Sign in with
Sign up | Sign in
Your question

Finally Anand benchmarks Amd's 65nm brisbanes.

Last response: in CPUs
Share
December 15, 2006 1:07:15 AM

Some slight performance issues there, 65nm is actually 5% slower in one of the tests.
December 15, 2006 1:16:09 AM

Well I think alot of people were hoping that the 65nm shrink would mean immediate higher OC potential. As of today the shrink just means Amd can tout lower power consumption at idle and reduced manufacturing cost. Not shabby but a bit late.
Related resources
December 15, 2006 1:22:15 AM

Quote:
Well I think alot of people were hoping that the 65nm shrink would mean immediate higher OC potential. As of today the shrink just means Amd can tout lower power consumption at idle and reduced manufacturing cost. Not shabby but a bit late.


It's a bit of a 'nothing' shrink really for both consumers and enthusiasts.

Consumers don't know nor care if they are running a 65nm or 90nm CPU. Enthusiasts are hardly gonna go 65nm K8 in the wake of Core2.

In the end this shrink mainly benefits AMD and not much else.
December 15, 2006 1:34:22 AM

I guess this was a "dumb" shrink after all. I really hope they do the right thing and drop those prices down a bit. As it stands today a 5000+ price wise is comparable to an E6600... that has to change. They should drop the 5000+ down to an E6400 to compete. Unfortunaly more and more people are OCing and that makes thing even thornier for Amd since "most" enthusiast can OC an E6300 to match an FX62. Luckly for them not alot of people know how to OC yet.
December 15, 2006 2:03:46 AM

They didn't OC'ed it like how the other guy OC'ed his brisbane in the other thread.. They just benchmarked it and check if the shrink made changes in the performance per watt ratio.. which it did.. but not enough to overpower c2d

But I am amazed how AMD is still able to squeeze in some little improvements and tweaks to their aging and beaten K8 architecture
December 15, 2006 2:36:14 AM

I don't think it's a "dumb" shrink. They've managed to cut power comsumption considerably, and I think it's better that they implement 65nm in K8 rather than start it with K8L.

Of course, it still makes much more sense to go with C2D, unless you have an AM2 motherboard and are looking to upgrade your CPU.
December 15, 2006 3:23:26 AM

Quote:
I don't think it's a "dumb" shrink. They've managed to cut power comsumption considerably, and I think it's better that they implement 65nm in K8 rather than start it with K8L.

Of course, it still makes much more sense to go with C2D, unless you have an AM2 motherboard and are looking to upgrade your CPU.

I might go AMD if this thing easily clocks around 3.5ghz. Until then, forget it. I reallllllllly need a new CPU/mobo/ram combo, I'm still using a 3200 and 1gig of ddr1 =[
December 15, 2006 3:28:33 AM

Gurious,

I am not sure where this idea that the 65nm shrink would produce some outrageous OC when it first came out.

130nm scaled from 1800-2600mhz.
90nm scaled from 1800-3000mhz.

The original 90nm parts did not scale much better than the 130nm parts, it wasn't until Venice that they really started scaling. So, why are we expecting a massive jump, when it has NEVER happend in the past?

Well, with a little better cooling, a stock FX cooler, we have seen a 65nm chip scale to 3.1 So, the earliest versions of the 65nm parts are working fine, they are doing just what AMD has done with past die shrinks. So, if history repeats, they should eventually make it up to 3.5 to 3.6 or maybe a little higher.

Intel has been a pretty good overclocker in the past years. And, until now, they needed to be a good overclocker in order to even compete. But, just like increasing the clock of the netburst architecture still didn't make it a good competitor to K8, it made it close, but K8 was the better choice IMO. It will take more than just increasing the core speed of K8 to make it a better choice than Core2. That is where K8L comes in.

So, this die shrink, does two things, makes them more power efficient, and allows AMD to produce the cpu's at a lower cost. This is a good thing of course. As the process matures, it will clock higher. Even thought you might not be aware of it, only makes the processors sell better in one market, our market. Our market is a very small market, so, they won't have any financial problems if more enthusiasts buy Core Cpu's.

