Hello, from another newbie

[VBCurtis]

Yep, that makes a lot more sense!

Also, note that with your powerful cooler keeping things under control, the risk posed by adding CPU voltage is negligible. If you were running at 70C+, the added voltage might cause more problems from heat than solutions from cleaner signals in the chip.

Think of heat as vibrations that create background noise in the chip, and voltage as a sort of volume control on the signals passed around in the chip. Errors happen when the signals can't be heard over the noise- so with your great cooler you have less such vibrational noise, but it's possible a bit more voltage is needed for all the signals to be clean and error-free.

I've had older machines that, when I added a great cooler to drop temps from 70 to 50C, I found I could reduce voltage on the CPU- which then created less heat, leading to even lower temps and less power usage by a few watts.

[mathematizer]

Is there a way to see whether my proof file was uploaded for a recent PRP that I did on gpuOwl v7.1-11?

I submitted the copied/pasted composite result into the manual testing results page, then copied the .proof (v2) file into the Prime95 directory.

I'm not sure if there's more I ought to do on my end.

[moebius]

Originally Posted by mathematizer

Is there a way to see whether my proof file was uploaded

Look in the prime.log file of prime95.

[kriesel]

Originally Posted by mathematizer

Is there a way to see whether my proof file was uploaded for a recent PRP that I did on gpuOwl v7.1-11?

I submitted the copied/pasted composite result into the manual testing results page, then copied the .proof (v2) file into the Prime95 directory.

I'm not sure if there's more I ought to do on my end.

It can take a while. Wait at least 2-3 hours for prime95 to do the upload. (Longer if it's having problems accessing the server reliably.) Check the prime95 results.txt file for error messages such as shown in https://www.mersenneforum.org/showpo...0&postcount=26
Check prime95's allowances for upload traffic rate, and adjust expectations.

Check the exponent's status. When it's indicated as assigned as a Cert to someone, or lists a Cert result, the upload was completed successfully.
Another way to tell is to attempt standalone uploader.exe upload. If it replies the file was already uploaded, you have your answer. Use that one sparingly, to not waste server resources.

[kriesel]

Originally Posted by moebius

Look in the prime.log file of prime95.

Cert file upload successes get logged there, and factoring or primality test result records. I don't see anything there indicating logging of successful uploads of PRP proof files.

For the same exponent 164000087, which successfully uploaded a proof file, and was subsequently verified by someone else:

results.txt

(no corresponding entry)

results.json.txt

[Sat Oct 24 16:40:38 2020]
{"status":"C", "exponent":164000087,  "worktype":"PRP-3", "res64":"8BFDAB6FD520EBCC", "residue-type":1,  "res2048":"(redacted)", "fft-length":9175040, "shift-count":79095424,  "error-code":"00000000", "security-code":"DAAA46D2",  "program":{"name":"Prime95", "version":"30.3", "build":6, "port":4},  "timestamp":"2020-10-24 21:40:38", "errors":{"gerbicz":0},  "proof":{"version":1, "power":7, "power-multiplier":2, "hashsize":64,  "md5":"(redacted)"}, "user":"Kriesel", "computer":"ostrich",  "aid":"(redacted)"}

prime.log

[Sat Oct 24 14:00:59 2020 - ver 30.3]
Sending result to server: UID: Kriesel/ostrich, M164000087 is not prime.  RES64: 8BFDAB6FD520EBCC. Wh4: DAAA46D2,79095424,00000000, AID: (redacted)

PrimeNet success code with additional info:
CPU credit is 1076.2506 GHz-days.

[mathematizer]

Thanks again, just want to make sure I have the right workflow down.

I recently checked the exponent and see it was assigned as a cert to someone, so it looks like things are running smoothly.

I love the speed the GPU churns through PRPs. gpuOwl seems to be working well, though I see there are a handful of other options out there for me to look into. I mainly want to contribute to the project with work that offers a small chance at finding a prime.

Still working on tweaks to the CPU, closing in on 1 day with +.05v and XMP off, without errors. Hoping to go 4-7 days before adding the other two sticks of RAM and turning XMP back on.

[moebius]

Originally Posted by kriesel;561711prime.log
[CODE

[Sat Oct 24 14:00:59 2020 - ver 30.3]
Sending result to server: UID: Kriesel/ostrich, M164000087 is not prime. RES64: 8BFDAB6FD520EBCC. Wh4: DAAA46D2,79095424,00000000, AID: (redacted)

PrimeNet success code with additional info:
CPU credit is 1076.2506 GHz-days.[/CODE]

I uploaded one .proof with Prime95 30.3b6 and received a message in the communication thread: proof file successfully uploaded. Shouldn't this already be visible in prime.log?

[Runtime Error]

We match on 100989029!!!

I'd still recommend finding the source of the error because you might want to return the faulty hardware while still under warranty. But it seems that it did not affect the end result on this exponent. Happy hunting!

[kriesel]

Originally Posted by moebius

I uploaded one .proof with Prime95 30.3b6 and received a message in the communication thread: proof file successfully uploaded. Shouldn't this already be visible in prime.log?

Open your prime.log file in a text editor or use the command prompt find command to search for the string "proof" in prime.log.

Should, yes, I think it should be there in prime.log. That doesn't mean it IS there. George is free to make such design decisions.

[mathematizer]

Originally Posted by Runtime Error

We match on 100989029!!!

I'd still recommend finding the source of the error because you might want to return the faulty hardware while still under warranty. But it seems that it did not affect the end result on this exponent. Happy hunting!

Oh! Thats great news! Thanks!

I had one other double check that mismatched the original LL, and someone later matched me.

When I click into info on my CPU, it shows reliability at 0.30, 1. That seems quite low, wouldn't I want the first number to be close to 1?

Is that a function of my results (which all seem solid, so far!), or of the hardware errors that I occasionally get during PRP/LL double checks?

NB: I decided to try installing the other 2 sticks of RAM to see what my results are. I had one blurry blue screen when I tried to open up gpuOwl immediately after booting, but haven't had any errors or other issues yet. XMP is on, CPU is offset by +.05v.

[mathematizer]

Just following up.
I've been running v303b6 concurrently with GpuOwl 7.2-2 roughly 24/7 for the past week without errors. I had upped the voltage on the CPU by 0.05v, which I think is the step that fixed my Gerbicz/double-check hardware errors.

I had also turned XMP back on, added my other two sticks of ram (so I'm 4x8mb atm).

Aside from the one blurry blue screen when I tried to launch prime95 immediately after my first restart with the new RAM, I haven't had issues.

Appreciate all the help!