AXP 1700+ (Palomino) overclocking

Today I borrowed an Athlon XP 1700+ (Palomino) from my friend. The CPU is reported "model 6, stepping 2" in WCPUID and CPU-Z. I wanna know how far "stepping 2" Palominos overclock with cheap cooler and 1.85v?

Now here's my overclocking experience with this CPU. I used a cheap alluminium based coolermaster HSF (no thermal pad/paste) which is fairly silent. Rest of the system:

ABIT NF7 v2.0
2 x 256 MB Kingston Value RAM PC2700 (rated CL2.5 for PC2700, CL2.0 for PC2100)
Gigabyte Radeon 9000 (not overclocked)
300W "Generic" PSU
2 case fans

CPU Temp (full load): 54C at stock speed & voltage, 59C at 1540 MHz (1.85v)

Special note about CPU voltage: Working CPU voltage with this CPU was always lower than the voltage set in BIOS. For example, The CPU ran at 1.71 to 1.73v when it's 1.75v at BIOS. In case of 1.85v, it ran at 1.81 to 1.82v

I used mainly prime95 for stability test:

<b>1470 MHz (stock speed), 11 x 133 MHz, 1.75v (default), memory sync with FSB/2.0-2-2-5</b>
Running prime95 since last 1 hour , no crash. Ran few benchmarks (3DMark 2001/2003)

<b>1540 MHz, 11 x 140 MHz, 1.80v, memory sync with FSB/2.0-2-2-5</b>
prime95 torture test crashed within few seconds, benchmarked system with 3DMark 2001 (1 time), played Mafia for ~1 hour

<b>1540 MHz, 11 x 140 MHz, 1.85v, memory sync with FSB/2.0-2-2-5</b>
prime95 torture test crashed after 2 minutes

<b>1540 MHz, 11 x 140 MHz, 1.85v, memory sync with FSB/2.0-3-3-6</b>
prime95 torture test crashed after 12 minutes

<b>1575 MHz, 11 x 143 MHz, 1.85v, memory sync with FSB/2.0-2-2-5</b>
prime95 torture test crashed within few seconds, unusual windows explorer crash

<b>1575 MHz, 11 x 143 MHz, 1.85v, memory sync with FSB/2.0-3-3-6</b>
prime95 torture test crashed within few seconds

<b>1575 MHz, 11 x 143 MHz, 1.85v, memory sync with FSB/2.5-3-3-6</b>
prime95 torture test crashed within a minute. This time I also used 3DMark 2001 (inf. loop) for stability testing (640 x 480 x 16 bit, Software T&L for stressing particularly the CPU). It crashed after with 20 mins.


The CPU I regularly use is a Duron 1 GHz, overclocked @ 1.15 GHz (morgan core, model 7, stepping 0). Rest of the system spec is same (most possibly the HSF, too. Also didn't use thremal pad/paste). In case of this CPU, the working voltage doesn't vary from the voltage set in BIOS. CPU runs at 51C under full load (at 1.15 GHz, 1.8v). Here's the overclocking result of this CPU-

<b>1000 MHz (stock speed), 10 x 100 MHz, 1.75v (default), memory sync with FSB/2.0-2-2-5</b>
prime95 torture test ran stable for 12+ hours (didn't crash, I stopped the test)

<b>1100 MHz, 11 x 110 MHz, 1.75v (default), memory sync with FSB/2.0-2-2-5</b>
prime95 torture test ran stable for 12+ hours (didn't crash, I stopped the test)

<b>1150 MHz, 11 x 115 MHz, 1.80v, memory sync with FSB/2.0-2-2-5</b>
prime95 torture test ran stable for 12+ hours (didn't crash, I stopped the test)

<b>1200 MHz, 11 x 120 MHz, 1.90v, memory sync with FSB/2.0-2-2-5</b>
didn't run prime 95. Crashed while installing a driver


Now I want to know how far the AXP can with my cooling? Why the CPU voltage is always lower than the voltage set in BIOS (for this particular CPU)?

