How to run XP2600+ / 166 at 16x 133MHz on Gigabyte GA-7VRXP rev. 2

[enycing]

Is there a way to get this cpu to run at 16x 133MHz on a Gigabyte GA-7VRXP board? (rev. 2, BIOS F12, older versions available).

I don't really need or want to overclock it, just run it at its proper speed.

CPU is AXDA2600DKV3D AQUCA 0312RPAW, pre-week 43/46, should not be locked.

Currently, it is recognized as an XP2000+ (12.5 x 133 MHz).

Running this board with an FSB of 166 MHz seems to be technically impossible due to chipset limitations if I followed older threads correctly.

I could force 166 MHz by using the F6 BIOS and lose my COP, or go through the hassle with rebooting every time.

Currently I am not using it often enough to leave it on standby all the time.
16 x 133 would be my preference, if possible somehow.

Is there a way, like changing bridges on the cpu ?

[enycing]

I know I could set the multiplier to 8 or 8.5 x.

Then it is a matter of forcing the XP to use the high multipliers.
I just don't know how to do that. Anybody?

Enyc.

[enycing]

I finally got the time to check and sort out what I already downloaded from this forum.
Cutting L3-5 is not my preference, but connecting it through a 200 or 220 ohm resistor to Vcore should also work theoretically (if L3-5 is connected to Vss through 1K internally, which should be the case according to some information from a German forum (or this one, I' m losing track again).
(i.e. through pin AJ27 on the underside of the board).

I read in some of Hood's reply's this board will not let you adjust multi's from 13 and up.
So before I start soldering components to this board, I would like to know if somebody has ever tried this.

[enycing]

I connected pin AJ27 (BP_FID:4) on the underside of the board to pin AH26 (Vcc) with a 200 Ohm resistor. This activated all above 12.5 multipliers.

At 133 MHz, the only working switch settings are 9 and 9.5, resulting in 17x and 18x multipliers. All others resulted in a black screen, although accompanied by the usual disk activity. BIOS reports "AMD Athlon processor, 2267 MHz" (and 2400 MHz for the 9.5 muliplier) frequency.
At 100 MHz, I tested 8 up to 10, resulting in:
"AMD Athlon XP 1900+ up to 2200+".
133/2267 and 133/2400 should be roughly equal to 2700+ and 2900+.

I have been running prime95 torture test (windows98) for a few hours now at 2267 MHz, max cpu temperature was about 50 degrees C. (no case, board on tabletop, cheap TR2-M3 cooler with 8mm 2500rpm fan). I'll set the switch to 9.5, and try running prime95 for the next 10 hours or so. More results may follow.
(while writing this, I checked and prime95 halted on the 7th iteration with a fatal error message (hardware failure detected), so 2400 MHz appears to be to much with all settings at their standard values (voltage etc.). I'll continu to run prime95 a bit longer at 2667 MHz.

SEE follow-up for final results.

[enycing]

At the first run, lower multipliers below 8.5 (133 MHz) were not working.
Strangely, after trying all multiplier settings at a bus frequency of 100 MHz, they did (and still do).
Startup BIOS message was one of these two, depending if BIOS recognized the setting:

ATHLON XP (tm) XP ....+
ATHLON XP (tm) processor, .... MHz

switch|m.plier|100 MHz FSB cpu MHz|133 MHz FSB cpu MHz (-- = blank screen)

5 13 -- --
5.5 13.5 -- --
6 14 -- --
6.5 21 -- --
7 15 1800+ 1500 2400+ 2000
7.5 22 2800+ 2250 --
8 16 1900+ 1600 2600+ 2133
8.5 16.5 1900+ 1650 2700+ 2200
9 17 2000+ 1700 ,2267 2267
9.5 18 2200+ 1800 ,2400 2400
10 23 ,1000 1000 --
10.5 24 ,1050 1050 --
11 - -- --
11.5 19 2200+ 1900 --
12 - -- --
12.5 20 2400+ 2000 --
(&auto)

I have run Prime95's torture test for about 24 hours, temperature was about 43 / 51 degrees C (case / cpu) with the 17x multiplier (2267 MHz).
No problems at all :-)