Here's a little teaser....

**s7e9h3n** · 10-04-2007, 01:37 PM

Originally Posted by railmeat

agree plus on air\water too helps alot and not 24\7 worthy o\c on ln2\dice..ty
like to see the relaistic clockable points of these chips for the masses.

I could see how this server will perform with both cpu's @ -100C on my dual headed autocascade, but that wouldn't be too much fun @ default clocks

**JohannesRS** · 10-04-2007, 01:51 PM

Originally Posted by s7e9h3n

Also, I'd like to state for the record: Until we see ANY concrete results from the "BA" stepping Barcelonas, I'll still stick with what I've been told and assume that these B1's will offer the exact same performance in applications across the board.....(with the possible exception of a couple of VERY specific benches).......

This means that you and Dave Graham completelly disagree each other and go out to settle things?

:P

Just kidding. So, from what we can see, seems that you possibly still have B1s, and so no-one can know about any performance difference, right?

What's even worst: still, no dual-channel, neither NUMA?

At least, the performance bugs can be resumed to this two points?

@Dave: Dave, could you get any news from the split plane and dual-HT boards?

**s7e9h3n** · 10-04-2007, 01:54 PM

Originally Posted by JohannesRS

This means that you and Dave Graham completelly disagree each other and go out to settle things?

:P

Just kidding. So, from what we can see, seems that you possibly still have B1s, and so no-one can know about any performance difference, right?

What's even worst: still, no dual-channel, neither NUMA?

At least, the performance bugs can be resumed to this two points?

@Dave: Dave, could you get any news from the split plane and dual-HT boards?

All I'm saying is that until somebody shows some solid evidence that the "BA" cpu's will/do outperform the B1's, what evidence is there to base ANY performance numbers on?

**JohannesRS** · 10-04-2007, 02:01 PM

I know, I know. Just kidding, could not avoid it when I saw your post.

But again: now, we are still and "only" locked into NUMA and dual-channel bios problems, is that right?

**s7e9h3n** · 10-04-2007, 02:21 PM

hOLY $HIT...I think I may have it....I got my modded bios to flash on my Asus and it actually booted....time to install the quads....

**informal** · 10-04-2007, 02:34 PM

hehe finally some action

.Good luck!

**s7e9h3n** · 10-04-2007, 02:46 PM

Originally Posted by informal

hehe finally some action

.Good luck!

ok...looks like I may need to do some work to get this thing to run....so consider this a false/early alarm...sorry...for now

**s7e9h3n** · 10-04-2007, 02:56 PM

Oh, and just in case I forgot to mention, a BIG

to Asus [non] support for totally dropping the ball on this bios build.....

**flippin_waffles** · 10-04-2007, 03:06 PM

OMG, this thread has more ups and downs than a whores drawers!!

But we're a patient bunch, thanks for the work thus far!

**tiro_uspsss** · 10-04-2007, 04:00 PM

Originally Posted by flippin_waffles

OMG, this thread has more ups and downs than a whores drawers!!

But we're a patient bunch, thanks for the work thus far!

**Slay0r** · 10-04-2007, 04:33 PM

Thanks for the report s7e9h3n

im tuned to see what these can do when they're 100% fit

**tictac** · 10-04-2007, 10:03 PM

Originally Posted by s7e9h3n

Oh, and just in case I forgot to mention, a BIG

to Asus [non] support for totally dropping the ball on this bios build.....

.... ASUS need to fixed that....

**KTE** · 10-05-2007, 03:18 AM

Originally Posted by s7e9h3n

Also, I'd like to state for the record: Until we see ANY concrete results from the "BA" stepping Barcelonas, I'll still stick with what I've been told and assume that these B1's will offer the exact same performance in applications across the board.....(with the possible exception of a couple of VERY specific benches).......

Is this what AMDs been telling you?

I'd like to know bcause my uncles been working with AMD negotiating on behalf of his firm for quite a while, in particular for mass orders of Barcelona and they've repeatedly insisted to him quite the opposite (they had B1s for testing too and were not impressed). Only Friday was the last talk they had, and the guy working with him is manager of business development.

