htmlmasterdave
Golden Member
What is the fastest possible OS for seti? I have heard that win32 is faster, but which version of windows? Anyone know? Just curious 😉
-
We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.
Fastest OS for Seti
- Thread starter htmlmasterdave
- Start date
Windows NT/2K/XP will probably work out the fastest as they are not affected by VLAR WUs. Linux is also not affected but the Linux client is about 5% slower. Running the Windows client in WINE is as fast but does not give an advantage over running it in Windows.
This is only what I've heard feel free to correct me if I'm wrong.
This is only what I've heard feel free to correct me if I'm wrong.
htmlmasterdave
Golden Member
Ok, good, I don't run 9x much at all anymore anyways 😀
RaySun2Be
Lifer
Baldy,
If I remember correctly, Poof ran some benchmark tests, and she found that running the Windows client in WINE in Linux was actually slightly faster than even the straight Windows client. Which still causes me to shake my head in amazement :Q
But yea, Win2K seems to be good overall, especially when dealing with VLARs, or "icky" WUs as some people have recently called them. 😉😉
I seem to take but an hour or 1.5 hours hit when I get a VLAR in Win2KPro. 🙂
If I remember correctly, Poof ran some benchmark tests, and she found that running the Windows client in WINE in Linux was actually slightly faster than even the straight Windows client. Which still causes me to shake my head in amazement :Q
But yea, Win2K seems to be good overall, especially when dealing with VLARs, or "icky" WUs as some people have recently called them. 😉😉
I seem to take but an hour or 1.5 hours hit when I get a VLAR in Win2KPro. 🙂
A bunch of us had run a number of benches with the 3.x SETI clients. It all sortof started in <a target=new class=ftalternatingbarlinklarge href="http://arstechnica.infopop.net/OpenTopic/page?a=tpc=50009562=122097561=841095953
">this thread</a> and progressed through a number of additional threads when 3.03 came out. Basically the science in the 3.03 client (double of drift rate) doubled the times from the 3.00 client. We also discovered the VLAR problem with the win CLI. It does manifest itself on all win OSes but to a lesser degree with NT/2K/XP Pro. With all other OSes, VLAR WUs run faster than mid-angle ranges.
I found that running the CLI with WINE gave me about 5 or so minutes faster processing time than running the CLI natively. 😀
[EDIT: Once again, linking from here is borked. :| But here is a quote of the 3.00 results I reported in that thread (bench WU run on the same machine - dual boot P3 600@800):
"So you have, using the SAME hardware, ie., the SAME machine and the WINE & 98SE using the SAME 3.0 CLI:
WINE = 3:42:11 hrs
98SE = 3:46:13 hrs
Linux = 3:55:24 hrs"
">this thread</a> and progressed through a number of additional threads when 3.03 came out. Basically the science in the 3.03 client (double of drift rate) doubled the times from the 3.00 client. We also discovered the VLAR problem with the win CLI. It does manifest itself on all win OSes but to a lesser degree with NT/2K/XP Pro. With all other OSes, VLAR WUs run faster than mid-angle ranges.
I found that running the CLI with WINE gave me about 5 or so minutes faster processing time than running the CLI natively. 😀
[EDIT: Once again, linking from here is borked. :| But here is a quote of the 3.00 results I reported in that thread (bench WU run on the same machine - dual boot P3 600@800):
"So you have, using the SAME hardware, ie., the SAME machine and the WINE & 98SE using the SAME 3.0 CLI:
WINE = 3:42:11 hrs
98SE = 3:46:13 hrs
Linux = 3:55:24 hrs"
I used to have 2 processes of SETI on an old Compaq Proliant dual Ppro 200 (512K cache) server running Red Hat 6.1 (back in the 2.x days). If you can get a pair of PPros that have at least 1MB cache, then you should do pretty good with the times on both CPUs, when compared to other dual systems of the same speed. I currently have a dual Xeon 1MB cache system (the PPros being the predecessors of the Xeons) and have had 2 SETI processes running on it (one per CPU), with no loss of time compared to running singly. Dualies with less than 1MB cache will experience bus/RAM contention (ie., the SETI dataset is ~800K in size and would have to swap out of the cache and into the RAM), thus lengthening the time to do both WUs in dual configuration.
