New WUs splitted. When to expect app's binaries update?

It took some searching, but I've found one WU completed by one Windows app, and one Linux app. It's validated at the first attempt, without needing the third copy. Some of us have been a little worried by marginally different results between the two platforms, but it's clearly good enough for the validator.

WU 7572084

(though in this particular case, the Linux app chose avx_ChirpData_b, rather than the troublesome v_ChirpData)

WU 7572083 shows the same immediate validation between Windows and Linux, this time with the Linux host choosing v_ChirpData.

I'm seeing a vastly underestimated run time for the new batch of tasks.

CreditNew at its best.

might get better when the slow rigs start reporting in in bigger numbers. Also for an individual host, once you are past the 11 task APR use threshold

---

A person who won't read has no advantage over one who can't read. (Mark Twain)

Any idea when we might get GPU apps to test? Not pushing, just curious :-)

Any idea when we might get GPU apps to test? Not pushing, just curious :-)

Trying to get a definite timeline out of a coder has about the same chance of success as getting a politician to give a straightforward statement he can be nailed down to later on.

That said, a preliminary ATI app is already in pre-alpha and Jason is waiting for the splitter tweaks to finish before doing 'trivial app tweaks' to CUDA - you don;t want to start before they've finished laying the grounds.

Any apps go through alpha testing and anonymous platform beta testing before they get recommended to Eric for public (beta) release.

iow it could be a while. but since we are all volunteers and free time is most abundant at weekends and holidays I'd go out on a branch and suspect something in testing by mid January.
[and now invoke Murphy please]

---

A person who won't read has no advantage over one who can't read. (Mark Twain)

Just spotted an inconclusive - my Windows against Claggy's Linux. WU 7572111.

Just spotted an inconclusive - my Windows against Claggy's Linux. WU 7572111.

To spike or not to spike that is the question.

Sorry couldn't resist : D

3rd wing is another windows.

---

A person who won't read has no advantage over one who can't read. (Mark Twain)

Just spotted an inconclusive - my Windows against Claggy's Linux. WU 7572111.

and vChirp involved...
Hope Eric releases new binaries today.

---

News about SETI opt app releases: https://twitter.com/Raistmer

Building now....

---

Building now....

Good. How long we need to wait?

Building now....

Good. How long we need to wait?

If the only available tool at moment of building is screwdriver... pretty long :D
(joke)

Actually I think it's not "buiding" time-limiting but "signing". AFAIK it should be done on offline computer and not always goes as expected.

---

News about SETI opt app releases: https://twitter.com/Raistmer

Building now....

Good. How long we need to wait?

If the only available tool at moment of building is screwdriver... pretty long :D
(joke)

Actually I think it's not "buiding" time-limiting but "signing". AFAIK it should be done on offline computer and not always goes as expected.

And - since the final build fixes were committed well after normal lab closing time - I'm guessing Eric left it running an overnight test, without risking an unattended deployment.

Building now....

Good. How long we need to wait?

If the only available tool at moment of building is screwdriver... pretty long :D
(joke)

Actually I think it's not "buiding" time-limiting but "signing". AFAIK it should be done on offline computer and not always goes as expected.

And - since the final build fixes were committed well after normal lab closing time - I'm guessing Eric left it running an overnight test, without risking an unattended deployment.

Seti v8 8.00 apps have been deployed, but there's no Seti v8 work available yet to test the apps.

Claggy

Something is wrong with the latest splitter, and, of course, Jeff is out of the office this afternoon. Work will arrive when it arrives...

---

Building now....

Good. How long we need to wait?

New x86_64-linux and i686-windows builds are out. i686-linux should arrive shortly. x86_64-windows has been giving me fits. Every 64-bit build runs fine on wine64 and XP, but gives a 0xc0000022 exception on every windows version newer than that. Probably some DLL load is failing, but neither the error message or debugger give a clue as to which DLL is the problem.

---

but gives a 0xc0000022 exception on every windows version newer than that. Probably some DLL load is failing, but neither the error message or debugger give a clue as to which DLL is the problem.

0xc0000022 - это STATUS_ACCESS_DENIED: попытка открыть какой-то объект (файл, ключ реестра, объект синхронизации и т.д.), не имея на это нужных разрешений

But no idea what it could be in case of console app...

---

News about SETI opt app releases: https://twitter.com/Raistmer

New x86_64-linux version uses fftw-lib v3.2.1 which does not know about sse2 or avx. This will cost performance again.
Additionally the published version states to be built from svn revision 3232, but you did more changes to the 64bit codepath after that revision. Is 64bit linux not affected ?

setigraphics looks ok (on my distro at least) now.

---

_\|/_
U r s

Just spotted an inconclusive - my Windows against Claggy's Linux. WU 7572111.

To spike or not to spike that is the question.

The spikes have it, but all three are valid.

I'll try to get 3.3.4 installed for the next release.

The revision listed in the code is from the last time configure was run. I need to add a method to change it whenever an svn update has occurred.

New x86_64-linux version uses fftw-lib v3.2.1 which does not know about sse2 or avx. This will cost performance again.
Additionally the published version states to be built from svn revision 3232, but you did more changes to the 64bit codepath after that revision. Is 64bit linux not affected ?

setigraphics looks ok (on my distro at least) now.

---

Something is wrong with the latest splitter, and, of course, Jeff is out of the office this afternoon. Work will arrive when it arrives...

Looks like Jeff has got the splitter running. Work has arrived.

BOINC client failed to download s@h v8 client again! s@h v8 is now 2118kb!
Workunits are fine. My BOINC is version 7.6.9. What's the problem! Windows 10?

---

Luckiest in the world. WMD = Weapon of Mass Distraction