The advanced Gerbicz error checking in LLR2 is both a blessing and a curse. The fantastic news is that most errors that occur in LLR -- which would normally cause a task to fail validation and completely waste the entire calculation -- are now corrected and are able to complete.

The flip side of this is that if your computer is completing valid tasks even though its malfunctioning, so there's no easy way to know that you have a problem. The computer is making hardware mistakes, but we're correcting them and the problem is hidden from the computer's owner. If you never know the computer isn't working correctly, you can't fix it.

This was driven home this morning when Honza told me one of his teammates had seen a few of these Gerbicz errors, and I decided to look in the database and see how common they were.

I found that not only was Honza's teammate getting these errors, but so was Honza, and he didn't even know it.

Another person getting errors without knowing it? ME. Yes, my brand new Ryzen was having problems and I had absolutely no idea anything was wrong!

The results page now displays an unmistakable WARNING if a task had Gerbicz errors.
____________
My lucky number is 75898⁵²⁴²⁸⁸+1

Is it also noticeable by longer completion times on the tasks with errors?
____________
1281979 * 2^485014 + 1 is prime ... no further hits up to: n = 5,700,000

Nice, I take it the "warning" will be in the status column? Fortunately I just skimmed through the results I have remaining on the system and didn't see any. Maybe add a "warning" category for the state filter so it is easier to see if there are any such tasks?

Is it also noticeable by longer completion times on the tasks with errors?

At what point should we start to be concerned? I have 8 valid tasks currently listed and a single warning.
At some point could we add timestamps to each line in the stderr file? It would make it much easier to tie errors to what else might have been happening on the pc in question.

Nice, I take it the "warning" will be in the status column? Fortunately I just skimmed through the results I have remaining on the system and didn't see any. Maybe add a "warning" category for the state filter so it is easier to see if there are any such tasks?

At what point should we start to be concerned? I have 8 valid tasks currently listed and a single warning.
At some point could we add timestamps to each line in the stderr file? It would make it much easier to tie errors to what else might have been happening on the pc in question.

At what point should we start to be concerned? I have 8 valid tasks currently listed and a single warning.
At some point could we add timestamps to each line in the stderr file? It would make it much easier to tie errors to what else might have been happening on the pc in question.

Replace “warning” with “invalid task” and what answer do you get? If you suspect an outside event may have caused it, then just wait and see if more happen.

The same warning message now appears if certain errors appear occur in PPS-Sieve tasks on AMD GPUs.
____________
My lucky number is 75898⁵²⁴²⁸⁸+1

Hi I get 2 error message at task results which is "Error while computing WARNING!"
When I hover mouse over warning shows " Errors occurred and were corrected during this calculation. Your computer is not operating correctly. This is a hardware problem which you should fix." What is that mean ? I am not change any option or hardware and also 1 day ago completed TRP tasks. Any advice appreciated.

http://www.primegrid.com/result.php?resultid=1169128644
http://www.primegrid.com/result.php?resultid=1167613948

Firstly no OC no boost bios all default values I didn't change any values last few weeks. MB MSI mag 570x latest bios installed. CPU Ryzen 2700x with water cooling. avg temp 60-65 Celsius. Rig has 3 intake 1 out additional fans. No additional OC app installed like MSI dragon only for temps CPUID HW Monitor installed. I will try to see what's going on.

There is an opportunity suggested here to create a tool that is very sensitive to computer stability. During the run-up to the current challenge, I noticed these "Gerbicz errors" in one of the output logs on a Ryzen 3700X, while testing an offline WU, but only with certain task/thread combinations, notably 8 tasks of 2 threads. The problem is yet unresolved, but for now running 8 tasks/1 thread works, and also provides slightly higher throughput. So even though this PC is not overclocked, and can run Memtest all night with no errors, it has a problem!

There is an opportunity suggested here to create a tool that is very sensitive to computer stability. During the run-up to the current challenge, I noticed these "Gerbicz errors" in one of the output logs on a Ryzen 3700X, while testing an offline WU, but only with certain task/thread combinations, notably 8 tasks of 2 threads. The problem is yet unresolved, but for now running 8 tasks/1 thread works, and also provides slightly higher throughput. So even though this PC is not overclocked, and can run Memtest all night with no errors, it has a problem!

Not really disappointing, hyperthreading is slower for LLR on all processors.
I've had 4 different ryzen processors and not noticed any instabilities when not trying to push things - although the 3700x will run super hot if you leave PBO set to AUTO rather than switching it off.
No idea what's up with Michael's but I'd suspect the motherboard and/or bios rather than the CPU if the RAM won't run at advertised speed - I did have to return a b450 board that refused to post with more than 1 stick of RAM.