Sign in with
Sign up | Sign in
Your question

AMD's 65nm Preview Part 2 - The Plot Thickens

Last response: in CPUs
Share
December 21, 2006 5:11:05 AM

Not sure if anyone has posted this but things are not looking so good for Amd's 65nm transition. I can deal with it just being a "dumb" shrink and getting no performance from it, but now anand is saying that it actually performs worse then the 90nm chips on several applications due to memory latency. Well their goes the theory that Amd will only release a product when it is ready. Another interesting point is that their shrink is not that great considering that they have kept the same number of transistors....anyone have any ideas as to why? On the bright side it looks like Amd does have something good on the idle power consumption front.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=289...
December 21, 2006 5:17:00 AM

Quote:
the new core is slower. We couldn't figure out why AMD made the change and with most of our key AMD contacts on vacation due to the holidays, we still have no official response on the matter. Rest assured that if/when we learn more we will let you know.



AMD really lost it now. :lol: 
December 21, 2006 5:23:26 AM

Quote:
the new core is slower. We couldn't figure out why AMD made the change and with most of our key AMD contacts on vacation due to the holidays, we still have no official response on the matter. Rest assured that if/when we learn more we will let you know.



AMD really lost it now. :lol: 

You would think that they would be all over this and have someone at their PR department working the holidays to get their response to this. I guess they figure that only computer enthusiast really care about this subject. Well they do have dell now to sell to, who cares about the enthusiast market.
Related resources
December 21, 2006 5:26:53 AM

Quote:
Not sure if anyone has posted this but things are not looking so good for Amd's 65nm transition. I can deal with it just being a "dumb" shrink and getting no performance from it, but now anand is saying that it actually performs worse then the 90nm chips on several applications due to memory latency. Well their goes the theory that Amd will only release a product when it is ready. Another interesting point is that their shrink is not that great considering that they have kept the same number of transistors....anyone have any ideas as to why? On the bright side it looks like Amd does have something good on the idle power consumption front.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=289...
It looks like AMD just decided to release 65nm to slow the migration to C2D(before the process was up spec), and hope that loyalty would sell Brisbane.
December 21, 2006 5:31:37 AM

Quote:
the new core is slower. We couldn't figure out why AMD made the change and with most of our key AMD contacts on vacation due to the holidays, we still have no official response on the matter. Rest assured that if/when we learn more we will let you know.



AMD really lost it now. :lol: 

You would think that they would be all over this and have someone at their PR department working the holidays to get their response to this. I guess they figure that only computer enthusiast really care about this subject. Well they do have dell now to sell to, who cares about the enthusiast market.

A pattern here perhaps?

Quad FX
Brisbane

What next? K8L a major flop? I wouldn't be surprised.
December 21, 2006 5:33:54 AM

Quote:
Not sure if anyone has posted this but things are not looking so good for Amd's 65nm transition. I can deal with it just being a "dumb" shrink and getting no performance from it, but now anand is saying that it actually performs worse then the 90nm chips on several applications due to memory latency. Well their goes the theory that Amd will only release a product when it is ready. Another interesting point is that their shrink is not that great considering that they have kept the same number of transistors....anyone have any ideas as to why? On the bright side it looks like Amd does have something good on the idle power consumption front.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=289...
Things just keep getting worse for AMD. I think they are beginning to see that things aren't so rosy, as they are now planning on dropping the PR rating(finally).

http://www.hkepc.com/bbs/itnews.php?tid=715237&starttim...
a b à CPUs
December 21, 2006 6:20:45 AM

*waits for flood of amd fanboy... er baron...*
December 21, 2006 9:35:03 AM

Ok, this is pretty bizarre :? I'm no expert but I can't think of any good reason why they would increase L2 latency - anyone got any ideas?

You have to admit this has not been a good half year for AMD - underwhelming performance increases in the best case for AM2, and performance drops with less than top notch memory, 4x4 ended up being a gimmick, and now this.

