Athlon64 X2 3800+ Feedback need I WANT one but is it good ??

Actually the problem was that AMD was using defective toledo cores
which cost the same to produce as non-defective cores. The Manchester
was designed from the ground up to have 512KB L2 cache so it actually
does cost less to produce.

http://www.anandtech.com/cpuchipsets...oc.aspx?i=2484

On Tue, 02 Aug 2005 07:18:48 +0000, Wes Newell wrote:

On Mon, 01 Aug 2005 15:57:27 -0400, General Schvantzkoph wrote:

Don't count on overclocking X2s yet, they are brand new and there isn't a
lot of margin in AMD's process. I have a 4400+, it's not completely stable

Welll, the 4400+ is already at 2.4 GHz. The X2 3800+ is only at 2.0 GHz.
Shouldn't have any problem overclocking it to 2.4GHz.

The 4400+ is 2.2GHz, the 4600+ and 4800+ are at 2.4GHz. Look at the prices
of the 4600+ and 4800+, it's a clear indicator that they have almost no
yield at those clock speeds.

On Tue, 02 Aug 2005 09:13:09 -0400, General Schvantzkoph wrote:

On Tue, 02 Aug 2005 07:18:48 +0000, Wes Newell wrote:

On Mon, 01 Aug 2005 15:57:27 -0400, General Schvantzkoph wrote:

Don't count on overclocking X2s yet, they are brand new and there isn't a
lot of margin in AMD's process. I have a 4400+, it's not completely stable

Welll, the 4400+ is already at 2.4 GHz. The X2 3800+ is only at 2.0 GHz.
Shouldn't have any problem overclocking it to 2.4GHz.

The 4400+ is 2.2GHz, the 4600+ and 4800+ are at 2.4GHz. Look at the prices
of the 4600+ and 4800+, it's a clear indicator that they have almost no
yield at those clock speeds.

You're correct. I got the 4400+ and the 4600+ mixed up. Hopefully, the
newer manchester cores will go higher.

--
KT133 MB, CPU @2400MHz (24x100): SIS755 MB CPU @2330MHz (10x233)
Need good help? Provide all system info with question.
My server http://wesnewell.no-ip.com/cpu.php
Verizon server http://mysite.verizon.net/res0exft/cpu.htm

On Mon, 01 Aug 2005 23:09:39 -0400, General Schvantzkoph
wrote:


On Mon, 01 Aug 2005 15:57:27 -0400, General Schvantzkoph
wrote:

Ran Memtest86 for 10 hours, ran fine. BTW Memtest86 changes the the clock
rate to the auto rate of 167MHz.

First off, make sure you are running Memtest86+, the PLUS version, not
the original Memtest86.

Go he http://www.memtest.org/ and pick up the latest version 1.60

Second, Memtest does not change any memory or clock settings, it just
reports what it sees. Again, make sure you run the latest version.

Steve

On Tue, 02 Aug 2005 12:16:01 -0500, Steve wrote:

On Mon, 01 Aug 2005 23:09:39 -0400, General Schvantzkoph
wrote:


On Mon, 01 Aug 2005 15:57:27 -0400, General Schvantzkoph
wrote:

Ran Memtest86 for 10 hours, ran fine. BTW Memtest86 changes the the clock
rate to the auto rate of 167MHz.

First off, make sure you are running Memtest86+, the PLUS version, not
the original Memtest86.

Go he http://www.memtest.org/ and pick up the latest version 1.60

Second, Memtest does not change any memory or clock settings, it just
reports what it sees. Again, make sure you run the latest version.

Steve

I ran Memtest86 v3.2 which is the latest version on the
http://www.memtest86.com. Thanks for the pointer to Memtest86+. Memtest86
does change the clock speed, it doesn't give any indication that it's done
it but does. I had done a bunch of benchmarks on this system using 4G of
OCZ memory. It was able to run at 200MHz with the OCZ memory except for
the occasional kernel oops so I was able to get benchmarks at 133MHz,
167MHz and 200MHz. I then swapped 2G of the OCZ memory for 1G of Crucial
which seems to be a less stable combination. With the 1G of Crucial it
wouldn't boot at 200MHz so I set the DDR clock to 200Mhz, the CL to 3
and then put in Memtest86 and ran it for 10 hours. Memtest86 (the version
from http://www.memtest86.com), found no errors. I exited out of Memtest86
and booted right into the OS. I then reran my benchmark and got the 167MHz
performance number. I then rebooted in to the BIOS to check the setting,
it said 200MHz. I then reboot the OS and got a kernel panic, tried several
times with the same result. I then booted into Memtest86 and from there
boot the OS, no problems. I repeated several times, always got the same
result, booting directly got a kernel panic, booting from Memtest86
didn't. When booting for Memtest86 the benchmark number was exactly the
167MHz number, from this I concluded that Memtest86 sets the DDR clock
speed to the AUTO value which for a dual double sided memory system is
167MHz.
I'll give Memtest86+ a try, from their website it looks like they might be
better behaved.

General Schvantzkoph wrote in
news
Don't count on overclocking X2s yet, they are brand new and there
isn't a lot of margin in AMD's process. I have a 4400+, it's not
completely stable even at the normal clock rates. I tried a 5%
overclock and things seemed to work, then I got ambitious and tried a
15% overclock (the air conditioner in my server room was cranked up so
this should have been possible). With 15% the system was dead, no BIOS
screen or anything.


If you are interested in overclocking, have you thought about the DFI
board? Wesley Fink at Anandtech wrote that the 754 and 939 versions of
this board are the best OC boards he has ever seen (I think Wesley is one
of more respected mobo reviewers). I have the 754 version. I have not
overclocked it yet but it is exceptionally stable (knock on wood ).

I also use OCZ Platinum Rev. 2 memory sticks.

Good luck.

On Tue, 02 Aug 2005 07:25:13 GMT, Wes Newell
wrote:

On Mon, 01 Aug 2005 21:21:03 -0500, Steve wrote:

On Mon, 01 Aug 2005 19:05:37 GMT, Wes Newell
wrote:

On Tue, 02 Aug 2005 01:00:38 +1000, No One Realy wrote:

Whats the difference to the 4200. Which was the one i wanted but to
expensive. Why is the 3800 X2 less expensive ?

It's less expensive because AMD wanted it to be. The difference is the
3800+ X2 has a 10 multiplier for 2000MHz, and the 4200+ X2 has an 11
multiplier for 2200MHz. Even though they're both capable of going higher,
this has been the way CPU's have always been marketed. The actual
manufacturing cost of all of them are basically the same.

Actually the 3800+ X2 is a new core and is only 147mm2 vs 199mm2 for
the 4200+ X2. Thus it is less costly to manufacture.

And you think that only the 3800+ X2 is going to use the new manchester
core? :-) IOW's, the core doesn't matter in cost comparisons because the
other X2's will also use the manchester core.

That is incorrect. Currently the only X2 processor with a manchester
core is the 3800+ X2. The 4200+ X2 and the 4600+ X2 have half of
their cache disabled. In the future the 4200+ and the 4600+ will also
use the manchester core.

Steve