Brand new 2500XP-M Barton really amazing

Just replaced my XP2500 Barton with this new mobile and right away got it to 2555mhz w/ 1.875vcore. It runs faster and cooler than my old XP2500 Barton which I could only take up to 2480mhz.

Give me a few days to figure out why I can't go over 225mhz. I'll try the L12 mod to see if I can go higher.

what L12 mod. Are the mobile the same as the reg xp's. I was thinking of buying a mobile 2500+, but im worried they might have some things turned off or something. I dont want pin mod the cpu or whatever just to get it to work.

So are they same as a desktop cpu just labeld mobile ?

If u want a comparision to see if youre cpu is where it should be Here is pic of mine @ 2600mhz, 450 mhz more so ur thing is looking good.

They are the same as the desktop ones only on a desktop board will have a stock multiplier of x6 and FSB of 133. The only thing that is "turned off" (and its only on an nForce2 motherboard) is the powernow feature which is requiered on a laptop to change speed on the fly. this will not affect anything when running on a desktop motherboard tho, so you shouldn't worry
L12 trick is done on 133MHz FSB chips to make the mobo "think" its a 166MHz chip by default, on many NF7 motherboards it helps to achieve a higher FSB.

Well I got it up to 2.65Ghz w/ 2V. running cool w/ low ambient temps.

Still trying to decide if doing the L2 mod will be worth the extra work since I have to remove my mobo from the case in order to remove mu HSF.

About the CPU temperature, I'm pretty sure 33C is your CPU socket, not your CPU core. The core temperature should be around 46C idle in your case. That is a great OC.

Better keeping pushing it. Try and break 10,000 Mips.

{PMS}fishy said:

Better keeping pushing it. Try and break 10,000 Mips.

Click to expand...

Strange thing is that the highest I got was 9782mips w/ my Xp2100 tbred b running at 2.6ghz and now only 9233 w/ my mobile barton at 2.62ghz. I ran them both w/ CAS2-2-2 ram settings.

The Dhyrstone integer IPC (instructions per cycle) for XP/Barton is about 3.76 instruction per cycle.
Ref: What is IPC and how to compare cycle or Hz for different CPU architectures (page 19)


From your number, you are getting 9233 MIPS at 2656 MHz, which is 3.48 instructions per cycle, which is kind of low.
Have you checked whether there are some background tasks running. Memory CAS timing should have little effect on Dhystrone benchmark whose codes are mostly executed in cache for measuring CPU raw computing power.


For my mobile Barton 2400+, I am getting 10553 MIPS at 2784 MHz, which is 3.79 instructions per cycle (which is in line).
Ref: Overclocking a mobile Barton 2400+ to 2.6/2.7+ GHz on air (page 18)
From the Sisoft Dhyrstone Integer and Whetstone Floating Point benchmark,
it scores 7%+ higher than a P4C 3.2 GHz w/ 2 SMT, for both integer and floating point (w/o SSE2).
Since the Dhrystone IPC ratio for XP to P4C (w/ 2 SMT) is about 1.22:1,
so a 2.79 GHz Barton would perform about the same as a 3.4 GHz P4 in integer arithmetic.

Which is better, L12 MOD or increasing FSB? I want higher score on benchmarks.

L12 mod doesn't have nothing to do with performance. on NF7 motherboards it sometimes lets a 133MHz FSB chip reach high max FSB

Sniper.nkc said:

Which is better, L12 MOD or increasing FSB? I want higher score on benchmarks.

Click to expand...

Short answer: Higher FSB is better for performance, mainly for memory bandwidth between the memory controller (in NB) and CPU for XP motherboards. L12 is just a potential mean (may or may not work) to achieve that.

If aiming for only high overclocking of an (multiplier) unlocked CPU, FSB and potential L12 mod would NOT help CPU overclocking. FSB and CPU frequency are independent parameters (at leat to first degree).


If you have an unlocked XP CPU such as a mobile Barton or an early Tbred B/Barton (older than week 0338 or so), then higher FSB and higher CPU frequency can be achieved at the same time by appropiately setting the CPU multiplier

CPU_frequency = CPU_multiplier x FSB

So higher FSB is better provided the motherboard, in particular the North Bridge and its memory controller, and the memory allow such FSB overclocking. One may need to improve chipset cooling (SB, and NB in particular) and give enough Vdimm to memory (not uncommon to use 2.9-3.0 V for BH5, CH5 memory chips) to achieve that.

For nforce2 rev 2.0 (aka 400 Ultra) motherboards, such as NF7-S rev 2.0, there is a good chance to overclock the FSB to 200 - 230 MHz without extensive mod (chipset cooling and Vdimm).

There are certain NF7-S rev 2.0 (not every board) whose FSB overclocking can be greatly facilitated by the so called "L12 mod" which alters the CPU default FSB_sense[0] and FSB_sense[1] signals from 133 MHz to either 166 MHz or 200 MHz.

I personally tried on two NF7-S rev 2.0, one board improved by 10-15+ MHz on FSB (from 215 to 225-230 MHz), and one board did not at all (0 MHz, but that board does 227 MHz with stock everything).
Ref:
FSB_Sense mod to set Tbred B default FSB from 133 to 166 (page 12)
FSB_Sense mod to set Tbred B default FSB from 133 to 200 (page 12)

NF7-S rev 2.0 FSB_Sense 166 and 200 mod (from 133) test results (page 12)
NF7-S rev 2.0 FSB_Sense 166 and 200 mod (from 133) test results (part 2) (page 12)
Summary on overclocking the NF7-S rev 2.0 (with Tbred B 1700+ DLT3C and ...) (page 15)