7656 ETA 20:14
20260912 02:00:25 83796667 74300000 d6dcce2b2abbc2f6 7777 ETA 20:31
20260912 02:03:00 83796667 74320000 fd5e99881f1f73bf 7747 ETA 20:24
20260912 02:05:34 83796667 74340000 ca75f1665115dd39 7700 ETA 20:14
20260912 02:082026-09-12 11:02:48 (7496): wrapper: running prpll.exe (-d 0)
2026-09-12 11:02:48 (7496): wrapper: created child process 1792
20260912 11:02:49 PRPLL v8.0-2-g8fb01aa starting
20260912 11:02:49 config: -d 0
20260912 11:02:49 device 0, OpenCL 581.57, NVIDIA GeForce GTX 1050 Ti, unique id ''
20260912 11:02:49 83796667 No FFTs found in tune.txt that can handle 83796667. Consider tuning with -tune
20260912 11:02:49 83796667 config:
20260912 11:02:49 83796667 Using CARRY64
20260912 11:02:49 83796667 FFT: 2M 4:1K:4:256:101 (39.96 bpw)
20260912 11:02:49 83796667 File 'D:\BOINCproj\slots\8\ll-83796667\83796667-074460000.ll' : CRC: expected 2315453555, actual 3384082823
20260912 11:02:49 83796667 Bad savefile 'D:\BOINCproj\slots\8\ll-83796667\83796667-074460000.ll'
20260912 11:02:49 83796667 Removing bad savefile 'D:\BOINCproj\slots\8\ll-83796667\83796667-074460000.ll'
20260912 11:02:50 83796667 LL loaded @ 61000000 : 1bc3696a3d9b9109
20260912 11:05:30 83796667 61020000 0774df0727cc3aab 7991 ETA 2d 02:34
20260912 11:08:01 83796667 61040000 f04b72f7f8dec051 7577 ETA 1d 23:54
20260912 11:10:33 83796667 61060000 aae2d45c4ab42bff 7574 ETA 1d 23:50
20260912 11:13:06 83796667 61080000 6dcfbc95a5d0d537 7649 ETA 2d 00:16
20260912 11:15:38 83796667 61100000 8f3589d3290a2d62 7600 ETA 1d 23:55
20260912 11:18:09 83796667 61120000 d7bb9d5a8abb95e7 7575 ETA 1d 23:43
20260912 11:20:41 83796667 61140000 524c5b9595cf2b32 7579 ETA 1d 23:42

Looks like it is impossible to finish till tomorrow evening. Thank You.

extended, pls check because the hosts are hidden.

If there is a new Mersenne prime number hidden in the Lucas-Lehmer double-checks, what is the range of the number of digits for the discovered prime number?

I've tried crunching with two AMD gpu's (9070xt, 7900gre) and they fail. I've noticed others have had the same problem Is there a functional wrapper etc., for these two amd gpu's yet? I have absolutely no problems with my 40/50 series nvidias. O/S is win11.

Just for interest.

Does server have any deadline tolerance for WUs?

As I know, some other project have various strategies for this: no tolerance, few hours tolerance, tolerance till WU is resent to other user.

How about SRBase?

Today I first time got 2 040 000 for task. It was strange, because all previous was paid with 1 360 000.

I checked few top hosts in SRBase and not noticed any differences of runtime for both credits.

Can you explain? Thank You.

Here were any changes in this application runtime in last two weeks?

I got new task for my 1050ti, and in stderr.txt I see 17 days left instead of 6-7 days before.

Thank You.

If you are running v0.12 then temporary add -fft 4:1K:4:256 to your app_config. The FFT size was changed from 2M to 4M which doubled the runtime. v0.13 fixed that issue after reported the bug.