Let's hope that K8L makes up for this series of disappointments, or AMD is gonna be in big trouble.
December 21, 2006 10:28:41 AM

Quote:
AMD really lost it now. :lol: 

Before we get too ahead of ourselves, let us not forget the P3 1.13GHz, Itanium, BTX and Netburst, which is pretty much one long string of f*ck ups from the guys at Intel. Logic states that everybody will screw up. Its just how you bounce back.
December 21, 2006 10:52:54 AM

Quote:


A pattern here perhaps?

Quad FX
Brisbane

What next? K8L a major flop? I wouldn't be surprised.


Well K8L definately ISNT going to beat the core arch:
Q3 K8L debut clock ~2.8 Ghz
Q3 core arch debut clocks ~3.8 Ghz

K8L has to be 30%+ faster clock-for-clock JUST TO PULL LEVEL.

To whup intel it would need to be 60%+ faster clock-for-clock which just isnt going to happen.
December 21, 2006 11:04:25 AM

Quote:
Not sure if anyone has posted this but things are not looking so good for Amd's 65nm transition. I can deal with it just being a "dumb" shrink and getting no performance from it, but now anand is saying that it actually performs worse then the 90nm chips on several applications due to memory latency. Well their goes the theory that Amd will only release a product when it is ready. Another interesting point is that their shrink is not that great considering that they have kept the same number of transistors....anyone have any ideas as to why? On the bright side it looks like Amd does have something good on the idle power consumption front.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=289...


1. That's why ALL X2s/X4s will get L3. It will end all latency issues.

2. Even the X2 3800+ can play Oblivion at 1600x1200.

3. Though losing frames is not what you want, going from 139.5 to 133.9 will not alter the experience.

It's not as bad a picture as your painting.
December 21, 2006 11:07:11 AM

Quote:


A pattern here perhaps?

Quad FX
Brisbane

What next? K8L a major flop? I wouldn't be surprised.


Well K8L definately ISNT going to beat the core arch:
Q3 K8L debut clock ~2.8 Ghz
Q3 core arch debut clocks ~3.8 Ghz

K8L has to be 30%+ faster clock-for-clock JUST TO PULL LEVEL.

To whup intel it would need to be 60%+ faster clock-for-clock which just isnt going to happen.


There is no such thing as a flop in this business. Do you think gong from dual core to quad core will do LESS than 40%?
The last quote was Barcelona 70% than Opteron.
December 21, 2006 11:10:05 AM

Agreed.

This is a rough patch AMD is going through now.
The second generation 65nm should be be better.

Ever since Intel's Core 2 came out, we all more or less knew that we would have to wait until Barcelona to get a large increase in performance. This is just proving that we were right.

I want everyone to remember this:
Intel was lagging behind AMD ever since the first Athlon 64 CPUs came out, for years... from September 23, 2003 (First Athlon 64) to July 27, 2006

AMD was leading Intel until:
The first wave of Core 2 processors was officially released on July 27, 2006

Intel has been leading for LESS THAN 5 months.

Barcelona will be released in 2H '07.
We have yet to see the performance of this chip, but we should assume it will be quite a lot better than the current AMD chips.
Putting 2 x 4 cores in a 4x4 setup = 8 cores, this should at least be competitive, but will probably take the performance crown.

Oh by the way... before the first Athlon, AMD was behind Intel ever since it was established in 1969. AMD is still here. It will continue to be here. Even if AMD never get the performance crown again (I doubt it), AMD can offer better price/performance ratios, and compete in the budget/midrange market.

Don't write AMD off so soon....
December 21, 2006 11:18:26 AM

Quote:


There is no such thing as a flop in this business. Do you think gong from dual core to quad core will do LESS than 40%?
The last quote was Barcelona 70% than Opteron.


Are you really this stupid, is anyone?

Core for core comparison knumbskull.

