https://youtu.be/SGZqDzb__bw

Since this is first LLR2 challenge and it is only 5 days long, I will put all my CPU resources to this: there is no need to hurry , to throw resources to be first, now you are first all the time.
PG become fair even to oldest, and slowest CPU :)
____________
92*10^^1439761-1 REPDIGIT PRIME :) :) :)
314187728^¹³¹⁰⁷²+1 GENERALIZED FERMAT
31*332^³⁶⁷⁵⁶⁰+1 CRUS PRIME
Proud member of team Aggie The Pew. Go Aggie!

Make sure you check the best way to run these - the options page says multithreading is recommended but in my ryzen 3700 tests 8x single threads looks about 10% faster than 2x 4 threads. I've not tested 4x2 but prime95 indicates it's around the same as 2x4.

Looks like they're 384k FFT, or around 3MB/task for cache size consideration. Zen 2 with 4MB/core would have best throughput run one core per task. Most other CPUs would run better with 2 or possibly more cores per task.

Make sure you check the best way to run these - the options page says multithreading is recommended but in my ryzen 3700 tests 8x single threads looks about 10% faster than 2x 4 threads. I've not tested 4x2 but prime95 indicates it's around the same as 2x4.

What a nice view :20 cores 20 work units :)
I start challenge right now to see how nice is LLR2 :)
____________
92*10^^1439761-1 REPDIGIT PRIME :) :) :)
314187728^¹³¹⁰⁷²+1 GENERALIZED FERMAT
31*332^³⁶⁷⁵⁶⁰+1 CRUS PRIME
Proud member of team Aggie The Pew. Go Aggie!

Got "some" systems on a rolling start. It seems cool enough recently I might be able to put more power on than earlier this year.

Let's hope they don't fix it :-)

I saw. Debating what to do about it... it is a recent rebuild. CPU is stock. Cooler Noctua D15. Ram I haven't even turned on XMP yet so only running 2400, although these tasks will be in-cache so ram shouldn't affect it anyway.

The same CPU was probably fine in another mobo though, which I no longer have.

I might try reducing PPT limit from stock 88W to something lower.

As I mentioned on the other thread, the main thing is to make sure PBO is off. Your cpu will thank you just for that.
The other usual suspect when you start getting errors is the psu. Are you using something decent?

PBO is by default off as it is an overclock. Apart from some early testing, I never turn it on. It is not a factor in my case.

The PSU is a highly rated 850W unit so is unlikely to be a problem.

PBO is by default off as it is an overclock. Apart from some early testing, I never turn it on. It is not a factor in my case.

The PSU is a highly rated 850W unit so is unlikely to be a problem.

I've had 3 ryzen boards, from 3 different manufacturers - b450, x470, x570 and all have had PBO=on as the default

You really shouldn't be seeing errors from the standard boost as that should be well within tolerances.
These things can be hard to track down.
I'd give memtest a run, or put in some alternate memory if you have it, just to rule out a memory problem. I'm probably preaching to the choir be make sure you have enough airflow over the board as well.

Are you 100% sure PBO was on, not just PB? Depending on the CPU, it is an easy check to see what power draw is reported under a heavy load like prime finding. 65W TDP CPUs run at 88W PPT limit as stock. I find it hard to believe PBO enabled is default as that is an overclock. The only maybe is sometimes it might get tied in with enabling XMP but I haven't seen that on AMD side, only Intel side.

yes, here's a screenshot of the default settings

yes, here's a screenshot of the default settings

"Auto" does not necessarily mean "on". I'd look to verify what power it is actually limiting at to determine if you are actually running PBO or not. That's looking for consumption above 88W on 65W TDP CPUs, and I think 142W on 105W TDP CPUs.

It would be verification of the state. A drop in temp is suggestive, but doesn't entirely confirm it. It would be bad form for mobo makers would overclock CPUs as default, and I've not come across this before hence I'd like to be sure that is actually what's going on.


Anyway, back onto my 3700X system, it seemed to have rebooted not long after I disabled boost earlier. I've got it going again on no new work, and once it has drained I'll retire it from crunching. Don't have time to tinker with it at the moment as it is already past my bed time. Maybe I'd try a positive voltage offset tomorrow.

Edit: forget it, I just saw it return a unit early with error state, round off errors. This is a bigger problem now.

We've dropped one position :(

Yesterday someone release huge amount of bunkered workunit, drop by one place was imminent
____________
92*10^^1439761-1 REPDIGIT PRIME :) :) :)
314187728^¹³¹⁰⁷²+1 GENERALIZED FERMAT
31*332^³⁶⁷⁵⁶⁰+1 CRUS PRIME
Proud member of team Aggie The Pew. Go Aggie!

Yesterday someone release huge amount of bunkered workunit, drop by one place was imminent

Yesterday the server was down for a moment. It begs the question if bunkering had anything to do with it.

Will we make it to third ?
Hope so.
The gap to CNT seems to be shrinking.