220x10, or 200x11?

govtcheez75

Platinum Member
Aug 13, 2002
2,932
0
76
I think that this is a fairly easy decision, but I thought that I'd get some input. My 2500+ Barton ran Stable (Prime95 Torture for over 48 hours), for almost a month, but when I got some new RAM (Downgrading), I thought I'd try some other settings. It seems that my MoBo didn't like the pc3700 with bh5 chips.

It used to run:
200x11@1.85v
2.5-3-3-11@2.6v w/pc3700 (bh-5) running dual channel @ 100%
38 degrees idle, and 50 degrees load

Now it runs:
220x10 (the 10x multi didn't used to work) @ 1.75v
2.5-3-3-7@2.6v wpc3200 (ch-5) running dual channel @ 100%
34 degrees idle, and 41 degrees load


....I know that higher the FSB, the better the performance, but all the diagnostic tools (such as Sandra, CPU-Z, WCPUID, and even the Control Panel "System" specs) used to recognize my chip as a 3200+. Now most of them don't recognize it, and Sandra says 3034 (estimated).

....I'm assuming that this is normal, since AMD doesn't produce a processor with 440mhz FSB? Is my current setup MUCH better than my previous one? I'm running the Prime95 torture test now, and it seems to be chugging along without any problems for the past 2+ hours. I will let it "burn" for the next few days, as I won't be here. These Temps are Great though! Especially since they are all encased in my Lanboy (which have really sucky ventilation, but is heavily modded with Fans).


Here is my Complete Setup:
-Antec Lanboy w/350w SmartBlue (Heavily Modded with Fans)
-Biostar M7NCD Pro (tried 1208, flashed back to 0424)
-Barton 2500+ (2005 mhz) 220x10 w/1.75v (Running Prime95 Torture Test)
-Buffalo PC3200 256mbX2 running "dual channel" at 2.5-3-3-7 w/2.6v
-Radeon AIW 9700pro running 365/340
-Thermalright SLK900a (lapped) with AS Ceramique, blown with 92mm Panaflo (34idle/42load)
-Vantec Iceberq Chipset cooling, Active on Northbridge w/Ceramique, Passive Heatsink (lapped) on Southbridge w/AS Alumina Adhesive
-2x120mm Fans (1x120mmx38mm Panaflo on Side Panel Intake, 1x120mmx25mm Delta on Blowhole)
-2x92mm NMB Fans (1 Intake, 1 Exhaust)
-1 PCI Slot Exhaust Fan
-Vantec Nexus Fan Controller (Black)
-USRobotics 22mbps Wireless PCI NIC
-Liteon 16x DVD (black)
-Liteon 40x24x40 CDRW (black)
-Windows 2000 Professional (service pack 4)
 

MDE

Lifer
Jul 17, 2003
13,199
1
81
I'd bump the voltage on the memory and try to run it at 11-3-3-2.5 @ 220MHz. My Buffalo did about that at 2.8V on my NF7-S and 2500+.
 

govtcheez75

Platinum Member
Aug 13, 2002
2,932
0
76
Originally posted by: MonkeyDriveExpress
I'd bump the voltage on the memory and try to run it at 11-3-3-2.5 @ 220MHz. My Buffalo did about that at 2.8V on my NF7-S and 2500+.

I don't get it. Why would I do that when it's running at tighter timings now at 2.6v?
 

MDE

Lifer
Jul 17, 2003
13,199
1
81
There was a thread I posted in earlier where someone explained it much better than I did, but the nForce2 works better at 11-x-x-x.
 

pspada

Platinum Member
Dec 23, 2002
2,503
0
0
This ^ is certainly not true in my case - I tried it on 2 different mobos, and with memory at 11-x-x-x, it's much slower than at my default of 8-3-3-2
 

txxxx

Golden Member
Feb 13, 2003
1,700
0
0
FSB effects on Athlon not as noticeable as on P4's (its not as bandwidth greedy), so for Athlon's, its not necessarily a good rule of thumb. Why else is there such small performance increases from 166mhz DDR memory to 200mhz DDR memory?
 

