Barcelona Prediction Thread

MDme

Senior member
Aug 27, 2004
297
0
0
Okay, there have been benchies out there but no one really knows how this will play out. There has also been a big discussion on the scaling issue. BUT anyway, why don't we post our estimates and see who can come closest to the real thing.

We'll divide the estimates based on the following:

GAMING
GENERAL SYSTEM PERFORMANCE
ENCODING/MEDIA
ENCRYPTION
RENDERING

now since this is just for fun, let's not make it over complicated. just quote a +/- % vs conroe on each of these items (with +5% meaning barcy will be 5% faster than conroe).

Let the games begin.....

BTW....no flaming, just plain guestimates. :)

Let's go back to this when barcey benches are out....we'll use AT Benches then for reference.
 

Pabster

Lifer
Apr 15, 2001
16,987
1
0
I'm predicting anywhere from 5 to 15% slower than Conroe.

Please keep your receipt. :laugh:
 

SickBeast

Lifer
Jul 21, 2000
14,377
19
81
Is this comparison clock-for-clock? If it is, I'm gonna guess that it's at least 20% faster across the board, more or less.

As a workstation CPU my guess is that it will suck until there is a good platform for it. The same goes for gaming, and all of the above-mentionned categories except for encryption and perhaps general system performance.

As an upgrade CPU for a server, it will be incredible. It will have even better performance-per-watt than K8 and Conroe. It will be an easy and relatively cheap no-brainer upgrade for server admins.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
clock for clock, 25% faster than Conroe and 20% faster than Penryn...

BTW, we should look for the reviews that compare the 2 GHz Barcelona to the E5335 Xeon (1333 FSB, 2 GHz)...I do hope that Intel gets one to Anand for the review!
 

Borealis7

Platinum Member
Oct 19, 2006
2,914
205
106
here's my prediction: im not going to buy one no matter how it performs!

Nothing against AMD, but i just dont need a new cpu+mobo right now.

And I believe many people are on the same boat as I am having bought a Q6600 or a 1333MHz CPU and P35 mobos. Its simply too late.
 

Idontcare

Elite Member
Oct 10, 1999
21,118
58
91
Single-thread apps -> Barcelona is 5-10% faster clock-for-clock than Clovertown at 85% the power consumption at full single-thread load (system level at the wall)

Mutli-thread apps -> Barcelona is 15-20% faster clock-for-clock than Clovertown at 80% the power consumption at full multi-thread load (system level at the wall)

The 45nm upgrade to Yorkfield for Intel will eliminate the power consumption delta at load between K10 and Xeon systems at the wall, and will reduce the performance gap by 5% per-clock across the board.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Originally posted by: Viditor
clock for clock, 25% faster than Conroe and 20% faster than Penryn...

BTW, we should look for the reviews that compare the 2 GHz Barcelona to the E5335 Xeon (1333 FSB, 2 GHz)...I do hope that Intel gets one to Anand for the review!

Your good . The highest possiable figure for amd who is on 3rd stepping B01. The one tested that came out day after 30,000 3dmark lol.

Than for penryn a totally new process never before produced for the market on a chipset that Intel says its really not intended for.X38 being the correct chipset . On A0 stepping lol. This really doesn't apply to AMD as their fanbois and AMD themselves have stated all along the K10 will work great on the present sockets a chipset . Is that not so? So Intel penryn on A0 stepping using a chipset not really intended to be used with it but can be. Only is 5% faster than conroe. LOL

Here my prediction Conroe 10% faster in most apps than K10. Penryn 25-30% faster than K10 desktop part. @=clocks . Now AMDs top k10 part for desk top . Vs Intels penryns top clocked part . I give Penryn 40-50% performance advantage . Even tho the clocks are NOT=. Than when we talk O/C its a pure as-s beating of k10 by penryn.

I am sure that intel will supply a 50 watt c2d or maybe even a 35watt penryn to test against K10. Or do you want it all intel xeon with fbdimms vs K10. But that really wouldn't be fair now would it . You would want the exact same memory used wouldn't you? you know apples to apples? I am thinking you would like to pick the bench marks also . Ya that would be fair. No it wouldn't fair would be the best amd has to offer against the best intel will have to offer. At same clock .

So it should be AMD with the best DDR2 available @ DDR2 800 vs Intels best with x38 chipset against Intel DDR3 1333. Using the certified memory for intels X38 chipset. That would be real world.

But we know it won't be done this way, The reviewers to scew results will use AMDs best @ DDR2800 Vs Intel using DDR2 800. In the guis of apple to apple compare. Which isn't real world. Fact is if intels X38 chipset has settings for DDR3 1600 thats what should be used . but that won't happen.

You wouldn't want intel to run at a bandwidth disadvantage would ya. All this can be done without O/C anything. Just pick DDR3 800 with a 1;2 ratio presto. Amd can use 1066 if the reviewers so choose . But nothing can be overclocked.



