Correct, no SSE code, so there's more to come in the future. The performance of the recent Linux power App should be a rough indication what to expect on the Windows platform as well.
Bikeman
There are two additional wonderful benefits to this, which are saving weeks of processing time to complete this phase of the project, and reducing lots of greenhouse gas emissions with more efficient processing! :)
You perhaps have heard the management jargon term "work smarter, not harder". Not all code speedup tweaks just get more of the hardware working at once. Some find ways to avoid doing some of the computing. I had the impression that the linear trig function trick is meant to do just that.
Still, if temp drops from one version to another, that's an indication there might be a few transistors in the CPU that are just bored most of the time and could perhaps be forced to help in the calculation, so to speak.
EDIT: 4.25 to 4.26 comparision: about 25% less time needed. Not bad... Ca 10 % improvement over 4.15, IIRC the runtime of 4.15 on my system.
You perhaps have heard the management jargon term "work smarter, not harder". Not all code speedup tweaks just get more of the hardware working at once. Some find ways to avoid doing some of the computing. I had the impression that the linear trig function trick is meant to do just that.
Still, if temp drops from one version to another, that's an indication there might be a few transistors in the CPU that are just bored most of the time and could perhaps be forced to help in the calculation, so to speak.
CU
Bikeman
Now now... We can't have you EU types causing our fair country over here to emit more gas during the search for gravity waves... :-P
Anyway, my core is still as toasty as ever, with temp hovering around 57-58. This is due to current ambient being lower than some like it, at only 17... :-)
Edit: Note to non-Metric folks, that was in Celsius.... I really don't have this room at 17F... LOL
I just read on the poweruserpage: "looks a bit faster, but may not run correctly on Windows Vista".
Does this mean, there could be problems at validation? Or something else? Here at Vista it runs fine up to now.
Looks like 4.26 will cut about 8 hours off the run time on the first WU. It's a short one and they were taking around 40 hrs. It's running a little over 3% per hour. The second one is doing 2.5% per hour, which is what I would expect as it is a longer WU. It looks like it will take roughly 48 hrs to complete. With the old app they were taking 55-60 hrs.
As soon as FAH is done with the 8 way I'll load 4.26 on it and see how it works there.
I just read on the poweruserpage: "looks a bit faster, but may not run correctly on Windows Vista".
Does this mean, there could be problems at validation? Or something else? Here at Vista it runs fine up to now.
There was some concern about the app crashing... If it is running, I think you're ok...
It is reading the app_info.xml file but it is still running 4.15!?!
Help.
Thanks
It's bluffing :-). It's really running the new app, but as the app_info.xml tells BOINC that 4.26 should continue all tasks intended for 4.15, it acts as an alias for 4.15 (sort of).
You can verify this by looking into the stderr.txt file that gets written to one of the slots/* directories. It will contain a line about 4.26 starting.
1/22/2008 4:33:27 PM|Einstein@Home|Restarting task h1_0819.05_S5R2__382_S5R3a_0 using einstein_S5R3 version 415
It is reading the app_info.xml file but it is still running 4.15!?!
Go see what task is actually running. (Ctrl-alt-del will get you Task Manager, click on the processes tab). I think you will find you are running 4.26.
The message you quote uses a version that is determined at download time, and entered into the .xml. If you leave 4.26 installed, that info will be 426 for your future downloads.
[edit: I see Bikeman posted while I was typing--but I'll leave this as there is some non-overlap]
RE: Correct, no SSE code,
)
There are two additional wonderful benefits to this, which are saving weeks of processing time to complete this phase of the project, and reducing lots of greenhouse gas emissions with more efficient processing! :)
RE: You perhaps have heard
)
Still, if temp drops from one version to another, that's an indication there might be a few transistors in the CPU that are just bored most of the time and could perhaps be forced to help in the calculation, so to speak.
EDIT: 4.25 to 4.26 comparision: about 25% less time needed. Not bad... Ca 10 % improvement over 4.15, IIRC the runtime of 4.15 on my system.
