Pages: [1]
Francis
BAM!ID: 2333
Joined: 2006-06-18
Posts: 5
Credits: 0
World-rank: 0

2006-06-18 09:45:51

Hi, I had a PC standing idle and loaded it with Windows Vista (Ult.) beta 2. While everything seems to work just fine, I would have liked to compare my setup against other boxes, with same or different OS'ses. Am I alone in this / can it be done?

Perhaps allowing a number of criteria, to form a custom search or so?

Checking stats for a computer shows processor and OS, but not the memory available...

william
BAM!ID: 633
Joined: 2006-05-22
Posts: 79
Credits: 77,326
World-rank: 737,274

2006-06-18 21:31:56

You would have to take into account the front end bus, The G card, The Ram, The V ram, The disk spaces, Disk speed, and A very long list of variables. It would be Easier to bench test your computer with two diffrent OS on it.
Honza
BAM!ID: 109
Joined: 2006-05-10
Posts: 154
Credits: 3,493,742,539
World-rank: 555

2006-06-19 07:04:57

There would not be much difference between M$ OSes since BOINC apps are not 64-bits. When same HW, difference might be more significant using FAT32 vs. NTFS, drivers used and available memory (swapping) which may be more intensive in Vista.

There were tests aimed on impact of main memory - clock speed, cache latencies, DDR vs. DDR2, FSB:CPU and such.

One can't take all into account.
I would suggest that you pick up one aspect and measure (e.g. FAT32 vs. NTFS, i-RAM vs. software RAM drive).
Not sure how you deal with different BOINC projects. CPDN stress whole system I guess the most among all projects (CPU, RAM, HD), projects with a primitive code stress mainly CPU and it's cache (SZTAKI), some are somewhere between (Einstein?).
Francis
BAM!ID: 2333
Joined: 2006-06-18
Posts: 5
Credits: 0
World-rank: 0

2006-06-19 18:54:52

There would not be much difference between M$ OSes since BOINC apps are not 64-bits. When same HW, difference might be more significant using FAT32 vs. NTFS, drivers used and available memory (swapping) which may be more intensive in Vista.

Agreed... cannot remember a choice for NTFS, but HDD was already formatted NTFS. I do *not* see much swapping; HDD use by BOINC is not even 20 MB out of 100+ GB.


One can't take all into account.
I would suggest that you pick up one aspect and measure (e.g. FAT32 vs. NTFS, i-RAM vs. software RAM drive).
Not sure how you deal with different BOINC projects. CPDN stress whole system I guess the most among all projects (CPU, RAM, HD), projects with a primitive code stress mainly CPU and it's cache (SZTAKI), some are somewhere between (Einstein?).



OK, one stat: % of time running BOINC = 100%, time work is allowed = 99,6298%, average CPU efficiency = 0,98373

I was curious how that compares to (say) WinXP, but I may need to reinstall WinXP & BOINC if I really really really want to know! (Idon't... not that much :-)!)
Thanks for your input!
Francis
Honza
BAM!ID: 109
Joined: 2006-05-10
Posts: 154
Credits: 3,493,742,539
World-rank: 555

2006-06-20 07:08:43

CPU effeciency will go lower when other tasks running. It may be antivirus and other background task or (in my case) spooler doing large jobs everyday when making full-res PDF files, perhaps music and video playback or (not my case) playing games.
So, it has nothing much to do with OS but with other tasks running.
Francis
BAM!ID: 2333
Joined: 2006-06-18
Posts: 5
Credits: 0
World-rank: 0

2006-06-21 18:36:46

CPU effeciency will go lower when other tasks running. It may be antivirus and other background task or (in my case) spooler doing large jobs everyday when making full-res PDF files, perhaps music and video playback or (not my case) playing games.
So, it has nothing much to do with OS but with other tasks running.



:-) Basically you're saying BOINC is not the way to test then ;-)
Note: I unloaded and uninstalled McAfee (isn't working properly under Vista anyhow!) but there's no diff.

There are no other tasks running, not even screensaver and only Vista is installed.

Thanks for the comments,
Francis
Francis
BAM!ID: 2333
Joined: 2006-06-18
Posts: 5
Credits: 0
World-rank: 0

2006-07-27 15:58:21

CPU effeciency will go lower when other tasks running. It may be antivirus and other background task or (in my case) spooler doing large jobs everyday when making full-res PDF files, perhaps music and video playback or (not my case) playing games.
So, it has nothing much to do with OS but with other tasks running.


A bit of a surprise; I tested BOINC on my primary box as well, a Dell GX620. P iv - 3,4 GHz, 2 GByte, 250 GB Hdd. (Win XP Pro)
The good news: with hyperthreading ON it performs significantly better, BUT... with hyperthreading OFF, it does NO better than the "longhorn" box (P iv - 2,8 GHz, 1 GByte, 110 GB HDD). Go figure...
Pages: [1]

Index :: Comments and suggestions :: Stats & testing OS?
Reason: