Somebody stepped on my feet...
 

I had reach to an old PC, and I decided to choose worktype 1 - Trial factoring to low limits. I was then assigned 100 exponents on the range of 216,xxx,xxx from 68 to 69 bits through primenet on the PC named CS50X-C9.

After some days of worling, I had to leave my PC unattended, and as always happens on those cases, the PC stopped working.
20 days after, I came back and turned the PC back on, but the long list of TF exponents I had reserved were all taken to 70 bits by user Yxinity using the manual reserveation page.

How could that happen? Now I reverted to ECM on small numbers on such PC, but I keep wondering why I lost the 100 exponents I had reserved.

Yxinity commonly poaches TF work without getting assignments.

[QUOTE=Mark Rose;478729]Yxinity commonly poaches TF work without getting assignments.[/QUOTE]

Good to know... I will grab less exponents next time :smile:

[QUOTE=ET_;478779]Good to know... I will grab less exponents next time :smile:[/QUOTE]
GPUs are so much better at trial factoring than cpus, you didn't lose much anyway. Consider putting the old box on double-checking. The slower the box, the higher "category" you'll want so they don't expire or get unreserved before completion. The thresholds shift over time; determine the box's speed and plan accordingly. [URL]https://www.mersenne.org/thresholds/[/URL]

[QUOTE=kriesel;479145]GPUs are so much better at trial factoring than cpus, you didn't lose much anyway. Consider putting the old box on double-checking. The slower the box, the higher "category" you'll want so they don't expire or get unreserved before completion. The thresholds shift over time; determine the box's speed and plan accordingly. [URL]https://www.mersenne.org/thresholds/[/URL][/QUOTE]

You are completely right, but I have no box to use with my GTX 980, my TF position on the stats is lowering, my cloud server slice of power is 10%, and trial-factoring to low limits is the best I can assign to it, followed by PRP double-check on Mersenne cofactors. :smile:

In other words, when I have a slow machine, I prefer it doing ALL of the possible slow kind of works instead of just one.

Yxinity also causes me problems. It acts very chaotically, unpredictably.

Well does anyone knows why Yxinity does this toestepping? When he poach work, do I still get credit for work done or is the derived "not needed" equal to 0 credit?

For some reason, the server doesn't accept duplicate TF results. However, you can ask George to manually credit you. Keep in mind that it may be a while before you get a response.

[QUOTE=ixfd64;491859]For some reason, the server doesn't accept duplicate TF results. However, you can ask George to manually credit you. Keep in mind that it may be a while before you get a response.[/QUOTE]

Well I think Georges time is better spent on developing AVX-512 however I just don´t get it since I was and still am making huge progress on my range - that Yxinity feels like he/she has to do work that I have already reserved. There is litterally billions of GHz days worth of work remaining and what I had reserved was actually in no way in risk of stalling the completion of the 68 bit TF range for first time LL/PRP candidates - so maybe it is time that the original reserver (even if he/she uploads a result after the poacher) is the one that get the credit while the poacher loose out on the credit :smile:

In my experience, Yxinity poaches work from nearly done ranges. If you decide to take a batch of exponents up a bit level, I'd submit all your work all at once.

[QUOTE=Mark Rose;491914]In my experience, Yxinity poaches work from nearly done ranges. If you decide to take a batch of exponents up a bit level, I'd submit all your work all at once.[/QUOTE]

Well I did plan to take them up 4 bits, but now I´m changing the strategy and goes up 1 bit and then switch to pure PRP testing - however I do not have time or desire to have to micromanage any projects at all, so even when I get my new system build, I would really like to take up a month wort of work without having to keep eye for poachers :smile: ... but thanks for the advice, it at least gives me a lead to a poachfree strategy :smile: