Overclocking Barton 1833 (2500+)

rstlne wrote:

"SysCold" wrote in message
...

Larry Gagnon wrote:


On Fri, 06 Aug 2004 17:19:04 +0200, SysCold wrote:




I'm now running stable on 220MHz FSB, that's 2.43GHz. I don't know if my
Asus PC Probe is lieing but it said: 33C Idle, 43C CPU Max. I touched
cooler near CPU and it wasn't hot, I'd say it's mild warm. Still without
changing voltage. Memory timings: 2.5-3-3-8. I'll try to get it even

lower.

I'll let you know when I reach limit.... if anyone is interested?...



Have you actually run Prime95 for a few hours to test if your computer
really is stable? You are very lucky if you don't have to raise the CPU
core voltage. Also your temps are very low - I mean how long have you

run

it at 100% load? Also, what is your room's ambient temperature?

Larry Gagnon, A+ certified tech.


I've tested it in simple program I wrote. It loops some arithmetical and
logical operations for half of hour taking 100% of CPU. When I wrote
above post room temp was 25.5C. But there is something strange I
noticed. There is the difference between ASUS PC Probe and BIOS
readings. BIOS shows about 5 to 7C more then PC Probe. But still, 50C is
OK temp for 25.5C ambient temp and overclocked processor. I don't know
if it is wise or will it damage my processor/MB but I'll rise FSB until
processor temp reaches 60C at full load for hour. MB temp is 25C in BIOS
and PC Probe, so I hope I don't have to worry about it.
System is now working at 230 FSB, 2.54GHz, still NO change in core
voltage. I left it on 'AUTO' and during all testings, CV was at 1.696. I
ran FarCry for few hours (until my last post) at 230 FSB without any
problems. Processor is now 31C Idle, 38C after hours of FarCry. Temp in
the room is now 23.1C.



Maxing a processor at 100% is not 100% load, U should use prime95 to be
honest (Not doubting your ability at making something that works, but
thinking that you might be better off using something "known" to stress the
cpu)

Asus probe is CRAP to be honest, Get MBM5 and read the temp from the CPU
directly. The value that Probe gives you is what's on the motherboard and I
can tell you that it might will def be 4/5c off at most times (during idle)
and you might find it shocking to see how high it does go. (I have seen it
probably 30 off due to a "under rated HSF")
so 60c is "okay" might mean ur running the cpu @ 90c..

oh.. I would say to go about this in reverse of the way I mention'd it.. get
MBM5 first ;P


Thanks for info... I got prime95 and MBM5. You were right, PC Probe was
lieing! I'm now running prime95 at 220 FSB, stable so far:
Test 1,2,3,4 1024K FFT length - Self-test 1024K passed. It is now
testing 896K. Temperature is stable 56-58C. No voltage change. Computer
is stable.

If it passes all tests, without hanging, then I suppose it is stable? Am
I right?
After testing it with prime95 I'll try to get the same temp with my
program. To see if there is any good of it. I haven't tested it with
MBM5 to see what temps I'll get. I'll let you know after testing...

Bill wrote:

In article , says...

Larry Gagnon wrote:


On Fri, 06 Aug 2004 17:19:04 +0200, SysCold wrote:




I'm now running stable on 220MHz FSB, that's 2.43GHz. I don't know if my
Asus PC Probe is lieing but it said: 33C Idle, 43C CPU Max. I touched
cooler near CPU and it wasn't hot, I'd say it's mild warm. Still without
changing voltage. Memory timings: 2.5-3-3-8. I'll try to get it even lower.
I'll let you know when I reach limit.... if anyone is interested?...



Have you actually run Prime95 for a few hours to test if your computer
really is stable? You are very lucky if you don't have to raise the CPU
core voltage. Also your temps are very low - I mean how long have you run
it at 100% load? Also, what is your room's ambient temperature?

Larry Gagnon, A+ certified tech.


I've tested it in simple program I wrote. It loops some arithmetical and
logical operations for half of hour taking 100% of CPU.


Does your program make full use of the Level 1 and Level 2 cache?
If not you're nowhere near generating the max heat possible to
generate on the cpu.

Bill

I really don't know, I'm not that good but I'll look in it. If I
generate same/almost same temp as prime95, does that mean prog is good?

Michael.Carlsson wrote:
SysCold wrote:
[...]
BTW, just to make su is Barton die temperature 90C?

Barton die temperature 85C?

Correct. For a desktop Barton/TBred, the maximum die temperature is 85 deg
C. Dekstop Palominos are 90 deg C, TBirds are 95 deg C. Mobile/SFF chips
have die temperatures all over the place, depending on which one you get.

--
Michael Brown
www.emboss.co.nz : OOS/RSI software and more
Add michael@ to emboss.co.nz - My inbox is always open

In article , says...
Bill wrote:


snip

I've tested it in simple program I wrote. It loops some arithmetical and
logical operations for half of hour taking 100% of CPU.


Does your program make full use of the Level 1 and Level 2 cache?
If not you're nowhere near generating the max heat possible to
generate on the cpu.

Bill

I really don't know, I'm not that good but I'll look in it. If I
generate same/almost same temp as prime95, does that mean prog is good?



For the purposes of heating up the processor it's ok, but if it's not
excersizing all the cache, how do you know if the cache is going to be
reliable under stress? My understanding of Prime 95 is that it heavily
stresses the cache as well as the processor, as well as main memory.

Bill

Bill wrote:

In article , says...

Bill wrote:



snip

I've tested it in simple program I wrote. It loops some arithmetical and
logical operations for half of hour taking 100% of CPU.


Does your program make full use of the Level 1 and Level 2 cache?
If not you're nowhere near generating the max heat possible to
generate on the cpu.

Bill

I really don't know, I'm not that good but I'll look in it. If I
generate same/almost same temp as prime95, does that mean prog is good?




For the purposes of heating up the processor it's ok, but if it's not
excersizing all the cache, how do you know if the cache is going to be
reliable under stress? My understanding of Prime 95 is that it heavily
stresses the cache as well as the processor, as well as main memory.

Bill


More a case of checking the CPU and Memory, by verifying the Data it
wrote, is the same it receives back. Just loading the CPU won't do
bugger all, unless your verifying something...

My 50c *8)
Minotaur

So at 1.75 volts and at 2.5ghz did i just get lucky LOL LOL LOL LOL
LOL

Iam going for 3ghz, just need that water cooling kit



On Sat, 7 Aug 2004 14:36:04 +1200, "Michael Brown"
wrote:

Michael.Carlsson wrote:
SysCold wrote:
[...]
BTW, just to make su is Barton die temperature 90C?

Barton die temperature 85C?

Correct. For a desktop Barton/TBred, the maximum die temperature is 85 deg
C. Dekstop Palominos are 90 deg C, TBirds are 95 deg C. Mobile/SFF chips
have die temperatures all over the place, depending on which one you get.

HELLO NURSE.