What does matter, is that right now, they won't be able to clock them much higher for the OEM market. But, they will can be price competitive to core.
So, this is not big deal as far as I can tell, other than not out performing core, which was something it was already doing with the 90nm process.

wes
December 15, 2006 3:54:52 AM

What are you talking about? Venice was the first 90nm AMD chip. Also, they clocked high from the get-go.
Even if they hit 3.5-3.6, I'm really, really skeptical about them defeating 3.0ghz C2D chips let alone the ultra-overclocked 4ghz ones.
December 15, 2006 4:17:17 AM

Quote:
I reallllllllly need a new CPU/mobo/ram combo, I'm still using a 3200 and 1gig of ddr1 =[
Hey, I'm still on a 2500+ and also 1 gig ddr1 :cry: 
December 15, 2006 4:30:51 AM

Quote:
Bummer that it doesn't OC that well.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=288...


Excellent start for 65nm. The temps are lower than with the EE 3800+ which is rated at 35W.

OC is not the true value of a CPU, since most people never will.


Go AMD!

The temperature readings may not be accurate, the 65nm uses more power than both AMD EE and more than the faster E6600 as well.
December 15, 2006 4:34:20 AM

Quote:
I guess this was a "dumb" shrink after all. I really hope they do the right thing and drop those prices down a bit. As it stands today a 5000+ price wise is comparable to an E6600... that has to change. They should drop the 5000+ down to an E6400 to compete. Unfortunaly more and more people are OCing and that makes thing even thornier for Amd since "most" enthusiast can OC an E6300 to match an FX62. Luckly for them not alot of people know how to OC yet.


Yes and no. If you think in terms of increased perfromance, yes it was "dumb", but then, its the same Urarh, just a smaller node so expecting improved processing performance was unrealistic.

The reduced power usage is gravy. Even though it couldnt bring the AMDs in line with the Core2s in power/perfromance curve, its still an improvement over the 90nm parts.

The big win for AMD is in reduced die size. AMD needs this as desperately as they need to update their Uarch. Its 1/2 of the puzzle for them to compete with Intel, assuming they (and that is debatable assumption) can acheive mature yields in a short time. With increased yields, they should be able to drop prices, which is another thing they desperately need to do, rather than wasting time and money on the 4x4 PR project.
December 15, 2006 5:04:32 AM

Quote:
Bummer that it doesn't OC that well.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=288...


Hello,
I read the article from "Anand..." and, what can I say? Look's good, but the AMD 65nm have only 512+512 L2 cache. I think that is cause to obtain so power consumption. Realy, AMD have much to work for wining the crown of 2 Core.
All good for you!
December 15, 2006 5:21:13 AM

Based on this article:

BOOOOOOOOOOOOOOOOOOORING!

Wait for Brisbane! Wait for Brisbane! Wait for Brisbane!

I should have booked Qantas instead.

Bunch of friggin' lunatics running AMD lately. Only possible explanation. The time was right for AMD to pull a rabbit the size of a T Rex out of its hat and send Intel running back home with its tail between its legs. So what does it do? It comes out with a shrunk version of a 90nm chip that heralds 14 less W. Yippee Crap.

Time to go back to sleep. Goodnight AMD... :twisted:
December 15, 2006 6:01:59 AM

Quote:
What are you talking about? Venice was the first 90nm AMD chip.


Neg there were 90nm 3500+'s "Winchesters?" I know this cause I missed out and got a shitty Newcastle 130nm :(  BLAST!
December 15, 2006 6:19:39 AM

Quote:
Bummer that it doesn't OC that well.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=288...


Excellent start for 65nm. The temps are lower than with the EE 3800+ which is rated at 35W.

OC is not the true value of a CPU, since most people never will.


Go AMD!

OC potential of the new brisbane is a good indication of how far they can expect to push the K8 65nm frequency to compete with Core2. If they were able to clock them at say 3ghz from the get go, while maintaining low temps then they would be better able to compete with Core2 no?
December 15, 2006 6:51:38 AM

Venice was not the first 90nm chip. Venice was gen 2 of 90nm Winchester being the first Gen. Just read up on that a bit.