CU
Bikeman
RE: RE: You perhaps have
)
Now now... We can't have you EU types causing our fair country over here to emit more gas during the search for gravity waves... :-P
Anyway, my core is still as toasty as ever, with temp hovering around 57-58. This is due to current ambient being lower than some like it, at only 17... :-)
Edit: Note to non-Metric folks, that was in Celsius.... I really don't have this room at 17F... LOL
I just read on the
)
I just read on the poweruserpage: "looks a bit faster, but may not run correctly on Windows Vista".
Does this mean, there could be problems at validation? Or something else? Here at Vista it runs fine up to now.
Looks like 4.26 will cut
)
Looks like 4.26 will cut about 8 hours off the run time on the first WU. It's a short one and they were taking around 40 hrs. It's running a little over 3% per hour. The second one is doing 2.5% per hour, which is what I would expect as it is a longer WU. It looks like it will take roughly 48 hrs to complete. With the old app they were taking 55-60 hrs.
As soon as FAH is done with the 8 way I'll load 4.26 on it and see how it works there.
RE: I just read on the
)
There was some concern about the app crashing... If it is running, I think you're ok...
RE: There was some concern
)
Ah, thanks. Surely the problem with the app_info file. Wich is solved. ;)
Hi, I tried installing
)
Hi,
I tried installing 4.26 -
unzipped and placed the files into the einstein directory but
messages say that I'm still running 4.15. Here are the messages on startup:
1/22/2008 4:33:27 PM||Starting BOINC client version 5.8.11 for windows_intelx86
1/22/2008 4:33:27 PM||log flags: task, file_xfer, sched_ops
1/22/2008 4:33:27 PM||Libraries: libcurl/7.16.0 OpenSSL/0.9.8a zlib/1.2.3
1/22/2008 4:33:27 PM||Data directory: C:\\PROGRA~1\\BOINC
1/22/2008 4:33:27 PM|Einstein@Home|Found app_info.xml; using anonymous platform
1/22/2008 4:33:27 PM||Processor: 2 AuthenticAMD AMD Athlon(tm) 64 X2 Dual Core Processor 3800+ [x86 Family 15 Model 75 Stepping 2] [fpu tsc pae nx sse sse2 3dnow mmx]
1/22/2008 4:33:27 PM||Memory: 990.48 MB physical, 2.33 GB virtual
1/22/2008 4:33:27 PM||Disk: 232.88 GB total, 199.41 GB free
1/22/2008 4:33:27 PM|Einstein@Home|URL: http://einstein.phys.uwm.edu/; Computer ID: 528442; location: home; project prefs: default
1/22/2008 4:33:27 PM||General prefs: from Einstein@Home (last modified 2008-01-15 18:07:24)
1/22/2008 4:33:27 PM||Host location: home
1/22/2008 4:33:27 PM||General prefs: using separate prefs for home
1/22/2008 4:33:27 PM|Einstein@Home|Restarting task h1_0819.05_S5R2__382_S5R3a_0 using einstein_S5R3 version 415
1/22/2008 4:33:27 PM|Einstein@Home|Restarting task h1_0819.05_S5R2__381_S5R3a_1 using einstein_S5R3 version 415
It is reading the app_info.xml file but it is still running 4.15!?!
Help.
Thanks
Joe B
RE: It is reading the
)
It's bluffing :-). It's really running the new app, but as the app_info.xml tells BOINC that 4.26 should continue all tasks intended for 4.15, it acts as an alias for 4.15 (sort of).
You can verify this by looking into the stderr.txt file that gets written to one of the slots/* directories. It will contain a line about 4.26 starting.
CU
Bikeman
RE: 1/22/2008 4:33:27
)
Go see what task is actually running. (Ctrl-alt-del will get you Task Manager, click on the processes tab). I think you will find you are running 4.26.
The message you quote uses a version that is determined at download time, and entered into the .xml. If you leave 4.26 installed, that info will be 426 for your future downloads.
[edit: I see Bikeman posted while I was typing--but I'll leave this as there is some non-overlap]