• 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.

Overclocking a 2.8c Comparrison Questions

M

Magi

n00b
Joined
Sep 2, 2004
Messages
33
I have a 2.8c, p4c800e delux, 2x512 kingston value ram, antec true 480w, zalman all cu 92mm heatsink fan

Using AI Overclock i can get it to 3.079 stable at 1:1 fsb 220
I can get 3.36 at 5:4 fsb 240 memory at 192

I tried AI at 30% at 5:4 for 260 and no joy I also did a manual overclock with agp bus lock at the 66/33 ratio 5:4 ratio at fsb 250 so the memory is at 200, and the vcore at 1.775 (pcprobe shows 1.8 i think) and that doesnt run stable....does the vcore need to go up higher, since the ram limitation should take effect with it running at 200mhz. The chips i thought could go higher. If this chip is just a "bad" chip what are your guys opionion of a real life better performer 3.079 ddr 440 or 3.36 ddr 400?
 
1.7-1.8v on vcore is getting right in the 'scary' range. That's the kind of voltage that extreme OC's need. Be careful with that voltage.

Sounds to me like your ram is holding you back. The value select ram isn't known for its OC ability. How much voltage do you have going to your ram? Also, make sure your ram timings are nice and loose. I will usually loosen my timings on my ram all the way while trying to find a maximum OC. Also, don't be afraid of the 3:2 ratio. I've seen some very good OC's using this ratio.

Have fun messing with it and good luck!!
 
My ram timings are 2.5 3 3 8, for when the ram is at 192 w/ cpu at 3.36, and its 2.5 4 4 8 when its at 220 with cpu at 3.079. I thought if the divider was at 5:4, while the fsb was 250, the ram is only working at 200...., so doesnt that effectively eliminate the ram as holding me back or is it a tad more complicated that that?
 
Hey Magi,

I've got a similar rig, and from what I can gather - it's most probably your MB holding you back.

Board seems to have a volt 'droop' problem. If you load up MBM a take a look when you stress out the CPU you'll see a substantial drop (well below 1.5v). I believe this might be holding you back.

My 2.8 is currently at 3.43 (100% Prime stable) with just a touch more volts (1.575) to the core and a little more for the ram (1.65) - I'm running a 5:4 ratio.

You could try and 'droop' mod your board, there are plenty of guides. Me, I'm happy with what I'm currently getting, though I would have liked to hit 3.5!

Good luck!

Jelly :)
 
I will come back and try maybe 245 at 5:4 because 240 seams ok if I use AI overclock, but not tonight Labs take precident to computer....of course i could be kidding myself and youll see me back posting away in a couple hours anyway.

Hows about the trade offs between running 220 memory at 3.079 vs 192 memmory at 3.36 any opinions on what you guys would rather run?
 
Hiya, this is my first post on these forums, but I have a similar config so I thought I would let you know how its going for me.

I usually run my 2.8 at 3.5 (250 FSB with 5:4 timings) It runs steady and cool with only 1.575 volts on the CPU. Its stable (PRIME95 for 10+ hours) and I can take it to 261 FSB, 1.6 volts, but see almost no difference in the benchmarks I have been using (Aquamark, 3DMark03, Sandra) so I just leave it at 250.

I can take the system up to 223 with 1:1 timings but my ram holds be back from anything higher than that (like yours) and performance is better at the 250 5:4 timings. You have a better power supply than mine so thats probably not the problem. I would have to agree its most likely the Motherboard. All chips have a different threshold though, and even if you can't get past 240 FSB, thats a pretty significant gain over stock.
 
well it posts and gets into windows at 1.625 V with 245 fsb and 5:4 but fails when loading all my windows proggies. So its basically certain the voltage droop is killin me. I guess i could pump up the volatage higher so the droop wont go too low maybe. I have the zalman copper 92mm cooler on it at the 220 1:1 the idle temp is 27c. I forget what it is under load. Thats at 1.6V. Is the only thing that limits the amount of volatage the added heat that comes with it?

3.36 240 fsb @ 5:4 192 memory or 3.079 220 fsb 1:1 220 memory? Though its probably a much of a muchness huh.
 
27c idle?! You live in an igloo or something?! That pretty good - is that after 1hr sitting?

Anyway, with that GPU of yours, I'd be looking to get your chip as quick as possible, certainly 3.36 if you can. Anything less and you're most likely introducing a CPU bottleneck.

As for volts, probably best to keep it around 1.6 - that should be good for a good long CPU life.

All the best.

Jelly.
 
27c was on all night....its cold in my room though cause i leave the windows open. I also have 2 80mm in 2 80mm out. Then the zalman so i have pretty good cooling thanks for the replies guys. Im gonna run with jellys suggestion i think and go at 3.36, im gonna play around and see since my memory speed is effectively lowered by doing that if i can drop the timings on the ram as well.
 
Do some benchmarks both ways and see which is better.

I'm leaning towards the 3.3 also.

Good luck and let us know how it turns out.
 
3.36 is faster for my config here are some results if anyone cares or wondered whether memory bus vs straight clock means is better. I forgot to lower timings for sandra on 3.36 but you guys get the drift It runs stable at those timings for under memory 200, but at 200 those timings crap out. They are the timings as if it was pc2700 or so cpuz says. It runs them stable for 192 though Whats funny is the memory bandwidth in sandra for both is about the same, and i can only assume that should i run sandra with the lower timings that it would make up the miniscle difference.

3.079 FSB 220 1:1 Memory 220 2.5-4-4-8
3dmark 2k1: 18360
Sandra Arithmatic 13,073
Sandra Multimedia: 56962
Sandra Memory Bandwidth: 9566
Sandra Combined Performance Index: 3345

3.36 FSB 240 5:4 Memory 192 2.0-3-3-7 (for 2k1) and 2.5-4-4-8 (sandra)
3dmark 2k1: 19,619
Sandra Arithmatic 14,222
Sandra Multimedia: 62,237
Sandra Memory Bandwidth: 9568
Sandra Combined: 3468
 
You must log in or register to reply here.
Back
Top