AMD claimed 40% at the K8L demo, I was in the audiance ... better phone them up and tell them to correct their powerpoints :lol: 
December 21, 2006 11:33:50 AM

Quote:
I want everyone to remember this:
Intel was lagging behind AMD ever since the first Athlon 64 CPUs came out, for years... from September 23, 2003 (First Athlon 64) to July 27, 2006

AMD was leading Intel until:
The first wave of Core 2 processors was officially released on July 27, 2006

wasnt since athlon XP? actually, it was since thunderbirds i guess... around 2000? then that's 6 years behind.
December 21, 2006 11:36:53 AM

Ill Agree With That
December 21, 2006 11:38:15 AM

LOL
December 21, 2006 11:40:57 AM

I Can Agree Witha That Also
December 21, 2006 11:42:03 AM

AMD isn't in the business of Revolution. It's all about evolution. . . it takes time. Millions millions of years :roll:

*note the sarcasm in my voice*
December 21, 2006 11:45:31 AM

Yeah Definitely, Just Look At My System, It May Not Be Core2 Fast But Its Fast Enuff That It Will Beat Most High-End PCs $3,000 And Up In Hard Drive Thouroughput And Will Play Any Game I Throw At It, Even Games In 2007/Early 2008, Im Not Uppgrading CPU/Mobo And RAM Again (But Vista Will Definitely Be Had) Until AMD Quad-Cores Become Available In The $300.00 - $400.00 Range, I Also Have Said Before That ALL AMD Chips Will End-Up With A L3 Cache (Yes Even The Semprons, And They Will Even Be At-Least Dual-Core)
December 21, 2006 11:49:49 AM

Sh$T happens.

Look at any other industry. Bad products, recalls. etc. etc. etc.

Just read consumer reports


It happens...................

How about that battery recall for laptops...........
If you have ever owned a car I am sure that you have gotten a recall on it before.
December 21, 2006 11:53:43 AM

Quote:

I want everyone to remember this:
Intel was lagging behind AMD ever since the first Athlon 64 CPUs came out, for years... from September 23, 2003 (First Athlon 64) to July 27, 2006

I wouldn't say that, when the A64 was released, it had an advantage in most single-threaded apps and games, the P4 had the advantage in most encoding apps and multi-threaded apps.

It wasn't until the dual-core that the A64 gained a decisive advantage.

Quote:
wasnt since athlon XP? actually, it was since thunderbirds i
guess... around 2000? then that's 6 years behind.

Northwood As gained parity with the XP, the Northwood Bs took the performance crown and the Northwood Cs were quite dominant for their time.
December 21, 2006 12:01:22 PM

Quote:
Quote:

wasnt since athlon XP? actually, it was since thunderbirds i
guess... around 2000? then that's 6 years behind.

Northwood As gained parity with the XP, the Northwood Bs took the performance crown and the Northwood Cs were quite dominant for their time.
clock for clock? what about power consumption? way before this performance per watt stuff
December 21, 2006 12:08:46 PM

Quote:

clock for clock? what about power consumption? way before this performance per watt stuff

-Raw performance is more important than clock-for-clock
-Power consumption wasn't a important issue, few reviews ever had any power measurements
December 21, 2006 12:26:39 PM

Quote:


There is no such thing as a flop in this business. Do you think gong from dual core to quad core will do LESS than 40%?
The last quote was Barcelona 70% than Opteron.


Are you really this stupid, is anyone?

Core for core comparison knumbskull.

AMD claimed 40% at the K8L demo, I was in the audiance ... better phone them up and tell them to correct their powerpoints :lol: 

The presentation with Task Manager quoted "40% FP increase and 70% integer." Got to AMDs virtual showroom and see for your self. Someone posted a link to it.

The first numbers were 60% increase in integer and 40% FP. I even made a joke about the numbers steadily going up. Since there was not a peep about real perf increases for AM2, QFX or 65nm, this means that there will be a new Sheriff in town. If they get 80%+ increase over K8 by Jul/Aug, every other architecture will be pushed off the TPC-H charts. Even SuperDome(Itanium 2 - 64 chips) won't be able to handle a 32 proc Barcelona beast.
December 21, 2006 12:34:44 PM

Quote:

If they get 80%+ increase over K8 by Jul/Aug, every other architecture will be pushed off the TPC-H charts. Even SuperDome(Itanium 2 - 64 chips) won't be able to handle a 32 proc Barcelona beast.