Also, I don't think I said they would defeat core 2. All I was intending to say, was that no one should have expected them to clock much higher than the current 90nm, plus, at stock (in the OEM market) where the money is, cost wise they can compete to a certain exent, but maybe not for long. Who knows, we will find out over the next few months.

wes
December 15, 2006 6:56:13 AM

Gurious,

This first run is a good indication of how high the first run will OC. Later generations of the at the 65nm node should clock higher. By the time AMD has the 65nm node well ironed out, K8L should be here. And, then, with the higher frequencies, and the increased efficiency of the arch. it should be able to compete, atleast on some levels, and maybe all.

wes
December 15, 2006 6:59:24 AM

CaptRobert,

Come on man, this is the same way they have always done it, they work differently than Intel does, and Intel they get a little closer to Intel in manufacturing ability, they will continue to work this way. They make shrinks, then they refine and make updates at the new node. Stop being a flamer Capt.

wes

Edit: sorry for so many in a row, I should have made one reply for all three.
December 15, 2006 7:37:28 AM

Quote:
Bummer that it doesn't OC that well.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=288...
I found it interesting that the E6600 used similar or less power while running most apps.
And that's with the E6600 paired with a 680i based motherboard, the biggest power consuming Intel supporting chipset.
December 15, 2006 7:38:18 AM

tanker,

I just looked at that link, and the 65nm AMD chip used less power in all the scenarios that I saw on link.

wes

Edit, just checked it again, the E6600 used less power in two of the tests, missed them the first time I looked.
December 15, 2006 7:47:49 AM

Quote:
Bummer that it doesn't OC that well.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=288...
I found it interesting that the E6600 used similar or less power while running most apps.

I maybe wrong but Amd has the rep of only releasing parts when they are ready no? Intel on the other hand is said to rush part out before they are ready...total opposite this time around. The 4x4 was obiously not ready, and the 65nm well....they took sure took their sweet time, you think they could of made some kind of improvements. I know, I know the K8L will be the chip that will be saviour of the green team, but if their current track record continues then I really don't expect much.
December 15, 2006 7:56:00 AM

Quote:
tanker,

I just looked at that link, and the 65nm AMD chip used less power in all the scenarios that I saw on link.

wes

Edit, just checked it again, the E6600 used less power in two of the tests, missed them the first time I looked.


WME: Intel
DivX: AMD
Quicktime:Intel
iTunes: Intel
3dsmax: Intel
Cinebench:Intel
Quake 4: Tied(Really too close to declare a winner).
Oblivion:AMD
HL2: AMD

Intel: Wins 5/9 and a tie
AMD: Wins 3/9 and a tie

Then test the AMD on the 680a chipset to even things up. Basically it looks like the AMD is a little more energy efficient in games, and the Intel is more efficient in encoding/rendering. :wink:
December 15, 2006 7:56:41 AM

4X4, that wasn't anything special. How could it have been? It is a 3ghz AMD chip on the same board as another one. These two cpu's are going up against what basically amounts to 2 Conroe's on one die. How in this universe, or any other, would it beat it? Two cpu's that are K8 architecture, are going up against 2 other ones, that already beat them once. 4X4 won't show you any type of great performance in my opinion until the CPU's that occupy the sockets are K8L. And in all honesty, they probably didn't expect to sell a bunch of those systems.

AMD talked about 4X4 for a while, and then released it on paper. The longer they could keep people in the dark on this the better, because eventually K8L will be out. 4X4 is a stop gap, it is merely to say, well we have a "Quad" core system also. Otherwise, Intel is the only name you will be seeing in the headlines.

