Lower multi for A64?

[f4phantom2500]

I know a lot of people lower their multipliers to achieve higher FSB's and such, but from what I've learned that doesn't seem to matter on an A64; the major concern is the resulting CPU speed. I've found that a lot of people with lower-end processors (3000, 3200, etc) seem to get high overclock % yields compared to people with 3500's, 3800's etc. Is it because they use a lower multiplier, or is it because the architecture caps out at about 3GHz so they just inherently have that much more headroom? In some instances I even see the lower end ones clock higher than higher ones, not just relative to it's stock speed but the resulting frequency is actually higher. Is it because people with 3000's that get it to OC that high flaunt it and those who don't (the majority of 3000 owners?) just don't, or do most 3000's OC to like 2.6+ GHz? Basically what I'm trying to ask is if lowering the multiplier to that of a 3200 or 3000 give me a higher overclock overall, or will it still cap out at the same speed? If it won't make a difference, then do people with those chips that get in the high 2GHz range just happen to have good OCing chips? My 3500 Venice didn't seem to want to quite hit 2.6GHz at like 1.55v, so I was wondering if lowering the multi would have helped. Also I'm currently OCing an X2 3800, and at the voltage I'm using, ~1.48v, it seems to be capping out at 10x254, so I wanted to know if lowering it to 9 or even 8 would allow me to hit a higher overclock (assuming my motherboard can take it...I know I can boot into windows at at least 285, so lowering it to 9 may be a viable option if it will mean a higher CPU OC).

*Note I know the differences among the cores, but I generally mean among processors with the same core, I have Venice in mind.

 

[Nickel020]

I think lowering the multi and increasing HTT (FSB) will not usually give you a higher overclock, but will give you a higher memory speed, which may be better than a slightly higher overclock with a lower HTT, depending on what you're doing. Best way to find out though is give it a try Lower your multi by 1 or 2 and try for max. HTT, then run some benchmarks.

 

[GuitarDaddy]

For some odd reason some AMD chips seem to O/C better with a lower multi. I had a 3200+ that could not go past 2.48ghz with the stock 10x multi, but I could hit 2.61 with the 9x.

 

[f4phantom2500]

Originally posted by: GuitarDaddy
For some odd reason some AMD chips seem to O/C better with a lower multi. I had a 3200+ that could not go past 2.48ghz with the stock 10x multi, but I could hit 2.61 with the 9x.

Hm...interesting, I'll definitely give it a try; I was hoping I'd be able to at least break 2.6GHz, and if that's the case with this processor I think that'd do it because it's only 60MHz away right now.

 

[LittleNemoNES]

Originally posted by: f4phantom2500

Originally posted by: GuitarDaddy
For some odd reason some AMD chips seem to O/C better with a lower multi. I had a 3200+ that could not go past 2.48ghz with the stock 10x multi, but I could hit 2.61 with the 9x.

Hm...interesting, I'll definitely give it a try; I was hoping I'd be able to at least break 2.6GHz, and if that's the case with this processor I think that'd do it because it's only 60MHz away right now.

In this case most likely it was the mobo that was limiting the OC.

 

[GuitarDaddy]

Originally posted by: gersson

Originally posted by: f4phantom2500

Originally posted by: GuitarDaddy
For some odd reason some AMD chips seem to O/C better with a lower multi. I had a 3200+ that could not go past 2.48ghz with the stock 10x multi, but I could hit 2.61 with the 9x.

Hm...interesting, I'll definitely give it a try; I was hoping I'd be able to at least break 2.6GHz, and if that's the case with this processor I think that'd do it because it's only 60MHz away right now.

In this case most likely it was the mobo that was limiting the OC.

It takes much higher FSB to overclock with a lower multi, how can the mobo limit the overclock at a lower FSB and not the higher FSB?

10x248 limited by mobo???

yet it will do 9x290

actually with an even lower multi my max stable FSB was 325mhz

Try Again, thanks for playing

 

[LittleNemoNES]

lol smartazz, i meant that it was a quirk of his particular mobo.

 

[f4phantom2500]

Originally posted by: gersson
lol smartazz, i meant that it was a quirk of his particular mobo.

Oh so you're saying that for some weird reason his mobo could OC the chip higher if a lower multi was used.

I tried lowering the multi to 9 and 285 wasn't stable. I didn't even want to mess with 284 because I had to go out of town this weekend and figured there was a good chance it'd run for a few hours and then crash, plus it's just 16MHz faster than 10x254, plus I can get the RAM closer to 200MHz on 254.

 

[GundamSonicZeroX]

Lowering the LDT seems to produce more of a result in my case on my Opty 165.