The estimate is 70% for a 8-core K8L vs 4-core K8 system in TPC-C, which would roughly tie an existing Clovertown system.
December 21, 2006 12:41:06 PM

Quote:
AMD really lost it now. :lol: 

Before we get too ahead of ourselves, let us not forget the P3 1.13GHz, Itanium, BTX and Netburst, which is pretty much one long string of f*ck ups from the guys at Intel. Logic states that everybody will screw up. Its just how you bounce back.

Agreed. While the first spin of 65 nm is apparently disappointing, AMD typically improves their processes with major revisions. The clock will get better in 6-9 months from now I suspect, in time for K8L.

Jack

You should have realized that that's why Brisbane is just a dumb shrink. They have at least 4 runs (2 months prod time) - more if you count monthly starts - to get the process as good as 90nm was after 6-8 months. I think they will really need to get the new mobile chip out fast since Anand's is reporting that laptops will overtake desktops next year.

I mean sure they can just shrink Turion X2, but they need to - if nothing else - get clocks up slightly in the same envelope.
December 21, 2006 12:56:03 PM

Quote:
The last quote was Barcelona 70% than Opteron.


You don't specify if that is per clock or the gains from going to quad.



http://www.amd.com/us-en/assets/content_type/Downloadab...
Projected ~ 40% performance increase*
Projected ~60% performance per watt improvement over previous AMD product
December 21, 2006 1:41:02 PM

Do you guys loose sleep at night thinking/worrying about this stuff. :roll:
December 21, 2006 1:54:00 PM

Quote:

clock for clock? what about power consumption? way before this performance per watt stuff

-Raw performance is more important than clock-for-clock
-Power consumption wasn't a important issue, few reviews ever had any power measurements
- hmmm i dont think so, coz now intel is claimed to have better technology because it achieves higher performance at lower clock speeds, as it used to be with amd.
- power consumption was an issue, but reviewers translated that into heat dissipation at that time. Now there's this whole story about performance per watt.
December 21, 2006 1:56:43 PM

Jack,

I am a little disappointed about the performance decrease. It really isn't that bad, or something you would even notice, but, not something you would expect from a new product. First, AM2 takes a hit with DDR2 latencies, now the next cpu shrink has higher latency cache on it, which causes a similar performance hit that we saw at the release of AM2.

Well, we know the process is in need of refinement(figured it would be prior to release, just not like this). This shows because it doesn't even scale as high as current 90nm offerings. We know, that for whatever reason, they put higher latency cache on the die which caused the performance hit. They didn't disclose any benchmarks, only spoke of power. AMD obviously knew the first parts would not be as good as the best 90nm parts. Which taking into account low yields, the above would explain why they didn't release this to the enthusiast sector. They are probably using these chips to make more money on the OEM contracts, or using them to sell at a lower price to keep the OEM contracts.

The performance hit isn't very substantial, but it is still a newer product performing worse than the older product. Other than power consumption, and possibly lower price in the future, they have no way to promote the first run of these cpu's. I expect the next revision to be better, as Jack stated, since the revisions are typically major to an extent. Lets hope they get this sorted out, as I am sure they will.

wes

Edit: any thoughts on this Jack/everyone else?? Please refrain from starting a fanboy flame thread.... both side.(not you Jack, just the typical hecklers)
December 21, 2006 1:58:25 PM

Quote:
What next? K8L a major flop? I wouldn't be surprised.


I'm going to go out on a limb and say that K8L is going to be a good change for AMD. I think it will get them back in the game and running neck and neck with some benchies going to AMD and some benchies going to Intel. That's the way I'd like to see things with fierce competition.
December 21, 2006 3:13:17 PM

Ive said before that AMD is currently hogtied to oems.We have all read that they were putting their efforts towards the server market, and with the oems,THEY are bottlenecked. Sounds like thats fine so far. Now lets assume the new facility, plus the familiarity of the 65 nm process increases production, then by the time K8L rolls out AMD will have product for the channel et el. If that IS the plan, then theyve made a fair start. The 65 nm process I believe was a rush, as been said, and revisions are to come. imho
December 21, 2006 3:15:20 PM

Quote:
What next? K8L a major flop? I wouldn't be surprised.


I'm going to go out on a limb and say that K8L is going to be a good change for AMD. I think it will get them back in the game and running neck and neck with some benchies going to AMD and some benchies going to Intel. That's the way I'd like to see things with fierce competition.

