• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Windows SMP > Linux SMP!

Y

Yamaki

Gawd
Joined
Dec 27, 2002
Messages
741
I recently downclocked one of my 3GHz E2160s to 2.8GHz because it was getting a bit hot for my liking. It's running XP32 with the SMP client and SMP affinity changer (all 3 processes set to Normal priority). While making the downclock and getting my other 3GHz machine up and running, I noticed something:

3GHz, WinXP, Affinity Changer = 1570PPD
2.8GHz, WinXP, Affinity Changer = 1360PPD (huge drop O_O)
3GHz, Linux (notfred's CD) = 1330PPD (!)

I don't know if this is for real or there's just something wrong with my linux client...but it appears that windows with affinity changer performs better than Linux.

These two machines have the same hardware pretty much, except the windows one has PC6400 RAM and the linux one has PC5300.

Has anyone run into a situation like this that could back up my claim/resolve my issue (if there is one)?

Thanks.

-Yama
 
Which protiens where you folding on all three tests ?

I'm getting a +200 PpD running Linux vs Windoze.
Running p2653 on the same hardware after swopping to Linux (~1700 PpD) from Windoze (~1500 PpD).
I can get a big variation of between 200-330 PpDpGhz when running the SMP client on 2 cores depending on which protien I'm folding.
So without knowing exactly which protien your folding, its hard to tell if its the OS or protien that affecting the variation the most.

Luck ............... :D
 
Your folding slow on all of them ........... :mad:

p2653 at around 310 PpDpGhz is a little slower then p2605 at around 330 PpDpGhz.
Your only getting 220-260 PpDpGhz at best.

With my Linux boxen at 2.65 Ghz with p2605, I'm getting 1,700 PpD. 320 PpDpGhz.
Scale that upto 3.00 Ghz and I'd expect +1,900 PpD.
My Windoze boxen at 3.00 Ghz with p2653 gives me 1,850 PpD. 310 PpDpGhz.

Luck ............... :D
 
Don't forget this is done on a nForce 7050/7100 motherboard which doesn't support dual channel ram. This may explain why it's a bit slow.

 
Tigerbiten said:
Your folding slow on all of them ........... :mad:

p2653 at around 310 PpDpGhz is a little slower then p2605 at around 330 PpDpGhz.
Your only getting 220-260 PpDpGhz at best.

With my Linux boxen at 2.65 Ghz with p2605, I'm getting 1,700 PpD. 320 PpDpGhz.
Scale that upto 3.00 Ghz and I'd expect +1,900 PpD.
My Windoze boxen at 3.00 Ghz with p2653 gives me 1,850 PpD. 310 PpDpGhz.

Luck ............... :D
Click to expand...

Even with the good proteins you have to keep in mind he's using 1 meg cache processors and they fold a bit slower on the SMP units.

That said, they are folding a bit slow. My X2@3.0Ghz is doing 1368 PPD on a 2605. This is Linux installed onto the hard drive. It's mostly just a folding box but also runs FahMon to monitor all my systems. You should be able to match or beat that a little bit with those processors at 2.8Ghz on the same protein.

 
Xilikon said:
Don't forget this is done on a nForce 7050/7100 motherboard which doesn't support dual channel ram. This may explain why it's a bit slow.

Click to expand...

That might also explain some of it. The nForce chipsets don't seem to fold as well as the Intel chipsets. I know a couple of people have found that at the same clocks the P35 chipset folds faster than the 680i. Considering that's supposed to be the high end chipset of the last generation I would expect the lower end chipsets to possibly fold even slower.

 
First things first for all the PPDpGHz numbers being questioned and compared......E2xxx CPUs do not fold at the same PPDpGHz as E4xxx and E6xxx CPUs. It's been shown over and over. E4 and E6 are nearly identical, but there's a pretty big hit going to the E2s.

When you dropped the system to 2.8, did you wait long enough for Affinity Changer to work it's magic? It takes 10 minutes for it to check and set the affinity for the first time.

As far as your linux vs windows question, linux should still be a little faster than Windows when installed on a hard drive. I can't imagine that folding from a CD is going to be the most efficient way to do it, so that may have something to do with your dramitic difference in Window's favor. That said, there still seems to be a problem. Maybe not enough ram in that system or it's thermal throttling.

 
Well, the system has 2GB of ram (2x1gb). Maybe I should remove one stick.

EDIT: Removed one stick and tightened the timings from 5-5-5-15 to 4-4-4-12. Now it's up to 1400PPD...it seems like it should be around 1600, especially considering that my 2.8GHz windows machine is still pushing 1360PPD.
 
Lemme finish my 610i/E2160 box and I will report back the results. I'm right now installing Windows XP on it.

 
Xilikon said:
Lemme finish my 610i/E2160 box and I will report back the results. I'm right now installing Windows XP on it.

Click to expand...

I just took my E4500 that was in a P35 and put it into my 610 board. I dropped from 2100 to 1950 ppd.

This does not make me happy, but I had to make room for a new addition :p

 
So far, I'm able to overclock to 2.8 GHz and I now believe I may have a poor overclocker because the default VID of my E2160 is 1.325, which is pretty high. However, it's rock stable at 2.8 GHz and the temps never broke 50C with the Q6600 stock cooler.

I'll report the PPD in 1-2 hours since it just started right now (I had to pull a lot of hairs to make the Endian Firewall work as VM and to be able to use RDC from my Vista box to monitor/tweak remotely).

 
Kendrak said:
I just took my E4500 that was in a P35 and put it into my 610 board. I dropped from 2100 to 1950 ppd.

This does not make me happy, but I had to make room for a new addition :p

Click to expand...

This may be the reason that my PPD sucks so much. It seems the 610i chipset, or any nVidia chipset for that matter, is far inferior to the Intel chipsets.

Looks like I'll be buying P35 next time.
 
610i chipset doesn't support dual channel memory and that's why it sucks (beside the weaker optimisations versus the Intel chipsets). I'm also considering going to P35 next time as well but it's hard to beat a motherboard with integrated graphics for 50$ :(

 
You must log in or register to reply here.
Back
Top