OFFICIAL "SERVER PROBLEMS" THREAD

[Dobri]

Originally Posted by Dobri

M101969381 is unassigned for a LL/PRP test and does not appear as 'Avaiable' in the Work Distribution Map.

Originally Posted by Dobri

The same server problem that occurred in March of this year in the 101-xxxxxx range is now observed for the current lowest untested exponent 103313627 in the 103-xxxxxx range.
It remains unassigned since mid-May, see the attached images.

P.S. Note that said server problem in the 101-xxxxxx range still needs a closure in the Work Distribution Map, because 54316 < 35141 + 14718 + 4458.
54316 | 35141 14718 4458

This time the current lowest untested exponent M104044933 remains unassigned since 2021-01-25, see
<https://www.mersenne.org/report_exponent/?exp_lo=104044933&exp_hi=&full=1>.

[Prime95]

Originally Posted by Dobri

This time the current lowest untested exponent M104044933 remains unassigned since 2021-01-25, see
<https://www.mersenne.org/report_exponent/?exp_lo=104044933&exp_hi=&full=1>.

I've put in a fix. The problem had to do with the server not wanting to hand out PRP assignments for exponents that had previously been assigned for LL.

[mrh]

One of my systems finished a PRP and got a couple of TFs assigned for some reason. It finished 222072449 before I noticed and was working on 222072469.

Didn't hurt anything, but seems really odd since it has WorkPreference=150. Probably just an odd server issue?

[Zhangrc]

Originally Posted by mrh

Probably just an odd server issue?

Really odd. Yesterday Ben Delo did several TF assignments too, like M222075827 and M222075551.

[slandrum]

And in this post https://www.mersenneforum.org/showpo...postcount=5189 someone is complaining that there are no TF assignments available for 73-74.

[Prime95]

Originally Posted by Zhangrc

Really odd. Yesterday Ben Delo did several TF assignments too, like M222075827 and M222075551.

My bad -- bug fix went awry.

[Prime95]

Originally Posted by Zhangrc

Really odd. Yesterday Ben Delo did several TF assignments too, like M222075827 and M222075551.

My bad -- bug fix went awry.

GPUowlers beware -- TF assignments may have been added to your worktodo.txt. I've no idea if that is handled gracefully.

[kriesel]

Originally Posted by Prime95

GPUowlers beware -- TF assignments may have been added to your worktodo.txt. I've no idea if that is handled gracefully.

I'm skeptical that gpuowl's primenet.py even would pass them.

workTypes = dict(PRP=150, PM1=4, LL_DC=101, PRP_DC=151, PRP_WORLD_RECORD=152, PRP_100M=153, PRP_P1=154)

But, tests by manual insertion into worktodo shows gpuowl will ignore them if TF assignments make it into a worktodo file:

2021-07-27 10:58:00 gpuowl v6.11-364-g36f4e2a
2021-07-27 10:58:00 config: -user kriesel -cpu asr2/radeonvii0 -d 0 -maxAlloc 13000 -proof 10 -use NO_ASM -cleanup -block 2000 -log 100000 -use OUT_WG=256,OUT_SIZEX=8,OUT_SPACING=2
2021-07-27 10:58:00 device 0, unique id ''
2021-07-27 10:58:00 asr2/radeonvii0 worktodo.txt line ignored: "Factor=(aid),122863843,74,76"

(continues on to next worktodo entry)

2021-07-27 11:03:00 GpuOwl VERSION v7.2-53-ge27846f
2021-07-27 11:03:00 config: -user kriesel -cpu asr2/radeonvii2 -d 2 -maxAlloc 14G -proof 10 -use NO_ASM -autoverify 10 -block 1000 -autoverify 11
2021-07-27 11:03:00 device 2, unique id ''
2021-07-27 11:03:00 asr2/radeonvii2 worktodo.txt line ignored: "Factor=(aid),122863843,74,76"

(continues on to next worktodo entry)
You'd have to be running ~v3.7 on Linux with ROCm for TF to be supported in Gpuowl.
I think any TF that did make it into a recent Gpuowl version's worktodo file would need to be either manually unassigned and deleted, or manually relocated to mfaktx.

[Viliam Furik]

On the manual results submission page, when I submit X newly found factors and Y already known, it displays them in the table at the bottom as Y new, X old.

After the next submission, I can add a picture.

I am doing TF on factored 10M exponents from 67 to 68 bits. Some of the factors are known or known composites, but some are new.

[James Heinrich]

Originally Posted by Viliam Furik

On the manual results submission page, when I submit X newly found factors and Y already known, it displays them in the table at the bottom as Y new, X old.
After the next submission, I can add a picture.
I am doing TF on factored 10M exponents from 67 to 68 bits. Some of the factors are known or known composites, but some are new.

I'm not sure if this is a comment, question, suggestion, or bug report. Perhaps a screenshot would help, with description of what you would like changed (if applicable).

[Viliam Furik]

Originally Posted by James Heinrich

I'm not sure if this is a comment, question, suggestion, or bug report. Perhaps a screenshot would help, with description of what you would like changed (if applicable).

A bug report. The numbers of factors seem to be misplaced. Screenshot coming soon.