Performance drop by lower HTT?

S

ScHpAnKy

2[H]4U
Joined
Jan 11, 2003
Messages
3,332
I was wondering what performance drop is associated with changing the A64's HTT multiplier from 5 to something along the lines of 2 or 3?

Does anybody have the time required to benchmark their Athlon 64's with these supplied multipliers?
Right now I'm not at home, and probably won't be for a few days, and I'd love to find this out and make a definate answer, because I couldn't find this answer for the life of me,
 
You want to keep it under 1000....so just use 3x...

I dont think there is a huge performance drop between 5x,4x,3x etc...you just want to keep it under 1000.
 
mine is at 4x with a fsb of 258. Is this ok? Or will i get better perfomance if i drop the HT?
 
i feel like a noob for asking, but i really know nothing about the 64s yet

what is the reason for keeping "it" under 1000?
 
A64's have something called HTT, which is basicly FSB. 1000 HTT bus is the max some A64 boards support.
 
Circuitbreaker8 said:
A64's have something called HTT, which is basicly FSB. 1000 HTT bus is the max some A64 boards support.
Click to expand...

yeah i get that. but everyone says to stay under 1000 for no apparent reason. i mean if you can get it to 1100 and everything is stable would that not be a good thing?
 
distr0 said:
yeah i get that. but everyone says to stay under 1000 for no apparent reason. i mean if you can get it to 1100 and everything is stable would that not be a good thing?
Click to expand...

Its not a good thing b/c the board isnt meant to go over 1000 HTT
 
Circuitbreaker8 said:
Its not a good thing b/c the board isnt meant to go over 1000 HTT
Click to expand...

and cpus aren't MEANT to go above their stock speeds... but they do :)
 
My HTT is running at 1125. Seems to be very stable. I also noticed a big drop in performance if it is set to 4x. I lost almost 3000 points in Aquamark 3.
 
I already know all of what this thread has to offer, yet nobody even hit on the original question... what is the performance drop associated with lowering the HTT multiplier?
 
ScHpAnKy said:
I already know all of what this thread has to offer, yet nobody even hit on the original question... what is the performance drop associated with lowering the HTT multiplier?
Click to expand...
none....
 
ScHpAnKy said:
I already know all of what this thread has to offer, yet nobody even hit on the original question... what is the performance drop associated with lowering the HTT multiplier?
Click to expand...
centvalny said:
try any benchmark program, if it not stable try 3x
Click to expand...
you overlooked, someone did touch on the question... sometimes you just have to try things for yourself..
 
ScHpAnKy said:
I already know all of what this thread has to offer, yet nobody even hit on the original question... what is the performance drop associated with lowering the HTT multiplier?
Click to expand...
The perfrormance difference of the htt between 800-1100 is not even noticable. The only thing I notice is when I start to get close to 1100 my board gets unstable (abit and epox).
 
Is stablility the only reason to keep the mulitplier low? I have a NF3 (800Mhz HTT) at over 1000Mhz for a while now.
 
in an earlier thread I posted these benchmarks

http://www.hardforum.com/showthread.php?t=846802

system
amd64 3200 @ 270*10 1/1 (2.5,3,3,10,1t)
6800u @ 461/1.16

3X=810
superpi 1m = 31sec
3dmark03 = 14197

4X=1080
superpi 1m = 31sec
3dmark03 = 14190

The htt multiplier seems to have no impact for me.
I'd keep it at the lower setting and go for more of an overclock.
Just my opinion.

My sisoft sandra scores were like 7650 vs 7620 or something thereabouts...

Hope this helps a little...
 
karnick said:
you overlooked, someone did touch on the question... sometimes you just have to try things for yourself..
Click to expand...

I still fail to see how that relates to my original question, I also cannot try this experiment because my pc is currently down and will be for a few days more.

spin02 said:
Is stablility the only reason to keep the mulitplier low? I have a NF3 (800Mhz HTT) at over 1000Mhz for a while now.
Click to expand...

Not really, I dropped my multi because my FX-55 wouldn't get passed 215 FSB with a 4x/5x multi, but I'm still pretty disappointed because the FX-53 I'm selling had a better overclock :(

jiminnyjellickerz said:
http://www.xtremesystems.org/forums/showthread.php?t=39559
Click to expand...
wormburner said:
in an earlier thread I posted these benchmarks

http://www.hardforum.com/showthread.php?t=846802

<snip>
Click to expand...

THIS is what I was looking for! Thank you so much!
 
ScHpAnKy said:
Not really, I dropped my multi because my FX-55 wouldn't get passed 215 FSB with a 4x/5x multi, but I'm still pretty disappointed because the FX-53 I'm selling had a better overclock :(
Click to expand...
hahaha! that makes my day... proves teh fact that its not modle that matters with a64 its luck of draw/stepping... guess thats what you get for getting a oem eh?

why not just keep the fx-53 sean
 
ryuji said:
hahaha! that makes my day... proves teh fact that its not modle that matters with a64 its luck of draw/stepping... guess thats what you get for getting a oem eh?

why not just keep the fx-53 sean
Click to expand...

OEM? It was retail
And I already sent in the FX-55 to see if I get back a better CPU
SOI is overrated :(
 
You must log in or register to reply here.
Back
Top