| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
| World Community Grid Forums
|
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 13
|
|
| Author |
|
|
RAID 0
Cruncher Joined: Mar 17, 2006 Post Count: 7 Status: Offline |
Anyone here know which one is faster? XP or Ubuntu 7.10 x64? Or does it not make a difference?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
There should be no significant difference.
However, we don't have the opportunity to perform side by side tests, so we will never know for sure. |
||
|
|
RAID 0
Cruncher Joined: Mar 17, 2006 Post Count: 7 Status: Offline |
So it does not matter that Ubuntu is x64 and XP is 32 bit?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
My experience with both has been that Ubuntu 7.10 64bit crunches slightly faster than XP 32bit, or at least it seems to produce a small increase in points anyway using the same hardware.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
There's Ubuntu 32bit and XP 64bit, too. It makes no difference at all; WCG sends 32bit science applications to your computer.
|
||
|
|
RAID 0
Cruncher Joined: Mar 17, 2006 Post Count: 7 Status: Offline |
Thanks for the responses all. :-)
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I have heard that switching from Linux x32 to Linux x64 produces a noticeable speedup, even though the application binary is the same. The only explanation that I can think of is that a fair amount of time must be spent in the support library routines. The 32-bit library routines for Linux x64 ought to be much faster because they can assume a modern CPU with full SSE support in addition to 64-bit operations. This speedup should also show when comparing Linux x64 against Windows x32. On the other hand, the Windows compilers have a reputation for producing better optimizations than the Linux compilers (more development money) so the speedup should be less than that observed when switching from Linux x32 to Linux x64.
This is all speculation on my part. Lawrence |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
My experience with both has been that Ubuntu 7.10 64bit crunches slightly faster than XP 32bit, or at least it seems to produce a small increase in points anyway using the same hardware. Maybe suggest for BOINC to maintain benchmarks at the 32 and 64 bits and compute the appropriate bit level claim depending on the science and hash that to make manual modding the values less easy.... it's Sunday before coffee ![]()
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Morning Sek
- the Boinc benchmark with Ubuntu 64bit does seem to be far higher in comparison to 32bit on the integer side of things, not being knowledgable of these things I have no idea why that is the case.I have heard that switching from Linux x32 to Linux x64 produces a noticeable speedup, even though the application binary is the same. The only explanation that I can think of is that a fair amount of time must be spent in the support library routines. The 32-bit library routines for Linux x64 ought to be much faster because they can assume a modern CPU with full SSE support in addition to 64-bit operations. This speedup should also show when comparing Linux x64 against Windows x32. On the other hand, the Windows compilers have a reputation for producing better optimizations than the Linux compilers (more development money) so the speedup should be less than that observed when switching from Linux x32 to Linux x64. This is all speculation on my part. Lawrence Similar to what I observe Lawrence - those over at XtremeSystems also report noticable speed increases when using XP 64bit too. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
It sure would fix an issue where a 64 bit claim would cause to be so much 'over' that it is considered an outlier and the credit defaulting to the machine that is closer to its grant history. Given that chance will bring a number of quorums together without creating an outlier condition, the average is bound to work in favour of the 64 clients even when running 32 bit science, who on top punch out more jobs per day because of greater system performance from a much more capable OS (Vista is much more efficient in my book in the memory management and throughput department, purely from observation and recording results).
----------------------------------------Now i had my coffee and all cylinders or on operating temp. See the HCC thread for an on-topic finding. ![]()
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Jan 27, 2008 9:34:35 AM] |
||
|
|
|