I think that by the time K8L hits the street Intels next generation core will be out, and, while K8L is running neck and neck with C2D, the new procs will be head and shoulders above both K8L and C2D (or C2Q for that matter).

I still maintain that Intel is going to continue to push, releasing products ahead of schedule, that, this time, perform as advertised. They're not going to rest on their hindquarters anymore, I think the few years of getting thier asses handed to them by AMD opened their eyes and have caused them to gain focus that they've never had before. Doesn't mean I want AMD to go away - I wouldn't want either company to die...one CPU manufacturer is bad. :) 
December 21, 2006 3:22:02 PM

Intel said a new generation every two years. I think they'll be busy shrinking and increasing the clock and FSB on Core 2's rather than coming out with something new. That'll give AMD a chance to regain some ground.
December 21, 2006 3:28:37 PM

Quote:

If they get 80%+ increase over K8 by Jul/Aug, every other architecture will be pushed off the TPC-H charts. Even SuperDome(Itanium 2 - 64 chips) won't be able to handle a 32 proc Barcelona beast.

The estimate is 70% for a 8-core K8L vs 4-core K8 system in TPC-C, which would roughly tie an existing Clovertown system.

Why would they rate increases any other way? Doing that is a joke. of course 8 cores should be a lot faster than 4.
December 21, 2006 3:32:32 PM

Quote:
Why would they rate increases any other way? Doing that is a joke. of course 8 cores should be a lot faster than 4.


Anyone who has some sort of wildly threaded application is going to want 8-cores. But I have a feeling that there will be very few benefits with 8 cores no matter who does it, AMD or Intel. Until software catches up I don't even have use for a quad core (AMD or Intel). I barely make use out of my dual core.
December 21, 2006 3:36:49 PM

Quote:
CPU CPU-Z L2 Cache Latency ScienceMark 2.0 L2 Cache Latency
AMD Athlon 64 X2 5000+ (65nm) 20 cycles 20 cycles
AMD Athlon 64 X2 5000+ (90nm) 12 cycles 12 cycles

The original K8 core, in both 130nm and 90nm flavors, had a 12-cycle L2 cache. With Brisbane, as reported by both CPU-Z and ScienceMark, 65nm K8 now has a 20-cycle L2 cache. Generally speaking you move to a higher latency cache if you're planning on introducing a larger cache size, but a quick glance at AMD's roadmaps doesn't show anything larger than a 1MB L2 per core for the next year. The argument for higher clock speeds isn't valid either as the highest clock speed on AMD's roadmaps thus far is only 3.2GHz.

Luckily the performance impact of the higher latency L2 cache isn't noticeable in all applications, thanks to the K8's on-die memory controller, but make no mistake - the new core is slower. We couldn't figure out why AMD made the change and with most of our key AMD contacts on vacation due to the holidays, we still have no official response on the matter. Rest assured that if/when we learn more we will let you know.


NOw y did AMD increase the L2 cache latency from 12 cycles to 20 cycles?

Will that help those 65nm processors to clock higher?
December 21, 2006 3:37:57 PM

This is nothing new. When amd switched from 130 to 90 it happened.... I was lucky enough to get my 64 3000+ clawhammer before they started making the crappy 90nm ones.... and I really mean crappy, look at tom's cpu charts and compare the dif between clawhammer and all the other versions of the 3000+
December 21, 2006 3:51:58 PM

Quote:
The last quote was Barcelona 70% than Opteron.


You don't specify if that is per clock or the gains from going to quad.



http://www.amd.com/us-en/assets/content_type/Downloadab...
Projected ~ 40% performance increase*
Projected ~60% performance per watt improvement over previous AMD product


B..B..But ...
\
:oops:  <= Our local resident AMD fanboi
a b à CPUs
December 21, 2006 4:05:17 PM

Quote:
AMD really lost it now. :lol: 

Before we get too ahead of ourselves, let us not forget the P3 1.13GHz, Itanium, BTX and Netburst, which is pretty much one long string of f*ck ups from the guys at Intel. Logic states that everybody will screw up. Its just how you bounce back.
And...
Quote:
Agreed.