As far as 65nm, I still stick to, you should not have really expected major improvements yet. AMD generally likes to do one thing at a time. Make the die shrink, get it running properly, then make core revisions. This is just the way they have been doing it for a while now. Not sure if it is the best way for AMD to operate, but it seems to have worked so far.

wes
December 15, 2006 7:58:57 AM

Quote:

Then test the AMD on the 680a chipset to even things up. Basically it looks like the AMD is a little more energy efficient in games, and the Intel is more efficient in encoding/rendering. :wink:

I'd imagine the only reason the 65nm 5000+ is using less power in games is that it's unable to utilize the 8800GTX as much as the E6600, too.
December 15, 2006 8:02:34 AM

Yeah, I missed a bunch of the benchmarks, I just checked AGAIN, for the third time, and saw all the rest of them. The power difference is very minimal in both directions.

What is interesting to me is, the AMD cpu is clocked 200mhz higher, has the memory controller on the cpu, and power consumption tends to be on par with the E6600. Now, if they can just get the performance up, they will be doing alright :wink:

wes
December 15, 2006 8:04:11 AM

Accord, 100% cpu usage is 100% cpu usage. It shouldn't matter either way.
Just for perspective, that is like saying, since a Prescott didn't perform like K8 in games, it should have lower power consumption while gaming.

wes

Edit: I do see what you are saying though, maybe the 8800gtx isn't working as hard, not the CPU. Thought about it after I made the post.
December 15, 2006 8:06:21 AM

Quote:

Then test the AMD on the 680a chipset to even things up. Basically it looks like the AMD is a little more energy efficient in games, and the Intel is more efficient in encoding/rendering. :wink:

I'd imagine the only reason the 65nm 5000+ is using less power in games is that it's unable to utilize the 8800GTX as much as the E6600, too.Yeah. It's not like there's a big difference between the 2 now, in most scenarios, it's just that all the AMD lovers bragged and bragged how close the 90nm AMD was to the 65nm Intel... and were c0ck-sure that 65nm AMD would embarrass Intel. Not the case. :wink:
December 15, 2006 8:15:45 AM

Quote:
CaptRobert,

Come on man, this is the same way they have always done it, they work differently than Intel does, and Intel they get a little closer to Intel in manufacturing ability, they will continue to work this way. They make shrinks, then they refine and make updates at the new node. Stop being a flamer Capt.


First of all it's fairly chilly today so I need to flame a bit to keep my toesies warm. :) 

Ok, your point is well taken. But let's look at it from the standpoint of people on the street who don't spend hours a day reading these forums and articles thus are up to the minute on what's going on. They read the occasional PC magazine and have been building up anticipation for these fabled Brisbane cores since springtime or even earlier. They watched C2D/Q trounce AMD into the enthusiast market dirt, stomp on it and relieve itself on its head before moving on. They naturally figured that it was a momentary setback and AMD would come back out swinging to reclaim the crown it had worn for so long. So in the period of just over a month, AMD:

- Introduced QuadFX that's just a head-shaking PC nightmare by Rube Goldberg.

- Launched 5600+ on an outdated, oversized Toledo!!!

- Finally brings out Brisbane and not only do you need a manual to read the number to figure out if you have an old 90nm or a new 65nm, but it turns out that the performance differences between the two sizes is a c___ hair.

I have been a ferocious AMD fan for many years, and spent many an hour drinking beers and pi$$ing on Pentium 4s. But in my many years, I've never seen a company disembowel one of it's own most active and loyal market segments as AMD is doing with its enthusiasts.

K8L? If what's been happening lately with AMD's staggering missteps is any indication, it will come out on a Newcastle core and benchmark competitively with a good Sempron! 8O
December 15, 2006 8:20:47 AM

Quote:
CaptRobert,

Come on man, this is the same way they have always done it, they work differently than Intel does, and Intel they get a little closer to Intel in manufacturing ability, they will continue to work this way. They make shrinks, then they refine and make updates at the new node. Stop being a flamer Capt.


First of all it's fairly chilly today so I need to flame a bit to keep my toesies warm. :) 

Ok, your point is well taken. But let's look at it from the standpoint of people on the street who don't spend hours a day reading these forums and articles thus are up to the minute on what's going on. They read the occasional PC magazine and have been building up anticipation for these fabled Brisbane cores since springtime or even earlier. They watched C2D/Q trounce AMD into the enthusiast market dirt, stomp on it and relieve itself on its head before moving on. They naturally figured that it was a momentary setback and AMD would come back out swinging to reclaim the crown it had worn for so long. So in the period of just over a month, AMD:

- Introduced QuadFX that's just a head-shaking PC nightmare by Rube Goldberg.