[EDIT: Let me add that you may not see a speedup when running dual, but you should NOT see an increase in time. 🙂]
[EDIT: Let me add that you may not see a speedup when running dual, but you should NOT see an increase in time. 🙂]
even if I find it that will not increase the processing time (if I understood correctly).
To be more explicit, it won't necessarily run the WU faster in dual mode than if run singly (although interestingly, due to the microcode in the Xeons and the PPros, WUs in general run faster than on say a regular P2 or P3), but it WON'T run slower as a dual, which happens with say dual P3s. Ie., running SETI on dual P3s that are not Xeons (or even a dual Pally) will give you results that take longer than if run on only one CPU of that kind.
To be more explicit, it won't necessarily run the WU faster in dual mode than if run singly (although interestingly, due to the microcode in the Xeons and the PPros, WUs in general run faster than on say a regular P2 or P3), but it WON'T run slower as a dual, which happens with say dual P3s. Ie., running SETI on dual P3s that are not Xeons (or even a dual Pally) will give you results that take longer than if run on only one CPU of that kind.
I was looking at your farm, very interesting herd you have there. I just come back from a little vacation that I took deep in the woods in North Carolina and I took all of my magazines that I need to read (no time, alot of work) and 2 of them have articles about Linux, I'm thinking to give a try in that Compaq PPro and learn from it, for what I read there are 2 main contenders (Mandrake and Suse), so I will do more research and see wich one to go for.
Eltano
Eltano
Cool!
I need to update my farm page since I have upgraded some systems - plus I think you can only list 10 machines so my other 4 don't show up.
I have a Mandrake machine here (an older distro - 6.5) but use SuSE or Red Hat primarily. The only thing with SuSE is that they don't offer .iso images over the net like most of the other commercial distros, but they will let you install it over the net using FTP (which is how I did my latest installs). I've also purchased the boxed set which includes the books, CDs (or DVD), and 30 days of phone support.
Good luck! 🙂
I need to update my farm page since I have upgraded some systems - plus I think you can only list 10 machines so my other 4 don't show up.
I have a Mandrake machine here (an older distro - 6.5) but use SuSE or Red Hat primarily. The only thing with SuSE is that they don't offer .iso images over the net like most of the other commercial distros, but they will let you install it over the net using FTP (which is how I did my latest installs). I've also purchased the boxed set which includes the books, CDs (or DVD), and 30 days of phone support.
Good luck! 🙂
My SetiQueue
If you look through the history a bit, you can see WU times for Windows CLI on Windows 2000, Windows CLI on Windows 98 SE, Windows CLI on WINE, and native Linux.. This might provide the statistical info you need. 🙂
EDIT: Never mind. I just remembered that I moved the setiqueue to my only remaining Windows box, which happens to be behind the firewall. 😱
If you look through the history a bit, you can see WU times for Windows CLI on Windows 2000, Windows CLI on Windows 98 SE, Windows CLI on WINE, and native Linux.. This might provide the statistical info you need. 🙂
EDIT: Never mind. I just remembered that I moved the setiqueue to my only remaining Windows box, which happens to be behind the firewall. 😱
This might provide the statistical info you need. 🙂
But in order to get the true goods, you would need to run the same WU on the same machine as a multiboot to account for effects from the hardware itself (eg., CPU temp, etc). This is what I did.
However, a number of us also confirmed that on dedicated boxen, there is a less than 1% difference between times on the same machine, running different WUs, but all with the same angle range.
But in order to get the true goods, you would need to run the same WU on the same machine as a multiboot to account for effects from the hardware itself (eg., CPU temp, etc). This is what I did.
However, a number of us also confirmed that on dedicated boxen, there is a less than 1% difference between times on the same machine, running different WUs, but all with the same angle range.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 25K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 24K
-
-
