Sorry it seems the taks are already deleted.
Since only 4 Wus got this error it seems that something went wrong on my side.
So far 2.200 workunits are fine.

The error in 203276984 was caused by a power outage.
So no problem with your app :)

So far no issues with Agbar's optimized application.

My little Intel i5 processes work units within 13 minutes on average.

I got a few errors on my i7 2600K (error 619), but I suppose this is on my own.

Most of the WU are ok (actually 432 WU ok and 8 errors (1.8% WU fail)).

Here is one of the tasks which failed : http://www.enigmaathome.net/result.php?resultid=207621378

Stderr output

<core_client_version>7.6.22</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code -1073741205 (0xc000026b)
</message>
<stderr_txt>
Wrapper v5.26 build 8: starting
01:28:52 (4088): wrapper: running enigma_0.76.exe (-R -o results.txt 00trigr.cur 00bigr.cur 00ciphertext)
Enigma Optima v1.0.0-beta.3 Windows64
Best ISA: AVX
Seed set to: 1469230132.
2016-07-23 01:28:52  enigma: working on range ...

</stderr_txt>
]]>

And here is my computer : http://www.enigmaathome.net/show_host_detail.php?hostid=154232

It may be connected with this issue with BOINC client. I will continue to investigate...

It must have been BOINC client's error:

 > certutil -error 0xc000026b 
0xc000026b (NT: 0xc000026b STATUS_DLL_INIT_FAILED_LOGOFF) -- 3221226091 (-1073741205)
Error message text: {DLL Initialization Failed}
The application failed to initialize because the window station is shutting down. 
CertUtil: -error command completed successfully.

Hello Agbar,

I've noticed that the files have been recently updated to "1.0.0 release", but no executables posted. Is it ready for release, or still some bits to do?

Cheers,

-jj-

I planned to release it after testing. Builds are uploaded and wait on server. So far I tested Win-x64 and Win-x86 on going.

To be honest v1.0.0 does not change anything important for you. All changes are limited to non-performance stuff. I removed 1. most of not used code (not used by E@H) with almost all command line switches and 2. updated copyright screen.

All in all it is same as beta 3. Mentioned changes will release me from supporting some obscure functions of original enigma app. Formally semantic versioning requires to bump major version on removal of any function(-ality) so this is the last moment.

I have no spare time recently and there is a rather minor issue with any version of Windows app (esp. including standard enigma app.) I have prepared v1.1, but I should better talk to TJM first.

PS. I received a message from Alex VE3NEA regarding updated hill climb algorithm. I'm gonna review it within few days. I guess it would make another version, probably as opt-in, enabled by a command line switch.

Given the project is almost done, and the possibility of bugs in third-party applications, should we really be making optimized applications?

---

-Dan Q

danq.co

This project is "almost done" for 3 years, at least. I mean when I started tinkering TJM said it was almost done.
No real progress since then IMO - this is what I would expect from this type of attack on cipher. At some point there will be a "big bang" and message will become broken.

Another point of doing optimized apps is to review the source code. So far I have found one (yes, 1) error that affects this project. It doesn't make results invalid, but cripples real throughput of Enigma@Home.

And it is fun.

This project is "almost done" for 3 years, at least. I mean when I started tinkering TJM said it was almost done.
No real progress since then IMO - this is what I would expect from this type of attack on cipher. At some point there will be a "big bang" and message will become broken.

We haven't been working on the fifth message that long, have we?

---

-Dan Q

danq.co

Thanks a lot for the information Agbar!

Cheers,

-jj-

And it is fun.

Main interest of this this project :D

Finally I released v1.0.0.

If you are using beta3, no need to update as changes include:
- removed most of code unused by this project.
- copy of GPLv2 is now included in downloadable files as it is (probably) required
- copyright info updates in info listings
- help redirected to project on GitHub (replaced Stefan Krah's e-mail, why bother him with my mistakes ;)

There is a bug in AVX2 code path preventing v1.0.0 from running. Please continue using beta3 (on AVX2 hardware, Haswell+) until release of bugfix.

AVX2 is fixed now in v1.0.1

Hell yeah.

Well, my x86 computers are already on the old beta 3, so I don't need to update to 1.01 if I understand correctly.

Hell yeah.

Well, my x86 computers are already on the old beta 3, so I don't need to update to 1.01 if I understand correctly.

You are right.

This still crashes for me, but only on AVX2 hosts:

Stderr output

<core_client_version>7.6.31</core_client_version>
<![CDATA[
<message>
process exited with code 2 (0x2, -254)
</message>
<stderr_txt>
Process creation (../../projects/www.enigmaathome.net/wrapper_5.32_i686-pc-linux-gnu) failed: Error -1, errno=2
execv: No such file or directory

</stderr_txt>
]]>

Actually I'm just dumb and I was lacking 32 bit libs on all the machines it wasn't working on. AVX2 was just a coincidence.

Remember to get all the dependencies on Ubuntu.

I was lacking 32 bit libs on all the machines it wasn't working on. AVX2 was just a coincidence.

Remember to get all the dependencies on Ubuntu.

To be more precise only a wrapper is 32 bit and so it requires 32 bit libs - ie. multilib on ArchLinux.

wrapper_5.32_i686-pc-linux-gnu is a standard file from Enigma@home and in general this project is 32-bit.

I will write a HOWTO on requirements on Linux.