- Launched 5600+ on an outdated, oversized Toledo!!!

- Finally brings out Brisbane and not only do you need a manual to read the number to figure out if you have an old 90nm or a new 65nm, but it turns out that the performance differences between the two sizes is a c___ hair.

I have been a ferocious AMD fan for many years, and spent many an hour drinking beers and pi$$ing on Pentium 4s. But in my many years, I've never seen a company disembowel one of it's own most active and loyal market segments as AMD is doing with its enthusiasts.

K8L? If what's been happening lately with AMD's staggering missteps is any indication, it will come out on a Newcastle core and benchmark competitively with a good Sempron! 8OLMAO!!! You are a funny dude, and have a good writing style. You should apply to Tom's for a reviewers/author's job. :p 
December 15, 2006 8:26:07 AM

Quote:
LMAO!!! You are a funny dude, and have a good writing style. You should apply to Tom's for a reviewers/author's job. :p 


Re: Journalist Position - THG

Dear Dr. Thomas Pabst:

Please find enclosed my resume and a letter of recommendation by "Ontario's Finest", Mr. 1Tanker.

I am available immediately, have my own transportation and only need to take one day a week off to have my various STDs medically attended to.

Sincerely,

Capt. Robert April
December 15, 2006 8:30:46 AM

Quote:
tanker,

I just looked at that link, and the 65nm AMD chip used less power in all the scenarios that I saw on link.

wes

Edit, just checked it again, the E6600 used less power in two of the tests, missed them the first time I looked.


Intel 65nm is uses 3-6W less then the Amd 65nm in; WMA, Quicktime, iTunes, 3D rendering. Both chips practially use the same power in Divx, and Quake 4 with a difference of only .5-1W. Amd does wins in Oblivion by using 3W less then the Intel chip and in HL2 Amd wins by a huge amount...27W!!! which Anand attributes to the Core2 feeding the GPU the fastest...hmm I wonder why we don't see this in Quake or Oblivion, oh well, the only thing for sure is that in the Intel still holds a huge advantage in the performance sector.
December 15, 2006 8:42:19 AM

I edited that post, I missed a bunch of the benchmarks the first time I skimmed through them, so went back and edited the post to correct myself.

But like I said, it has a 200mhz higher clock speed, and ondie memory controller. I am not using that as some kind of crutch saying the AMD chip actually performs better if you down clock it and remove the IMC. I am just saying, that all they have to do now, is release K8L with the thermals, and Conroe performance numbers. Interesting times ahead, as usual.

wes
December 15, 2006 8:50:01 AM

Quote:
I edited that post, I missed a bunch of the benchmarks the first time I skimmed through them, so went back and edited the post to correct myself.

But like I said, it has a 200mhz higher clock speed, and ondie memory controller. I am not using that as some kind of crutch saying the AMD chip actually performs better if you down clock it and remove the IMC. I am just saying, that all they have to do now, is release K8L with the thermals, and Conroe performance numbers. Interesting times ahead, as usual.

wes
I agree, and it's definetely a step in the right direction. I just don't think that it's the big difference that many people were hoping for. The only real benefactor is AMD, as they will either charge the same and make more/chip or drop prices to compete more with Intel(most likely)... and maybe still make a tad more/chip.
December 15, 2006 8:52:34 AM

Quote:
LMAO!!! You are a funny dude, and have a good writing style. You should apply to Tom's for a reviewers/author's job. :p 


Re: Journalist Position - THG

Dear Dr. Thomas Pabst:

Please find enclosed my resume and a letter of recommendation by "Ontario's Finest", Mr. 1Tanker.

I am available immediately, have my own transportation and only need to take one day a week off to have my various STDs medically attended to.