**linhvndiy** · 10-05-2007, 07:42 AM

s7e9h3n:
Thanks a lot for you info. Please test with LN164 as soon as you can.

And can you send me the bios(L1N) to vndiy@yahoo.com?

**Rickster_64** · 10-05-2007, 09:57 AM

This thread rocks haha. Talk about a big PITA to get those things working. It's almost like no ones had a K10 core in hand until 2 weeks B4 retail.

**hargen** · 10-05-2007, 07:05 PM

Originally Posted by s7e9h3n

Oh, and just in case I forgot to mention, a BIG

to Asus [non] support for totally dropping the ball on this bios build.....

lol. Now, I can talk with you in this thread, and my account is activated as hargen. I can upload the test file here. In this test. Whether B1 have some performance bugs, such as lack of out-of-order loading or lack of two 128-bit sse loads/cycle?

The source file and executable files are include in the following zip file.

**zpdixon** · 10-05-2007, 08:46 PM

s7e9h3n,

If you have the occasion to boot Linux, can you run a quick "openssl speed" benchmark ? Or anyone else ? I am very interested into how Barcelona performs on cryptographic algorithms and hash functions.

- Z

**Lightman** · 10-06-2007, 12:04 AM

Originally Posted by hargen

lol. Now, I can talk with you in this thread, and my account is activated as hargen. I can upload the test file here. In this test. Whether B1 have some performance bugs, such as lack of out-of-order loading or lack of two 128-bit sse loads/cycle?

The source file and executable files are include in the following zip file.

Welcome to XS!

I tried your util on my X2 F3 and it says it's not a Barcelona

!

Great little test though! I like it!

**informal** · 10-06-2007, 12:26 AM

Originally Posted by zpdixon

s7e9h3n,

If you have the occasion to boot Linux, can you run a quick "openssl speed" benchmark ? Or anyone else ? I am very interested into how Barcelona performs on cryptographic algorithms and hash functions.

- Z

Here you go,Openssl tested on Barcelona:
http://www.tecchannel.de/server/proz...28/index9.html

It rocks in this test.

**zpdixon** · 10-06-2007, 12:45 AM

Informal: I should have mentioned that I was aware of this bench... The pb is that they just tested RSA with 1 key size, while I am looking for complete "openssl speed" results (representing the 30 or so combinations of algorithms and key sizes.)

**s7e9h3n** · 10-06-2007, 01:04 PM

Originally Posted by hargen

lol. Now, I can talk with you in this thread, and my account is activated as hargen. I can upload the test file here. In this test. Whether B1 have some performance bugs, such as lack of out-of-order loading or lack of two 128-bit sse loads/cycle?

The source file and executable files are include in the following zip file.

**BrowncoatGR** · 10-06-2007, 01:27 PM

Originally Posted by s7e9h3n

I dunno if im misinterpreting the numbers but those results seem slow to me

**xPliziT** · 10-06-2007, 02:14 PM

I think it does not that bad.

This is from a Athlon 64 @ 3500 MHz.
Especially the sse load operation seems to be much much faster than the Athlon 64's. 2 loads/cycle vs 0.5 loads/cycle....

**~~DeepThought86~~** · 10-06-2007, 08:30 PM

Originally Posted by xPliziT

I think it does not that bad.

This is from a Athlon 64 @ 3500 MHz.
Especially the sse load operation seems to be much much faster than the Athlon 64's. 2 loads/cycle vs 0.5 loads/cycle....

That's FANTASTIC. This is what I get on a Sempron 3000+

:

**Carfax** · 10-06-2007, 08:36 PM

Those scores seem accurate for the K10, as it can do 2 128-bit loads per cycle.

The Core 2 can only do 1 load per cycle I believe.

Thread: Here's a little teaser....

Thread Tools

Search Thread

Rate This Thread

Display

Barcelona OpenSSL benchmarks

Bookmarks

Bookmarks

Posting Permissions