When all remaining problems are ironed out, the Windows (and other) apps will be updated to feature SSE(and Altivec??) vector arithmetic that is now already speeding up the Mac OS/Intel app. That should help to speed up things quite a bit, so finishing S5R3 in 2008 should be feasible.
When all remaining problems are ironed out, the Windows (and other) apps will be updated to feature SSE(and Altivec??) vector arithmetic that is now already speeding up the Mac OS/Intel app. That should help to speed up things quite a bit, so finishing S5R3 in 2008 should be feasible.
Bikeman
That is great news! Of course, this means that faster science work leads to quicker science reports. By the way, will there be any time lag between the completion of S5R3 and the start of S6 science run?
When all remaining problems are ironed out, the Windows (and other) apps will be updated to feature SSE(and Altivec??) vector arithmetic that is now already speeding up the Mac OS/Intel app. That should help to speed up things quite a bit, so finishing S5R3 in 2008 should be feasible.
Bikeman
That is great news! Of course, this means that faster science work leads to quicker science reports. By the way, will there be any time lag between the completion of S5R3 and the start of S6 science run?
Hi.
The S6 run can only begin when the "Enhanced LIGO" upgrade is ready for a science run, which is anticipated for 2009.
As the current S5R3 experiment is using the same raw data that was used for S5R2 (which was begun long before the S5 run had ended), it's obvious that even S5R3 is only looking at part of the S5 data. So there's more than enough S5 work left to be done for E@H before S6 data even becomes available.
When all remaining problems are ironed out, the Windows (and other) apps will be updated to feature SSE(and Altivec??) vector arithmetic that is now already speeding up the Mac OS/Intel app. That should help to speed up things quite a bit, so finishing S5R3 in 2008 should be feasible.
Bikeman
Any chance that AKOSF will share his secrets about how he got times below 10k sec with the E@H team? I wouldn't think he could have done it all by overclocking. I assume he had to make some code optimizations. I could be wrong an all accounts.
How many participants?
)
http://einstein.phys.uwm.edu/server_status.php
Under Computers. Divided by Users and Hosts.
RE: http://einstein.phys.uw
)
As an aside, 488 days???
Do they really need that much time to get prepped for S5R4 / S6? ;)
RE: RE: http://einstein.p
)
Where did you see that?
Physics is for gurls!
In the Work still remaining
)
In the Work still remaining column of the S5R3 search progress. Although it's now down to 486.1 days.
S5R3 though? ;-)
When all remaining problems
)
When all remaining problems are ironed out, the Windows (and other) apps will be updated to feature SSE(and Altivec??) vector arithmetic that is now already speeding up the Mac OS/Intel app. That should help to speed up things quite a bit, so finishing S5R3 in 2008 should be feasible.
Bikeman
RE: When all remaining
)
That is great news! Of course, this means that faster science work leads to quicker science reports. By the way, will there be any time lag between the completion of S5R3 and the start of S6 science run?
RE: RE: When all
)
Hi.
The S6 run can only begin when the "Enhanced LIGO" upgrade is ready for a science run, which is anticipated for 2009.
As the current S5R3 experiment is using the same raw data that was used for S5R2 (which was begun long before the S5 run had ended), it's obvious that even S5R3 is only looking at part of the S5 data. So there's more than enough S5 work left to be done for E@H before S6 data even becomes available.
CU
Bikeman
RE: When all remaining
)
Any chance that AKOSF will share his secrets about how he got times below 10k sec with the E@H team? I wouldn't think he could have done it all by overclocking. I assume he had to make some code optimizations. I could be wrong an all accounts.