Sincerely,

Capt. Robert AprilYour hired if you can just bring a bottle of Kwellada to work, and apply it in the washroom or (go to the free clinic on your own time). :wink:
December 15, 2006 9:23:25 AM

Quote:
Your hired if you can just bring a bottle of Kwellada to work, and apply it in the washroom or (go to the free clinic on your own time). :wink:


Dear Sir.

Thank you for your prompt reply. Unfortunately Kwellada has not had any effect on my various body louses thus I have been instructed by my doctor to bathe daily in hydrochloric acid. You do not have to concern yourself unduly with contamination of your office staff as I have my own Speedy 75SP Spray Gun and 7.0 CFM @ 40 PSI 1.5HP Air Compressor to apply several heavy coats of protective latex to my genitalia three times a day.

I look forward to contributing to your august journal and hope that the receptionist is under 19, blonde, stacked and appreciates the refined aesthetics of the doggy position.

Sincerely,

Capt. Robert April.

Now that we have completely and totally hijacked this thread, let's see if we can make a half-hearted effort to get back on track.

Qurious has stated that the overall power consumption difference at anywhere between .5 and 6 W in various benchmarks. That is going to contribute less than a degree C in temperature and a wholly insignificant amount to the power bill. Which still leads me to wonder "Where's the beef?"
December 15, 2006 9:32:31 AM

Where is the beef? Well, Capt. didn't know you liked "the beef". To each his own. :wink:

STOP, STOP, STOP, comparing to Core 2, and compare to the 90nm A64's. From looking at it, there is generally a 10-20 watt better power consumption of the 65nm 5000 over the 90nm 5000. So, that alone is a pretty good improvment. Of course, not saying that you will save a bunch of money by switching to AMD!!! Just that, the process shrink was successful. When have simple process shrinks ever made a cpu faster when all they do is shrink the process and nothing else?

wes
December 15, 2006 9:33:43 AM

Quote:
Your hired if you can just bring a bottle of Kwellada to work, and apply it in the washroom or (go to the free clinic on your own time). :wink:


Dear Sir.

Thank you for your prompt reply. Unfortunately Kwellada has not had any effect on my various body louses thus I have been instructed by my doctor to bathe daily in hydrochloric acid. You do not have to concern yourself unduly with contamination of your office staff as I have my own Speedy 75SP Spray Gun and 7.0 CFM @ 40 PSI 1.5HP Air Compressor to apply several heavy coats of protective latex to my genitalia three times a day.

I look forward to contributing to your august journal and hope that the receptionist is under 19, blonde, stacked and appreciates the refined aesthetics of the doggy position.

Sincerely,

Capt. Robert April.

Now that we have completely and totally hijacked this thread, let's see if we can make a half-hearted effort to get back on track.

Qurious has stated that the overall power consumption difference at anywhere between .5 and 6 W in various benchmarks. That is going to contribute less than a degree C in temperature and a wholly insignificant amount to the power bill. Which still leads me to wonder "Where's the beef?"I guess it's meant to be more like an appetizer....to hold you off til the main meal(K8L).
December 15, 2006 10:20:47 AM

Quote:
Where is the beef? Well, Capt. didn't know you liked "the beef". To each his own. :wink:

STOP, STOP, STOP, comparing to Core 2, and compare to the 90nm A64's. From looking at it, there is generally a 10-20 watt better power consumption of the 65nm 5000 over the 90nm 5000. So, that alone is a pretty good improvment. Of course, not saying that you will save a bunch of money by switching to AMD!!! Just that, the process shrink was successful. When have simple process shrinks ever made a cpu faster when all they do is shrink the process and nothing else?


I definitely like my beef and have enjoyed placing my beef in many a hot taco... :lol: 

OK, let's look at these alternative marketing strategies in place of their "bummer acid trip" launches of the past couple of months.

A) AMD today releases 65nm 6000+ @ $399 street. Not the top performing CPU in the world, but definitely competitive and worthy of regaining respect for the brand in this particular market.

B) AMD continues to take its lumps and doesn't waste any more R&D time and money in "who gives a flying f***" processors like the 65nm 5000+ or the 90nm 5600+. At the beginning of Q2 '07, K8L comes out in quad, dual and single core configurations providing 20%+ performance advantage over Intel's offerings and priced within a couple of bucks of them.