----------------
<b><A HREF="http://geocities.com/spitfire_x86" target="_new">My Website</A></b>

<b><A HREF="http://geocities.com/spitfire_x86/myrig.html" target="_new">My Rig & 3DMark score</A></b>

well..you know that the fastest Palomino is the 2100+@1733Mhz, since you are using a cheap cooling solution, I wouldn't expect much.
I used to have a 2000+ Palomino, it crashes when I set the FSB more than 140Mhz, no matter how high the voltage I set
I need to set the voltage at 1.80V to make it run at 139x12.5=~1733Mhz(2100+)

it crashed in Windows when FSB>140Mhz

If you have anything that would improve those temps I would try it. Perhaps you have some grease (any will work better than nothing) that you could use, it doesnt damage anything.

<A HREF="http://service.futuremark.com/compare?2k1=6752830" target="_new">Yay, I Finally broke the 12k barrier!!</A>

Unfortunately, I can't use thermal pad/grease. It's very hard to find thermal grease here.

Can anybody answer why my CPU voltage (with this CPU only) is unstable?

----------------
<b><A HREF="http://geocities.com/spitfire_x86" target="_new">My Website</A></b>

<b><A HREF="http://geocities.com/spitfire_x86/myrig.html" target="_new">My Rig & 3DMark score</A></b>

ok, I just re-read ur post, looks like it happens only with ur AXP
well, i don't know the exact reason, but my voltage isn't stable either, and the reason for me, it's because I enabled <b>CPU Halt Command Detection</b> in Bios, and therefore my CPU's voltage varies between 1.6~1.7 instead of 1.65.
but it stays at 1.65 when I disabled it.

I don't have this option at BIOS, but I had "CPU Disconnect" option enabled. I think it's the same option as your's, and disabling it isn't stabilizing my voltage.

Currently I'm running it at <b>1531 MHz, 11 x 139 MHz, 1.80v, memory sync with FSB/2.0-2-2-5</b>. prime95 torture test is running for 6.5 hour (haven't aborted it yet) . Don't know disabling "CPU Disconnect" feature in BIOS helped or not. Will try FSB 140 MHz and up again.

----------------
<b><A HREF="http://geocities.com/spitfire_x86" target="_new">My Website</A></b>

<b><A HREF="http://geocities.com/spitfire_x86/myrig.html" target="_new">My Rig & 3DMark score</A></b>

CPU Temp. is same, and this time it's more stable with 140 MHz FSB (generates error in prime95 after 6 minutes, compared to immediate error of pervious time). My PSU may be guilty for unstable voltage, this can also be a cause of unstable OC in 140 and higher FSB

----------------
<b><A HREF="http://geocities.com/spitfire_x86" target="_new">My Website</A></b>

<b><A HREF="http://geocities.com/spitfire_x86/myrig.html" target="_new">My Rig & 3DMark score</A></b>

I may try again with 1.85v. If it again gives error, then I'll switch back to 139

----------------
<b><A HREF="http://geocities.com/spitfire_x86" target="_new">My Website</A></b>

<b><A HREF="http://geocities.com/spitfire_x86/myrig.html" target="_new">My Rig & 3DMark score</A></b>

I have a 1600+ Pally with a Volcano 9 HSF. It does 1.73GHz on stock voltage, or 1.80GHz with the voltage raised a wee-bit. You probably just have a crappy stepping, or it's inadequate cooling. Mine idles just under 50C, and under load it's just over 50C (at 1.8GHz). That's with a case that has practically no air flow (32C)

<b>Qui habet aures audiendi audiat</b><P ID="edit"><FONT SIZE=-1><EM>Edited by markgun on 12/10/03 11:10 AM.</EM></FONT></P>

I know! It happens to me the first time. From that time I always write on an excel ALL that is on the CPU (even the most stupid mark) in case I need it later!

I was thinking that you can ask (maybe new thread) for a substitute for standard thermal compound ... it's not fair you can't go further due to cooling (IMO)

Good luck!


Still looking for a <b>good online retailer</b> in Spain