This is a rough patch AMD is going through now.
The second generation 65nm should be be better.

Ever since Intel's Core 2 came out, we all more or less knew that we would have to wait until Barcelona to get a large increase in performance. This is just proving that we were right.

I want everyone to remember this:
Intel was lagging behind AMD ever since the first Athlon 64 CPUs came out, for years... from September 23, 2003 (First Athlon 64) to July 27, 2006

AMD was leading Intel until:
The first wave of Core 2 processors was officially released on July 27, 2006

Intel has been leading for LESS THAN 5 months.

Barcelona will be released in 2H '07.
We have yet to see the performance of this chip, but we should assume it will be quite a lot better than the current AMD chips.
Putting 2 x 4 cores in a 4x4 setup = 8 cores, this should at least be competitive, but will probably take the performance crown.

Oh by the way... before the first Athlon, AMD was behind Intel ever since it was established in 1969. AMD is still here. It will continue to be here. Even if AMD never get the performance crown again (I doubt it), AMD can offer better price/performance ratios, and compete in the budget/midrange market.

Don't write AMD off so soon....


Good way to keep this whole thread in context and give the discussion the proper perspective. You both make good points.

I'm anticipating some interesting developments from AMD and look forward to the back and forth between both companies. I just wish I had enough money to build both Intel and AMD based machines to really enjoy the technology.
December 21, 2006 4:26:40 PM

Quote:
Yeah Definitely, Just Look At My System, It May Not Be Core2 Fast But Its Fast Enuff That It Will Beat Most High-End PCs $3,000 And Up In Hard Drive Thouroughput And Will Play Any Game I Throw At It, Even Games In 2007/Early 2008, Im Not Uppgrading CPU/Mobo And RAM Again (But Vista Will Definitely Be Had) Until AMD Quad-Cores Become Available In The $300.00 - $400.00 Range, I Also Have Said Before That ALL AMD Chips Will End-Up With A L3 Cache (Yes Even The Semprons, And They Will Even Be At-Least Dual-Core)


Now was there any need for that?
December 21, 2006 4:47:57 PM

Quote:
The last quote was Barcelona 70% than Opteron.


You don't specify if that is per clock or the gains from going to quad.



http://www.amd.com/us-en/assets/content_type/Downloadab...
Projected ~ 40% performance increase*
Projected ~60% performance per watt improvement over previous AMD product


What an a-hloe. Can you read where it says SPECFP at the top of the slide. I said

40% FP increase.

go to the Virtual Mall or whatever and it says 70% in the video for Integer.
December 21, 2006 4:51:15 PM

Quote:
Why would they rate increases any other way? Doing that is a joke. of course 8 cores should be a lot faster than 4.


Anyone who has some sort of wildly threaded application is going to want 8-cores. But I have a feeling that there will be very few benefits with 8 cores no matter who does it, AMD or Intel. Until software catches up I don't even have use for a quad core (AMD or Intel). I barely make use out of my dual core.


Down below some idiot trying to prove I said something wrong posted that the increases are based on OLTP (on-line Transaction processing read:D atabases).

Those will use as many cores as you can give them. There are 64 socket clusters running Oracle and SQL.
December 21, 2006 4:56:42 PM

Quote:
The last quote was Barcelona 70% than Opteron.


You are implying that the 70% is the improvement overall. I quoted AMD's general improvement of 40%. If you meant integer, then you should have said 70% integer. Not like it's something to brag about anyways when it's barely ahead of a 6 month product.

I have a suggestion. Quit name calling and acting like an idiot and people will respect you :roll:
December 21, 2006 5:01:25 PM

Quote:
Baron, I knew it was a shrink of the K8 core -- you have seen me post about that several times, but it was not a dumb shrink as you think. The fact that they went to 1/2 multipliers means some redesign was needed to set the PLLs correctly.

