Here's a little teaser....

Printable View

Show 100 post(s) from this thread on one page

10-04-2007, 08:32 AM
lukija

Hm, if you can, please do a register dump in CPU-Z (About -> Register Dump) for this revision and previous revision and attach both txt files. This should help, since CPU-Z and other similar tools analyze cpuid intruction results and then decide what revision it is.
10-04-2007, 09:06 AM
KTE

Bugger! I've seen B0 and B2, and they were very clearly marked as that. AMD says you should not be concerned with B1's since they are only ES, and were bugged, but were not any release or retail samples. I'm 100% it's only BA shipping, but I don't remember seeing a CPU-Z ss for it. You can check online if you have a BA

Comes back to the first round of comments: why did they give out B1s that have performance penalties for testing if BA is release!?! :shakes:

As far as CPU-Z goes, it recognizes B0, B1, B2 properly, so I'm not sure why it would not recognize BA properly either. It should just read the core register values for family, model, stepping. Since those have the same values, they really do look like B1s going off that.

These are the only two 2350 chips available and they're BA

http://products.amd.com/en-us/Optero...il.aspx?id=234
http://products.amd.com/en-us/Optero...il.aspx?id=235

OPN Tray: OS2350WAL4BGE
OPN PIB: OS2350WAL4BGEWOF
10-04-2007, 09:16 AM
MAS

Quote:

Originally Posted by s7e9h3n

.maybe there really isn't a difference between BA/B!?

cpu-z can't show you "sub-revision"

in fact there are only two steppings of K10 - B1 and B2

B1 consists of B1 (engineering samples) and BA (Opterons at 1,9 and 2 GHz). B1 and BA obviously are very similar, BA is just "more advanced" and latest version of B1 revision

B2 is high-clock revision (they promise up to 3GHz)

S7, you should show us something new. Remove one of the cpus and run 3dmark on single-socket configuration (let the bios stay old, that doesn't matter)
10-04-2007, 09:48 AM
KTE

BIOS and software developer guides stated F-2-A was core rev BA and F-2-2 is core rev B2 (ATs 2.5GHz B2), and we already knew F-2-1 is core B1.

s7: are your chips 0S2350WAL4BGC or 0S2350WAL4BGE (OPN)?
10-04-2007, 09:49 AM
s7e9h3n

Quote:

Originally Posted by MAS

S7, you should show us something new. Remove one of the cpus and run 3dmark on single-socket configuration (let the bios stay old, that doesn't matter)

Your wish is my command.....:p:

http://img107.imageshack.us/img107/1...061cpuaec0.jpg

http://img107.imageshack.us/img107/4...061cpubeq5.jpg
10-04-2007, 09:51 AM
s7e9h3n

Quote:

Originally Posted by KTE

BIOS and software developer guides stated F-2-A was core rev BA and F-2-2 is core rev B2 (ATs 2.5GHz B2), and we already knew F-2-1 is core B1.

s7: are your chips 0S2350WAL4BGC or 0S2350WAL4BGE (OPN)?

They are in fact "C"'s....I haven't checked until now...that's what I get for believing the good 'ol boys @ AMD :rolleyes: I'm working on gettin replacements as we speak.....;)
10-04-2007, 09:57 AM
Thesavage

Quote:

Originally Posted by s7e9h3n

They are in fact "C"'s....I haven't checked until now...that's what I get for believing the good 'ol boys @ AMD :rolleyes: I'm working on gettin replacements as we speak.....;)

Any truth to the rumour that there b3's arround? :cool:
10-04-2007, 10:04 AM
s7e9h3n

Quote:

Originally Posted by Thesavage

Any truth to the rumour that there b3's arround? :cool:

None that I've seen......yet ;)
10-04-2007, 10:37 AM
MAS

thanx
s7e9h3n, can you adjust HT via bios setup?
10-04-2007, 10:39 AM
s7e9h3n

Quote:

Originally Posted by MAS

thanx
s7e9h3n, can you adjust HT via bios setup?

