Windows S5R4 App 6.06 available for Beta Test
27 Nov 2008 14:50:35 UTC
Topic 13719
(moderation:
[Edit] It looks like this App has a serious bug in the checkpointing (see below) Only use it if you are sure of what you are doing, e.g. keep the App in memory all the time through a task. [/Edit]
A new Windows App is available from our Beta Test Page.
As the preparation for S5R5 is taking a bit longer than expected I decided to produce a full-featured S5R4 App with the build process used for the 6.05 "Power" App for you to test. All three executables in thie App package are built using the new gcc build process, which should allow for direct speed comparison with their x87 and SSE counterparts of the standard 6.04 App.
In addition the "crash after done" bug reported e.g. here should be fixed in this App.
The zip archive contains an app_info.xml that should allow to switch tasks in progress by 6.04 and 6.05 App, however this is also subject to the Beta test.
Please test and report.
BM
BM
Windows S5R4 App 6.06 available for Beta Test
)
Thanks for the note! Fixed.
BM
BM
Thanks for the report. If
)
Thanks for the report.
If the file is not renamed it means that an old checkpoint is not overwritten by a new one, i.e. the checkpointing is actually completely broken.
There will probably soon be another release.
BM
BM
There shouldn't be any
)
There shouldn't be any difference in speed comparing to the 6.05 if the 6.05 ran on your computer.
I'd be curious to know how the 6.04_0 compares to the 6.06_0 and the 6.04_1 to 6.06_1 (does s.o. get what I mean by this?).
BM
BM
The checkpointing problem
)
The checkpointing problem apparently arises from compiling the BOINC library for Windows with gcc. I've seen these "couldn't rename" messages occasionally with 6.05, but apparently the change I made to fix this in 6.06 made things rather worse than better. I'll continue to work on this.
I think for most people that have SSE2 machines for the time being it would be best to step back to 6.05, I hope to get a 6.07 out over the weekend, but am not sure.
I'd still be curious about performance comparisons between equally featured Apps of the 6.04 vs. 6.06 packages, e.g. if someone could test 6.06 on an old Athlon that is not capable of SSE2.
I actually wonder if there are still non-SSE machines attached to Einstein@home.
BM
BM
I issued a SSE2 part of a new
)
I issued a SSE2 part of a new App 6.07 that is intended to fix the checkpointing problem. However my possibilities for testing are very limited during the weeked, so I rely more or less on you for that. Please take a look and report.
BM
BM