Second, if you haven't been paying attention, I have been down on the AMD 65 nm process since Feb of this year. Now, to be frank, it turned out worse than I exptected. Based on AMD's original 65 nm data that I keep linking to.... these processors should OC to 3.2-3.4 range. They are struggling to break 3.0 -- there is something very wrong here. I will almost bet money that SiGe did not make it in for this revision. So I expect it to improve, but for now this release is not much better than 4x4 -- the postive note is that power consumption is now on par with C2D.

Third, even after they fix it -- the max clock on the 65 nm in it's lifetime will be no better than 3.4 GHz (stock) and 3.6 GHz( overclocked) if they do no better than the quality of the transistor they published a year ago.



The problem I have is that you make it sound like you're producing better 65nm out of your garage.
Making chips isn't easy. If the damn thing works it's better than nothing. Only people who actually care about each instruction and each cycle will even notice a small decrease, but even you will notice when they get the process matured.
December 21, 2006 5:14:10 PM

Quote:

I want everyone to remember this:
Intel was lagging behind AMD ever since the first Athlon 64 CPUs came out, for years... from September 23, 2003 (First Athlon 64) to July 27, 2006

I wouldn't say that, when the A64 was released, it had an advantage in most single-threaded apps and games, the P4 had the advantage in most encoding apps and multi-threaded apps.

It wasn't until the dual-core that the A64 gained a decisive advantage.

Quote:
wasnt since athlon XP? actually, it was since thunderbirds i
guess... around 2000? then that's 6 years behind.

Northwood As gained parity with the XP, the Northwood Bs took the performance crown and the Northwood Cs were quite dominant for their time.

I'm glad someone else remembers how it really went down.
December 21, 2006 5:21:38 PM

Quote:
The last quote was Barcelona 70% than Opteron.


You are implying that the 70% is the improvement overall. I quoted AMD's general improvement of 40%. If you meant integer, then you should have said 70% integer. Not like it's something to brag about anyways when it's barely ahead of a 6 month product.

I have a suggestion. Quit name calling and acting like an idiot and people will respect you :roll:

Why don't you go nitpick in the video card area? If you knew OLTP was 70% (you posted a pic) then there should have been no ambiguity.

I even stated exactly that right below the first post. 70% integer, 40% FP.
December 21, 2006 5:24:55 PM

Quote:
AMD really lost it now. :lol: 

Before we get too ahead of ourselves, let us not forget the P3 1.13GHz, Itanium, BTX and Netburst, which is pretty much one long string of f*ck ups from the guys at Intel. Logic states that everybody will screw up. Its just how you bounce back.
And...
Quote:
Agreed.

This is a rough patch AMD is going through now.
The second generation 65nm should be be better.

Ever since Intel's Core 2 came out, we all more or less knew that we would have to wait until Barcelona to get a large increase in performance. This is just proving that we were right.

I want everyone to remember this:
Intel was lagging behind AMD ever since the first Athlon 64 CPUs came out, for years... from September 23, 2003 (First Athlon 64) to July 27, 2006

AMD was leading Intel until:
The first wave of Core 2 processors was officially released on July 27, 2006

Intel has been leading for LESS THAN 5 months.

Barcelona will be released in 2H '07.
We have yet to see the performance of this chip, but we should assume it will be quite a lot better than the current AMD chips.
Putting 2 x 4 cores in a 4x4 setup = 8 cores, this should at least be competitive, but will probably take the performance crown.

Oh by the way... before the first Athlon, AMD was behind Intel ever since it was established in 1969. AMD is still here. It will continue to be here. Even if AMD never get the performance crown again (I doubt it), AMD can offer better price/performance ratios, and compete in the budget/midrange market.

Don't write AMD off so soon....


Good way to keep this whole thread in context and give the discussion the proper perspective. You both make good points.

I'm anticipating some interesting developments from AMD and look forward to the back and forth between both companies. I just wish I had enough money to build both Intel and AMD based machines to really enjoy the technology.To get a large increase....yes. To get a decrease.....i don't think so!! :oops: 
December 21, 2006 5:29:12 PM

Wombat2 was talking about per clock performance. You responded with 70%, when AMD has stated 40% general gains. They also claim 40% FP gain, so no, just because you said 70%[ambiguous], it's hard to distinguish to understand whether your 70% is per clock gains or gains from doubling the cores.
!