Crucial

Diamond Member
Dec 21, 2000
5,026
0
71
Originally posted by: MonkeyDriveExpress
There was a thread I posted in earlier where someone explained it much better than I did, but the nForce2 works better at 11-x-x-x.

I noticed an imediate improvement in my benchmark scores after making this change. I have a 2100+@2200mhz at 200x11 on an Epox 8rda.

 

myocardia

Diamond Member
Jun 21, 2003
9,291
30
91
Originally posted by: kevinthenerd
Originally posted by: PorBleemo
Originally posted by: VIAN
fsb speed over processor speed

Good rule of the thumb.

Agreed. My P75 ran slower than my P66. (P75 = 1.5 * 50 )
Kevin, you really shouldn't try to be a smarta$$ about something that you don't understand. And govtment, here's the way that it works: with two processors, running at exactly the same speed, the proc. with the higher fsb will
always be faster.
 

Jeff7181

Lifer
Aug 21, 2002
18,368
11
81
Assuming 2.2 Ghz is your maximum overclock, and your processor is stable with a 220 Mhz FSB, and your RAM can run in sync, of course 220x10 would be better.
 

Richdog

Golden Member
Feb 10, 2003
1,658
0
0
Originally posted by: Jeff7181
Assuming 2.2 Ghz is your maximum overclock, and your processor is stable with a 220 Mhz FSB, and your RAM can run in sync, of course 220x10 would be better.

Yup. And as a general rule, Nforce2 boards do indeed perform slightly better with a Tras of 11.:beer:
 

Killrose

Diamond Member
Oct 26, 1999
6,230
8
81
I don't get this...

It used to run:
200x11@1.85v
2.5-3-3-11@2.6v w/pc3700 (bh-5) running dual channel @ 100%
38 degrees idle, and 50 degrees load

Now it runs:
220x10 (the 10x multi didn't used to work) @ 1.75v
2.5-3-3-7@2.6v wpc3200 (ch-5) running dual channel @ 100%
34 degrees idle, and 41 degrees load


This should not have changed this much, you had other issues/scenerios before which could be responsible for what you are able to accomplish now. Could be a summer/winter difference, or maybe when you bought new ram, you cleaned your CPU hs/fan or applied new thermal compound.

Buying new ram should not require CPU core voltage changes.



 

Acanthus

Lifer
Aug 28, 2001
19,915
2
76
ostif.org
Originally posted by: Killrose
I don't get this...

It used to run:
200x11@1.85v
2.5-3-3-11@2.6v w/pc3700 (bh-5) running dual channel @ 100%
38 degrees idle, and 50 degrees load

Now it runs:
220x10 (the 10x multi didn't used to work) @ 1.75v
2.5-3-3-7@2.6v wpc3200 (ch-5) running dual channel @ 100%
34 degrees idle, and 41 degrees load


This should not have changed this much, you had other issues/scenerios before which could be responsible for what you are able to accomplish now. Could be a summer/winter difference, or maybe when you bought new ram, you cleaned your CPU hs/fan or applied new thermal compound.

Buying new ram should not require CPU core voltage changes.


Hes right.
 

Richdog

Golden Member
Feb 10, 2003
1,658
0
0
Originally posted by: Killrose
I don't get this...

It used to run:
200x11@1.85v
2.5-3-3-11@2.6v w/pc3700 (bh-5) running dual channel @ 100%
38 degrees idle, and 50 degrees load

Now it runs:
220x10 (the 10x multi didn't used to work) @ 1.75v
2.5-3-3-7@2.6v wpc3200 (ch-5) running dual channel @ 100%
34 degrees idle, and 41 degrees load


This should not have changed this much, you had other issues/scenerios before which could be responsible for what you are able to accomplish now. Could be a summer/winter difference, or maybe when you bought new ram, you cleaned your CPU hs/fan or applied new thermal compound.

Buying new ram should not require CPU core voltage changes.



A higher multiplier setting needs more voltage, therefore it is understandable that he needed 1.85v @ 11*200... whereas at 10*220 he needed only 1.75v.:beer: