Welcome Guest ( Log In | Register )

6 Pages  1 2 3 > » Bottom

Outline · [ Standard ] · Linear+

 [WTF] barcelona delayed.

views
     
TSedwin3210
post Jun 8 2007, 10:30 AM, updated 19y ago

lll
*****
Senior Member
808 posts

Joined: Jan 2007
doh.gif doh.gif doh.gif doh.gif doh.gif doh.gif doh.gifx doh.gif doh.gif doh.gif doh.gif doh.gif doh.gif doh.gif doh.gif doh.gif doh.gif doh.gif



speechless

This post has been edited by edwin3210: Jun 8 2007, 10:30 AM
ikanayam
post Jun 8 2007, 10:45 AM

there are no pacts between fish and men
********
Senior Member
10,544 posts

Joined: Jan 2003
From: GMT +8:00

Am i the only one who is NOT surprised?
superpc
post Jun 8 2007, 10:52 AM

Owned by Doggies
*****
Senior Member
954 posts

Joined: Jul 2005
From: Anywhere!


they gonna lose to intel seriously
my beloved AMD tongue.gif jus reli hope they can do somthing to cheer up AMD fans
AceCombat
post Jun 8 2007, 11:18 AM


Group Icon
Elite
5,434 posts

Joined: Dec 2006


but they always delay it and makes no point....
really fedup...
charge-n-go
post Jun 8 2007, 11:20 AM

Look at all my stars!!
*******
Senior Member
4,060 posts

Joined: Jan 2003
From: Penang / PJ

QUOTE(ikanayam @ Jun 8 2007, 10:45 AM)
Am i the only one who is NOT surprised?
*
I am not surprised as well biggrin.gif
Judging from the poor clock scaling, i believe there is some flaws in their 65nm ramp up (although AMD keeps saying no problem). Besides, there might be due to low yield rate.
TSedwin3210
post Jun 8 2007, 11:21 AM

lll
*****
Senior Member
808 posts

Joined: Jan 2007
QUOTE(superpc @ Jun 8 2007, 10:52 AM)
they gonna lose to intel seriously
my beloved AMD tongue.gif jus reli hope they can do somthing to cheer up AMD fans
*
not rili, and pray that it wont happen. from the info that have been surfaced, barcelona will match intel current core processor. as for the claim by AMD that barcelona will outperform core by 20% or more, i seriously doubt that doh.gif . but barcelona power consumption and its feature may post a shock to us.
X.E.D
post Jun 8 2007, 11:58 AM

curmudgeonosorus emeritus
******
Senior Member
1,955 posts

Joined: Jan 2006
From: Llanfair­pwllgwyngyll­gogery­ch



ikanayam
post Jun 8 2007, 12:33 PM

there are no pacts between fish and men
********
Senior Member
10,544 posts

Joined: Jan 2003
From: GMT +8:00

QUOTE(X.E.D @ Jun 7 2007, 10:58 PM)
Wait. This was expected already? I thought it was yearend delivery for both serverside and consumer...

(IMO Penryn wouldn't come Q3, most probs mid-end of Q4. I'll wager my FX5200. tongue.gif)

Hopefully they'll redo some marchtecture magick (don't think it's too late- Core Duo to Core 2 Duo was fast) after getting stable silicon. Achieving both is a challenge though.
*
there's no way you can make substantial changes to your marchitecture in less than half a year, so don't count on it. Core2 was not started after Core was finished, else they would have never completed it so fast. Intel has more resources to have more concurrent development teams.
toughnut
post Jun 8 2007, 12:41 PM

Look at all my stars!!
*******
Senior Member
3,239 posts

Joined: Jun 2005
not suprised. more like expecting it tongue.gif
afterall, ATi has infected AMD with their delay sickness.


Added on June 8, 2007, 12:54 pmmy condolence for Barcelona notworthy.gif
http://www.gearlog.com/2007/06/amds_barcel...z_and_disap.php
Hope it's not true...

another one,
QUOTE
According to Dean McCarron of Mercury Research, if AMD-or any other card maker for that matter-doesn't have a product ready to ship by Q4 in a given year, there generally isn't huge rush to get it out the following quarter.

http://www.extremetech.com/article2/0,1697,2099613,00.asp

This post has been edited by toughnut: Jun 8 2007, 12:54 PM
PhenomFX
post Jun 8 2007, 01:39 PM

New Member
*
Junior Member
9 posts

Joined: Jun 2007
Terrible news... but did AMD said to reuters that we can expect a demo somewhere around july? btw, read this another malaysian news website regarding barcy updates, its been well circulated in many forums.. http://xpentor.com/mamboserver/index.php?o...id=516&Itemid=1

BTW, i'm a newbie to this forum, XD

SUSjoe_star
post Jun 8 2007, 04:31 PM

Serving the Servants
******
Senior Member
1,810 posts

Joined: Mar 2007
I dont know how they got themselves into such a mess doh.gif doh.gif doh.gif At 1st I thought Barcelona would actually be good but this confirms to me that its gonna be just another R600. And Penryn will be launching like what.....a month or 2 later? Anyway I hope they get their act together coz having an Intel-Nvidia monopoly would definitely harm us consumers in the end sad.gif

This post has been edited by joe_star: Jun 8 2007, 04:34 PM
charge-n-go
post Jun 12 2007, 09:03 AM

Look at all my stars!!
*******
Senior Member
4,060 posts

Joined: Jan 2003
From: Penang / PJ

Anandtech wrote up something. The poor performance might be due to immature BIOS implementations. My guess wasnt wrong too, the clock scaling is not up to expectation and AMD might delay Barcelona launch.
TSedwin3210
post Jun 12 2007, 09:08 AM

lll
*****
Senior Member
808 posts

Joined: Jan 2007
QUOTE(charge-n-go @ Jun 12 2007, 09:03 AM)
Anandtech wrote up something. The poor performance might be due to immature BIOS implementations. My guess wasnt wrong too, the clock scaling is not up to expectation and AMD might delay Barcelona launch.
*
here is the whole story
ikanayam
post Jun 12 2007, 09:29 AM

there are no pacts between fish and men
********
Senior Member
10,544 posts

Joined: Jan 2003
From: GMT +8:00

That anand article looks suspect to me... performance doesn't scale exponentially with clockspeed lol
toughnut
post Jun 12 2007, 09:37 AM

Look at all my stars!!
*******
Senior Member
3,239 posts

Joined: Jun 2005
if performance does scale exponentially with clockspeed, i will consider LN2. really!

anyway, this time DAMMIT seem like in big mess. everything seem like didn't work out for them. even accusing intel of being unfair sweat.gif when there is a fair competition anyway?

edit: removed s

This post has been edited by toughnut: Jun 12 2007, 09:39 AM
soulfly
post Jun 12 2007, 09:41 AM

revving towards 10,000 rpm
Group Icon
VIP
15,903 posts

Joined: Jan 2003
From: Miri



QUOTE(toughnut @ Jun 12 2007, 09:37 AM)
if performance does scale exponentially with clockspeed, i will consider LN2. really!

anyway, this time DAMMIT seem like in big mess. everything seem like didn't work out for them. even accusing intel of being unfair sweat.gif when there is a fair competition anyway?
i thought it was intel who accused amd for being unfair comparing barcelona with clovertown
charge-n-go
post Jun 12 2007, 09:43 AM

Look at all my stars!!
*******
Senior Member
4,060 posts

Joined: Jan 2003
From: Penang / PJ

QUOTE(ikanayam @ Jun 12 2007, 09:29 AM)
That anand article looks suspect to me... performance doesn't scale exponentially with clockspeed lol
*
I guess the performance lost might b due to poor memory utilization by the BIOS (max of DDR2-667 and timing might not b good). It is a miracle if BIOS could make the CPU performance scale exponentially with clock speed LOL.
ikanayam
post Jun 12 2007, 09:49 AM

there are no pacts between fish and men
********
Senior Member
10,544 posts

Joined: Jan 2003
From: GMT +8:00

Assuming performance is not where it should be, my guess is that some things are still not working right and they had to microcode the fixes which makes it really slow.
charge-n-go
post Jun 12 2007, 10:08 AM

Look at all my stars!!
*******
Senior Member
4,060 posts

Joined: Jan 2003
From: Penang / PJ

Maybe it is some pre-fetchers problem, since they are new.
Besides this, maybe the performance is not up to expectations in single thread applications due to small L2 caches and 3 issue core. Intel has an absolute advantage when it comes to cache and prefetching.
ikanayam
post Jun 12 2007, 10:16 AM

there are no pacts between fish and men
********
Senior Member
10,544 posts

Joined: Jan 2003
From: GMT +8:00

I'm surprised that they did not show any encoding benchmarks which should show the largest gains due to SSE128. Looking at die shots, the additional 64bit of the SSE/FP engine seems to be "tacked on" to the existing K8. Doesn't seem to have major changes to the layout per core. So that could be broken, and if the original SSE/FP engine is working they might just have made temporary fixes (to this and a few other things) via microcode, which is why performance isn't so hot, and they can't really show encoding benchmarks because it will look crappy.

Cinebench (which DT ran) is hardly a single threaded app. It scales well up to 16 cores or so.

6 Pages  1 2 3 > » Top
 

Change to:
| Lo-Fi Version
0.0205sec    1.41    5 queries    GZIP Disabled
Time is now: 20th December 2025 - 09:25 PM