C) AMD doesn't even launch Quad FX, but holds back until they are able to introduce it in a jaw-dropping Spring '07 spectacle with K8L Quads sitting on it, demos it with a tuned multithreaded app and turns benchmark numbers that give you nosebleeds just reading them.

Come on. No one can convince me that their current marketing launch strategy is better than this. AMD's becoming the butt of jokes throughout the enthusiast segment. Given their background, you'd expect them to pull this off with some significant flair, not this amateur Marketing 101 flunkout crap.

Quote:
I guess it's meant to be more like an appetizer....to hold you off til the main meal(K8L).


Yeah, but what if AMD stays true to its current form and instead of K8L being a $9.95 king-sized slab of prime rib with a baked potato the size of your head, it comes out as a nouvelle cuisine delicate one ounce medallion fillet with a baby carrot, a truffle shaving and a couple of drops of Perigueux sauce on a big empty plate for 75 bucks?
December 15, 2006 11:47:05 AM

Quote:
Where is the beef? Well, Capt. didn't know you liked "the beef". To each his own. :wink:

STOP, STOP, STOP, comparing to Core 2, and compare to the 90nm A64's. From looking at it, there is generally a 10-20 watt better power consumption of the 65nm 5000 over the 90nm 5000. So, that alone is a pretty good improvment. Of course, not saying that you will save a bunch of money by switching to AMD!!! Just that, the process shrink was successful. When have simple process shrinks ever made a cpu faster when all they do is shrink the process and nothing else?


I definitely like my beef and have enjoyed placing my beef in many a hot taco... :lol: 

OK, let's look at these alternative marketing strategies in place of their "bummer acid trip" launches of the past couple of months.

A) AMD today releases 65nm 6000+ @ $399 street. Not the top performing CPU in the world, but definitely competitive and worthy of regaining respect for the brand in this particular market.

B) AMD continues to take its lumps and doesn't waste any more R&D time and money in "who gives a flying f***" processors like the 65nm 5000+ or the 90nm 5600+. At the beginning of Q2 '07, K8L comes out in quad, dual and single core configurations providing 20%+ performance advantage over Intel's offerings and priced within a couple of bucks of them.

C) AMD doesn't even launch Quad FX, but holds back until they are able to introduce it in a jaw-dropping Spring '07 spectacle with K8L Quads sitting on it, demos it with a tuned multithreaded app and turns benchmark numbers that give you nosebleeds just reading them.

Come on. No one can convince me that their current marketing launch strategy is better than this. AMD's becoming the butt of jokes throughout the enthusiast segment. Given their background, you'd expect them to pull this off with some significant flair, not this amateur Marketing 101 flunkout crap.

Quote:
I guess it's meant to be more like an appetizer....to hold you off til the main meal(K8L).


Yeah, but what if AMD stays true to its current form and instead of K8L being a $9.95 king-sized slab of prime rib with a baked potato the size of your head, it comes out as a nouvelle cuisine delicate one ounce medallion fillet with a baby carrot, a truffle shaving and a couple of drops of Perigueux sauce on a big empty plate for 75 bucks?

Were you waiting for a meal when you wrote your reply? LOL

Now, on the Brisbane, I think AMD has done a successful shrink, which may not have been tested to it's potential (there might be a need to a BIOS update to the motherboard, that we are unaware of, and that will come in the next weeks as the cpu availability grows). Comparing to 90nm it is an improvement in power comsuption and probably we'll see a revision in february (maybe sporting a fx running over 3 GHz - tough I wouldn't bet any money on it).

On the K8L, I think it might improve considerably on the K8, but I think everyone is making a mistake of not taking Intel into account on their predictions. Intel has been really aggressive lately, and we might see 45nm available by the time K8L comes out, so Intel can steal AMD's show.
December 15, 2006 11:59:31 AM

AMD better hope we don't drop Penryn before K8L. It's a monster.
December 15, 2006 12:07:17 PM

Quote:

Were you waiting for a meal when you wrote your reply? LOL

Now, on the Brisbane, I think AMD has done a successful shrink, which may not have been tested to it's potential (there might be a need to a BIOS update to the motherboard, that we are unaware of, and that will come in the next weeks as the cpu availability grows). Comparing to 90nm it is an improvement in power comsuption and probably we'll see a revision in february (maybe sporting a fx running over 3 GHz - tough I wouldn't bet any money on it).