Yep...but it doesn't seem to make a difference in any benchmarks.....
10-04-2007, 10:58 AM
MAS

dissapointing result =(
where's K10's power fpu?
10-04-2007, 11:04 AM
s7e9h3n

Quote:

Originally Posted by MAS

dissapointing result =(
where's K10's power fpu?

Stuck in single channel, low-fsb, non-ocing limbo.....
10-04-2007, 11:20 AM
s7e9h3n

Quote:

Originally Posted by MR_SmartAss

s7, you still haven't fixed the issues with your motherboard?

I thought you've done this benchmark on the QuadFX mobo... For a reference Q6600 will score around 3K. So this score is not bad, but we'll need some real-life benchmarks on an non-crippled mainboard before we can make any solid conclusion.

I 100% agree, only problem is that the first bios bricked my board but I was able to save it with an external eeprom programmer. I'm somewhat reluctant to try the new bios until I get confirmation about a specific question I had. Hopefully it won't be too long....;)
10-04-2007, 11:23 AM
MAS

Quote:

Originally Posted by MR_SmartAss

For a reference Q6600 will score around 3K.

3400 to be exact

quad-core opteron at 2,4GHz with single-channel buffered 667 MHz ram: 2907/1995*2400 = 3497 (real number - 3300 rather)

Phenom X4 at 2,4GHz with dual-channel low-latency 1066MHz memory - ?
10-04-2007, 11:29 AM
Sparky

Quote:

Originally Posted by MAS

3400 to be exact

OK so 3400 Q6600 at stock speed of 2.4GHz. This K10 is 2900 at 2GHz. And from the sound of it still B1, though I do wonder how much BA will make a difference.

Doesn't sound too bad but not a killer either. Hmm...
10-04-2007, 11:34 AM
s7e9h3n

Quote:

Originally Posted by SparkyJJO

OK so 3400 Q6600 at stock speed of 2.4GHz. This K10 is 2900 at 2GHz. And from the sound of it still B1, though I do wonder how much BA will make a difference.

Doesn't sound too bad but not a killer either. Hmm...

I'm still hearing that BA steppings are NOT going to bench much different/if at all than B1....I dunno....:shrug:
10-04-2007, 11:45 AM
Thesavage

Do you know if there will be any difference moving to ddr2 800 over ddr2 667? that is to say performance wise.
10-04-2007, 11:49 AM
s7e9h3n

Quote:

Originally Posted by Thesavage

Do you know if there will be any difference moving to ddr2 800 over ddr2 667? that is to say performance wise.

Of course there'll be a difference, but the biggest difference will show once NUMA is corrected on these rigs. Take, for example, what dual-dual rigs can currently get in bandwidth and then more triple that and you'll see what an octal AMD dual socket setup is truly capable of.....
10-04-2007, 11:50 AM
KTE

Quote:

Originally Posted by s7e9h3n

They are in fact "C"'s....I haven't checked until now...that's what I get for believing the good 'ol boys @ AMD :rolleyes: I'm working on gettin replacements as we speak.....;)

Gack, that's too bad. Those are B1s like the ones sent out for reviewing. :(

Thanks for taking all the hassle and feedback. Lets hope they get fed up and now send you a B2. :D
10-04-2007, 12:07 PM
s7e9h3n

Quote:

Originally Posted by KTE

Gack, that's too bad. Those are B1s like the ones sent out for reviewing. :(

Thanks for taking all the hassle and feedback. Lets hope they get fed up and now send you a B2. :D

b2's are coming with phenom....;)
10-04-2007, 12:08 PM
accord99

Quote:

Originally Posted by MAS

Phenom X4 at 2,4GHz with dual-channel low-latency 1066MHz memory - ?

Maybe 3% higher, 3DMark06 CPU has shown itself to be insensitive to memory bandwidth and latency.
10-04-2007, 12:21 PM
BadNizze

Wonder if BA clocks better than B1?
You don´t happen to know when Am2+ board will appear in retail do you? Wonder if AMD has begun producing Am2+ quad´s yet or if they only make Opterons...

Show 100 post(s) from this thread on one page