#0, Cake , le 23/05/07 8:29 AM

Can you ask AMD what the hell is going on?

[UberPulse response]
Just did! Read next post!
JB

#1, mac , le 24/05/07 4:14 AM

OMG, my power5, does 16000
pixels per second... I forgot
to mention I set the raytracing
depth to 0... how can you
compare test that you don\'t
know how they were performed,
thats unprofessional

#2, mark , le 24/05/07 6:38 PM

ummm what is this article saying really? because last time i checked the Barcelona core hasn\'t been tested by anyone; even AMD hasn\'t released numbers yet. I\'m just not seeing where the hell Intel got these numbers. AMD wasn\'t even at the Microprocessor forum and I highly doubt that they would loan Intel 4 Barcelona processors.

#3, JAT , le 25/05/07 6:20 AM

Penryn may be 40% faster if SSE4 software is used. SSE4-support will be found on all major applications when Penryn is launched. Fer sure.

"If you take 16 cores at 1.8Ghz, divided by 2 and add 60% to get to 3.0Ghz, it gives POV-Ray score of 3600 for 8 cores."

Where we need banchmarks, while we can CALCULATE processor speed?

#4, Max , le 25/05/07 6:36 AM

Amd responded and aperently intel crippled the system with 6gb of ram

#5, BULLCRAP , le 25/05/07 12:02 PM

how did intel get its hands on barcelona? It isnt even out yet. WTF was actually in the amd box? because intel is know for its fed up marketing with testing older crap from amd to newer stuff from their company which is a load of BULLCRAP

#6, fastpunk , le 25/05/07 4:51 PM

I wouldn\'t worry about it. Intel is not exactly the most reliable source.

#7, denis , le 26/05/07 4:11 AM

AMD said the barcelona chip will outperform intels Cloverton chip by 50% fp and 20% int , clock for clock !So here goes .....
3 ghz div by 1.8 = 1.666 recuring
about 2000 pixels X 1.666 = 3333.333 pixels per sec recuring
NOW ! a 50 % increase on 3333.3333 = 5000
Thats with AMD\'s prediction/assumption on its Barcelona chips performance over cloverton and doesn\'t take into account an increase in performance of Barcelonas memory subsytem vs fb-dimms.
Also do the same calculations for 20 % int and the result YOU WILL FIND is somewhere between 20 and 50 % over the dual cloverton score.

#8, Cheetah , le 26/05/07 7:14 AM

The author of this article is a complete .

How about you compare results on actual working AMD silicone when Barcelona/Agena are released?

dumb man really dumb...

#9, frew , le 26/05/07 8:24 AM

JAT: Barcelona also has SSE4a.

#10, JAT , le 27/05/07 10:02 AM

frew: I do not think that first versions will have, the ost likely have SSSE3, not SSE4.

#11, JAT , le 28/05/07 3:53 AM

AMD is faster do proper benchies and not 1 bench wiggaz

#12, Ivan , le 30/05/07 1:37 PM

Some math for you guys:
8 K8 cores scored 2200 in 3.7 beta.
From the FX74 SSE2 version of the 3.7 beta (4 cores total) which is 2200,we can conclude that Opterons were running at 1.8Ghz and doing the X87 LEGACY TEST! NOT THE SSE2 optimized (gives 17% better score than legacy,which fits perfectly with 1.8Ghz Opteron scores which should be ~2 times lower than 8 core 3Ghz hypotetical FX74)

Next:
16 K10 1.8Ghz cores scored 4100 PPS in same x87 legacy benchmark.

8 K10 1.8GHz cores would get around the same scores as Opterons ,\'cos scaling is not perfect(1.85%).So we can assume not worse than 2200.

Now the interesting part.
Scale to 3Ghz the K10s,we get ~3660 with LEGACY x87 test!

Now more interesting:SSE2 boost with K10.AMD said it would be 80% compared to K8(per core of course,so 360% per quad core VS Dual Core).

Now we have : scale K8s to 3Ghz~3600;boost for SSE2 is around 17% in K8 case for PovRay SSE2 bench compared to x87 Legacy ,lets take LESS, 15% for instance ---> we get : 3600x1.15=4140.

Now we calculate for the SSE boost Barcelona will bring relative to this K8 score, with improvements to SSE engine(80% better than K8 per core).

We get :4140x1.8=7452 PPS ! Let\'s assume 80% is a peak,and we get more realistic 60%.

4140x1.6=6630.Compare this projected SSE2 K10 score with Intel V8 3Ghz(4933):
Best case scenario for AMD: 7450/4933=1.51,so 51% faster K10 than Clowerton at the same clock(3Ghz).

Worst case for AMD(best for Intel) : 6630/4933=1.35,so 35 % is K10 faster in PoVRay at the same clock(3Ghz).
There you have it,no BS from intel fans,K10 WILL blow away Clowerton at the same clock,just as AMD said it(50%-max in fp workloads,20% in integer).

#13, Caleb , le 01/06/07 8:34 PM

this is dumb....they probably used 4 penryn chips compared to 8 penryn chips or some ish. Intel\'s propaganda has to stop and that company needs to get a grip on reality and stop worrying about AMD this AMD that. They should concentrate on the consumer for God sakes instead of trying to release second rate chips for bragging rights. IMO this whole article is pathetic.

#14, Matias , le 04/06/07 9:28 AM

What can tell us that old man?

All sucks!
give me the AMD, intel Sucks because they have MONEY! also they can\'t compete with AMD.

#15, who? , le 26/12/07 9:49 AM

now, we all know that the video was right.

who?