Manual Testing LL result syntax (where to find documentation)

[chalsall]

Originally Posted by preda

Although, it my opinion, GPU are more cost-effective than FPGA for the same computation power.

If I may son, Amazon wouldn't offer them if some couldn't use them effectively.

Sometimes in this game age (read experience) has value.

[Madpoo]

Originally Posted by preda

The result line of my program is exactly the same as CUDALucas save for the program name (because I see no reason to produce a different result). The program right now doesn't use offset, thus offset=0 always in the result.

The program name is "gpuowl", followed by a two-segment version string e.g. "gpuowl v0.1".

M( 77000201 )C, 0x5561a29ad401481c, offset = 0, n = 4096K, gpuowl v0.1, AID: FAF7794A597507B167BE68936175A48E

For a prime, I expect the result should be:
M( 77000000 )P, 0x0000000000000000, offset = 0, n = 4096K, gpuowl v0.1, AID: xxx

The program did check correctly M37,156,667 (to 0 residue, 0 all), and did double-check correctly a few 77M candidates, so it's working at least to some degree.

I've added an entry (at George's suggestion) to the server that lets us store an id # alongside any results coming in.

This has to be an exact match to what the client reports, in this case "gpuowl v0.1" is what I added (and yeah, it tracks different version #'s... each one gets its own id # in a table so we know what specific version of a program turned in the result).

Also... I didn't see an answer, but does the program use shift counts? If not there are just a few adjustments to be made to ensure assignments are handled appropriately (DC auto assignments shouldn't go out to a non-shift-count app if the first run was also done without a shift count). I believe manual assignments will automatically exclude any double-checks if the first one didn't have a shift count, because there's no way to know what program will be doing the check.

[ATH]

Originally Posted by Madpoo

Also... I didn't see an answer, but does the program use shift counts? If not there are just a few adjustments to be made to ensure assignments are handled appropriately (DC auto assignments shouldn't go out to a non-shift-count app if the first run was also done without a shift count). I believe manual assignments will automatically exclude any double-checks if the first one didn't have a shift count, because there's no way to know what program will be doing the check.

Originally Posted by preda

The result line of my program is exactly the same as CUDALucas save for the program name (because I see no reason to produce a different result). The program right now doesn't use offset, thus offset=0 always in the result.

It does not use shift counts "right now", so maybe an upcoming feature?.

[preda]

gpuOwl doesn't use shiftcount. It might be possible to add it in the future, though I don't have a plan for that right now (I need to finish a few other things before, to get this program off the ground).

[Madpoo]

Originally Posted by preda

gpuOwl doesn't use shiftcount. It might be possible to add it in the future, though I don't have a plan for that right now (I need to finish a few other things before, to get this program off the ground).

Cool. Well, if/when you do add that feature, be sure to change the version # reported by the client so it can be handled differently by the server in that regard.