x86 F@H client on XP Professional x64 Edition

Lancelot

Member

​

Joined

Feb 12, 2001

Location

the Netherlands

• Jan 22, 2006

• #1

I have made a dual boot with regular XP and XP x64 Edition. (fortunatly all my hardware has 64-bit driver support) x64 Edition runs noticeably faster but my folding clients are probably running on WoW64 (Windows32 on Windows64) What would be best to leave running overnight, regular XP or x64?!

 

V

veryhumid

Member

​

Joined

Apr 22, 2004

Location

New Hampshire

• Jan 22, 2006

• #2

i am not totally sure, but i think FAH is a 32bit application. I bet it wouldn't make a difference.

 

OP

Lancelot

Member

​

Joined

Feb 12, 2001

Location

the Netherlands

• Jan 22, 2006

• Thread Starter
• #3

That's why I'm asking this; F@H is a 32bit app but XP x64 sort of 'emulates' 32-bit. One would think 'direct' 32bit would be slightly faster...

 

TollhouseFrank

Senior Headphone Guru

​

Joined

Nov 29, 2004

Location

T3h Intr@tub3z!

• Jan 22, 2006

• #4

well... the only way to find out is to do a long-term test.

 

W

WillS

Registered

​

Joined

Oct 29, 2003

Location

Lawton, OK

• Jan 22, 2006

• #5

I haven't noticed any speed difference with FAH on x64 (in terms of PPD).

 

Misfit138

Member

​

Joined

Mar 23, 2005

Location

Jersey

• Jan 22, 2006

• #6

i was running x64 for about a week when my Raptor blew up.
(I ordered 2 deskstars for a SATA II RAID 0 as a replacement.)
I didnt notice any difference in folding. seemed about the same.

 

pscout

Senior Newbie Cannon-Fodder - R. I. P. good budd

​

Joined

Nov 22, 2004

Location

Toronto

• Jan 22, 2006

• #7

you could find out for sure easily enough ... just make a copy of the queue.dat and work subdirectory. Run xp or xp64 for a few (10?) frames. switch to the other os, delete the work dir and queue.dat, copy back the WU (work dir and queue.dat) you saved at the beginning. Let it run till you have at least the same number of frames as the first test.

Just make sure you don't do anything or have anything else running on the rig during either test and you should have a pretty exact comparison of times from the log.

Post your results ... it will benefit everyone using xp 64 to know.

 

Captain Helghas

Member

​

Joined

Nov 22, 2005

Location

Tallahassee, Fl

• Jan 23, 2006

• #8

Sorry to threadjack, but I think this also applies. Does anyone know if there is a difference in folding efficiency between one version of windows vs. another? Other than just 32 and 64 bit support. I'm mainly curious if 95/98/2000/ME/ or XP show any true difference.

 

Adak

Senior Member

​

Joined

Jan 9, 2006

• Jan 23, 2006

• #9

The biggest difference in the versions of Windows, would be in the stability of the system.

I would not recommend 95, 98 (although Second Edition of this wasn't too bad).
and DEFINITELY would not recommend ME, which wants to "do housekeeping" when I try to do serious and intensive work on it. Extremely frustrating.

Win2000 and XP are the two stars in this mix, certainly, and XP inherited many of 2000's good innards, just as 2000 inherited the best of WinNT.

I would bet money that the F@H software has been extensively tested on XP, and possibly 2000 and 98SE. I doubt if it's been tested on 95 or ME.

Adak

What? No Windows 3.0 or 3.1?

 

AlabamaCajun

Member

​

Joined

Apr 22, 2005

Location

A Labyrinth

• Jan 23, 2006

• #10

For folding, the OS should have little effect. The reason is the CPU is not changing it mode of operation just the OS. There may be some very small differce in the IO but it should be negligable. FaH runs it's own native code and spends 99.99& of it's time there. Only the GUI while open will have an impact.

 

Captain Helghas

Member

​

Joined

Nov 22, 2005

Location

Tallahassee, Fl

• Jan 23, 2006

• #11

I actually debated to ask about 3.0 & 3.1 as I recently found copies of 3.1. However, I don't think I can get a client to run on it, and even if I could 3.1 doesn't recognize past 64Mb of ram, so you couldn't fold well anyways.

 

pscout

Senior Newbie Cannon-Fodder - R. I. P. good budd

​

Joined

Nov 22, 2004

Location

Toronto

• Jan 23, 2006

• #12

AlabamaCajun said:

For folding, the OS should have little effect. The reason is the CPU is not changing it mode of operation just the OS. There may be some very small differce in the IO but it should be negligable. FaH runs it's own native code and spends 99.99& of it's time there. Only the GUI while open will have an impact.

Click to expand...

So if there is much of a difference, it would likely be due to contention with other background tasks the os includes ... which could of course vary widely depending on install defaults in the os and settings, and other stuff you have installed.

On my xp rigs where i look at task manager much, i don't see much else running and seldom does the core drop below 50% utilization (p4 HT).

 

dwschoon

Member

​

Joined

May 24, 2005

Location

Oklahoma City, OK

• Jan 23, 2006

• #13

Ive heard that windows 2000 is probably the best OS for folding only. Anything that predates that cant run bigpackets because it calls for the v4 client. You can run the v5 client on 98, but if you turn bigpackets on, some WU's will generate errors and not fold. Windows 2000 folds faster for me, especially using older hardware, because it has less services running in the background, plus it defaults to the classic interface.

 

pscout

Senior Newbie Cannon-Fodder - R. I. P. good budd

​

Joined

Nov 22, 2004

Location

Toronto

• Jan 23, 2006

• #14

yes the classic interface ... thats the first thing i do on an xp install is set the desktop and IE to classic! ... so that i can find everything and the stupid stuff that xp has which wastes desktop space is eliminated.