On the K8L, I think it might improve considerably on the K8, but I think everyone is making a mistake of not taking Intel into account on their predictions. Intel has been really aggressive lately, and we might see 45nm available by the time K8L comes out, so Intel can steal AMD's show.


Well, I was replying to questions about "beef" and "appetizer"! :) 

I have previously stated that the sum of all my knowledge on the Brisbane benchmarks are from Anand. If it turns out that there's some magical BIOS flash that turns this into a Speedy Gonzales on crystal meth, then I will definitely, and happily, retract my booing.

Quote:
AMD better hope we don't drop Penryn before K8L. It's a monster.


So, my dear Mr JK Intel... are you in a position to drop any tasty crumbs to add to the info in:

THG Penryn Article???

8)
December 15, 2006 12:12:03 PM

Quote:

Well, I was replying to questions about "beef" and "appetizer"! :) 

I have previously stated that the sum of all my knowledge on the Brisbane benchmarks are from Anand. If it turns out that there's some magical BIOS flash that turns this into a Speedy Gonzales on crystal meth, then I will definitely, and happily, retract my booing.



Well, I don't expect marvells to come out of a bios update, but we could see some improvements on power consumption and probably those benchmarks where the 65 nm had less performance than it's 90nm counterpart may change.
December 15, 2006 12:21:18 PM

Quote:
Well, I don't expect marvells to come out of a bios update, but we could see some improvements on power consumption and probably those benchmarks where the 65 nm had less performance than it's 90nm counterpart may change.


So the bottom line is still: BOOOOOOOOOOOOOORING!
December 15, 2006 12:42:00 PM

You can't be truly disappointed, everybody knew that this would be just a process shrink, no architectural changes were expected.
I think you're just trying to fan on the fire here to spice up the discussion, cause i see that you're overall a reasonable and unbiased guy ;) 
To the guys which are counting in how many tests the newcomer used less power than the E6600: you're nitpicking.
With the exception of HL2, in all the other tests the difference is withing +/- 5W, for the whole platform.. and on a total consumption of more than 200W, that's less than 3% of a difference.
So who cares?
Concerning the clocking potential.. i'm actually positively surprised that they could clock it past 2.9GHz with the stock cooler.. this means that the 65nm process is already roughly on par with the much more mature 90nm process.
December 15, 2006 12:44:05 PM

The performance will be on par with 90nm, give or take a 1-2% which can be within the margin of error of the test itself.
December 15, 2006 12:49:03 PM

Quote:
4X4, that wasn't anything special. How could it have been? It is a 3ghz AMD chip on the same board as another one. These two cpu's are going up against what basically amounts to 2 Conroe's on one die. How in this universe, or any other, would it beat it? Two cpu's that are K8 architecture, are going up against 2 other ones, that already beat them once. 4X4 won't show you any type of great performance in my opinion until the CPU's that occupy the sockets are K8L. And in all honesty, they probably didn't expect to sell a bunch of those systems.

AMD talked about 4X4 for a while, and then released it on paper. The longer they could keep people in the dark on this the better, because eventually K8L will be out. 4X4 is a stop gap, it is merely to say, well we have a "Quad" core system also. Otherwise, Intel is the only name you will be seeing in the headlines.

As far as 65nm, I still stick to, you should not have really expected major improvements yet. AMD generally likes to do one thing at a time. Make the die shrink, get it running properly, then make core revisions. This is just the way they have been doing it for a while now. Not sure if it is the best way for AMD to operate, but it seems to have worked so far.

wes



QFX systems are available RIGHT NOW at several e-tailers. It's just an Opteron system with no ECC. Opteron systems are good. This wasn't meant to compete with C2Q.

It was meant to give FX a shot in the arm.
!