 

Phynaz

Lifer
Mar 13, 2006
10,140
819
126
Originally posted by: Viditor

BTW, we should look for the reviews that compare the 2 GHz Barcelona to the E5335 Xeon (1333 FSB, 2 GHz)...I do hope that Intel gets one to Anand for the review!


Why?

We should compare AMD's best shipping cpu to Intel's best shipping cpu.

This whole thing about getting a "win" by having a higher ipc is a non-starter.
 

biostud

Lifer
Feb 27, 2003
18,193
4,674
136
http://www.vnunet.com/tv/?chan...lipid=1386_vnunet_0096

If you don't want to watch the video the scores revealed in the video is:
A 2.0Ghz Barcelona scores an average of SPECfp 78 vs a 2.33GHz Clovertown SPECfp 60, and the AMD system using ~330W while the Intel system use ~360W, giving the AMD system ~25% advantage in raw power and ~30% in Performance-per-watt.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Originally posted by: biostud
http://www.vnunet.com/tv/?chan...lipid=1386_vnunet_0096

If you don't want to watch the video the scores revealed in the video is:
A 2.0Ghz Barcelona scores an average of SPECfp 78 vs a 2.33GHz Clovertown SPECfp 60, and the AMD system using ~330W while the Intel system use ~360W, giving the AMD system ~25% advantage in raw power and ~30% in Performance-per-watt.



LOL LOL LOL LOl . I see you picked your benchmark . LOL LOL . I didn't bother watching I have heard about it for a year . Intel now has a 50 watt part out. You know. A presentation made by AMD. After the lies they have spewed over the last year only the purest of fan-boys would buy into that How many FB dimms did AMD have running in this video? LOL lOL.
How many of you guys buy servers. they would be the only ones interested in this BS. I bet Tigerton beats 4 socket AMD K10 hands down won't even be close. Dam I hope intel brings out that DDR3 server chipset soon . AMD is really using that memory against intel . But I am sure once Intel helps get it perfected AMD will jump in LOL.

You know I could always link to 4X4 vs. V8 than we could all have a good laugh.Of course to keep it fair we would use 2 c2d 6800 rather than the 6850. or we could for laughs do QC vs 4x4 . I mean if you don't want to do a apples to apples compare just come out and say it. You guys do know that intel has a Xeon cpu for 775 socket right . But na that wouldn't be fair would it?
 

nonameo

Diamond Member
Mar 13, 2006
5,949
3
76
I think that benchies will produce #'s that are confusing and don't make sense, with performance being rather uh, "random" across the board.

IOW, who the !@#% knows.
 

Keysplayr

Elite Member
Jan 16, 2003
21,209
50
91
If the Barc is much faster all around, only then would I be tempted to change my platform and depends on costs to do so. As it stands right now, my platform is ready to have a Penryn dropped in it with my current BIOS. So, it would take a mofo of a Barc to make me switch. I mean MOFO. I believe the Barc and Penryn will be competitive on a clock per clock basis. Trade blows. I'd say +/- 10%. Can go either way. IMHO.
 

SlowSpyder

Lifer
Jan 12, 2005
17,305
1,001
126
My guess (just a guess) is that it'll be comparable to Conroe in single threaded performance, both of them trading benchmarks but scale a bit better in the multi-threaded benchmarks. Just a guess.
 

Keysplayr

Elite Member
Jan 16, 2003
21,209
50
91
Originally posted by: ultra laser
Is there any word on how much heat these babies put out?

1.21 Jiggawatts last I heard. No seriously, I read they put out 90W. but don't quote me.
 

coldpower27

Golden Member
Jul 18, 2004
1,677
0
76
Originally posted by: ultra laser
Is there any word on how much heat these babies put out?

Depends, I believe since they use the existing Opteron TDP's your looking at 68W and 95W versions at first.
 

Zstream

Diamond Member
Oct 24, 2005
3,396
277
136
Originally posted by: keysplayr2003
Originally posted by: ultra laser
Is there any word on how much heat these babies put out?

1.21 Jiggawatts last I heard. No seriously, I read they put out 90W. but don't quote me.

Well if you compare a 90w amd vs Intel I am not quite positive they will generate the same heat. Correct me if I am wrong.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: coldpower27
Originally posted by: ultra laser
Is there any word on how much heat these babies put out?

Depends, I believe since they use the existing Opteron TDP's your looking at 68W and 95W versions at first.

Correct...there are 3 levels of quad core Barcelona TDP.

1. SP (Standard Power)...The SP (95W) Barcelona processors will launch on Sept. 10 at 2.0-GHz. They'll be released in high clock-speed versions in the fourth quarter of this year.

2. HE (Highly Efficient)...The 68-W, low power HE Barcelonas will launch on Sept. 10 at 1.9-GHz. They'll also ramp to high speed in Q4.

3. SE (Special Edition or High Performance)...The 120-W SE Barcelonas won't launch until the fourth quarter, but when they do, they'll ship at speeds of 2.3-GHz and higher