Intel To Eclipse A64 For Benchmarking ?

[perkam]

Intel seems to be in the top spots of all futuremark benchmarks, be it Yonah, Conroe, Merom or Dothan.

Does this spell the end for A64's for 3d benching ? Or will it make a comeback ?

With all the news of Conroe breaking 60k 01 and what not, how long can the enthusiast market wait for AMD's response before they switch to conroe ?

Perkam

[Shadowmage]

AMD's usually a few months behind on new stuff (architecture, process, etc), so I don't see this being a big deal.

[nn_step]

Actually AMD's response is going to arrive shortly after Conroe hits market..
but none the less Intel will take the lead.. but that is a good thing.. the transfer of power benefits us the Buyers..
Heck thanks to A64, I got a Killer deal on a Photoshop rig because of the drastic price cuts made by Intel.

[Pinnacle]

This isnt News, and Quite Frankly, its annoying already, what do you know of what the FX-64 will be capable of?

Please everyone stop with the AMD bashing already

[nn_step]

This isnt News, and Quite Frankly, its annoying already, what do you know of what the FX-64 will be capable of?

Please everyone stop with the AMD bashing already
Completely agree.. what makes everyone think that FX-64 isn't going to be K8L?

[fhpchris]

A general thread, but nonetheless a news one.

It seems the Conroe is ready to eclipse the A64's every milestone (60K being broken here and there with Conroe) and AMD's answer is 6 months away...

No FX-64 is going to save AMD from this...

Perkam

Every milestone?

Look, in the server arena, AMD is still holding strong.

In the desktop market, where 90% of buyers do not know the difference between Netburst and the K7 and K8 to begin with, AMD does not even need to be faster to make money.

However, where it matters is in the server arena, where performance matters.

K8L is going to put a serious beatdown on Core in the server arena, and even the antiquated, "slow" A64 is hurting intel.

Even Dell is selling A64 Servers now! ROFL!

[3NZ0]

This isnt News, and Quite Frankly, its annoying already, what do you know of what the FX-64 will be capable of?

Please everyone stop with the AMD bashing already

agreed, im getting tired of it too. :)

[Teflon]

This isnt News, and Quite Frankly, its annoying already, what do you know of what the FX-64 will be capable of?

Please everyone stop with the AMD bashing already
A big QFT here, we are all quite impressed by conroe's performance, I myself will be getting one, but a news thread like this is more "Inquirer" than XS

[keiths]

Completely agree.. what makes everyone think that FX-64 isn't going to be K8L?

If the FX-64 is slated for Q1 of 07, it could be K8L. Is that what you are defining as shortly after?

[keiths]

FX-64 is supposed to be 3 GHz dual core; if not a K8L, a plethora of examples of exactly what to expect here on xtremes of a64s clocked at 3 GHz and more. If a K8L, we'll find out next year.

[zakelwe]

Whatever happens AMD will have to drop it's prices on the desktop to sell anything so that looks like it's going to be good for the buyer:)

Regards

Andy

[perkam]

I asked everyone's opinion on whether they think AMD will make a comeback...

NOT whether this thread should be in news. Fact is, the news audience and the news posters do not visit wamps or at least not the majority of the people who view the news, hence I put this thread here.

Now, what I'm NOT going to do is delete those nice posts saying "This isn't news because I happen to know exactly what news is" ... but I did delete those ones calling me a troll...you're lucky I don't give warnings out that easily.

I have re-worded the first post to make it more specific...it's about benchmarking now with recent Conroe ORB Blitz.

Perkam

[ozzimark]

Whatever happens AMD will have to drop it's prices on the desktop to sell anything so that looks like it's going to be good for the buyer:)

Regards

Andy
indeed. low prices are always cool :toast:

somehow i wouldn't be surprised if K8L gets a surprise early introduction, but i'm not holding my breath for it.

[nn_step]

Does anyone know if AMD has pressed 1st silicon for K8L?
Yes they have, about the same time Core was announced

[Pinnacle]

Umm...this is getting funny now. HEHE

If you think this is News, the questions your asking thats AWSOME!!

[Drunner611]

nn_step, do you think k8l would be moved to as early as September, or weren't you thinking of that big of a move up?

[nn_step]

nn_step, do you think k8l would be moved to as early as September, or weren't you thinking of that big of a move up?
K8L has been release ready for a little while now.. but they are tweaking it while they wait to see what Intel is going to release.

[largon]

Don't know the answer to the topic, but...

One thing is clear:
The die pictured in this article (http://tweakers.net/nieuws/41908/AMD-toont-45nm-wafer-levert-eerste-90nm-chips-uit-Fab-36.html) is something very much like the core AMD showed off (http://www.theinquirer.net/?article=31761) few days ago (http://news.com.com//i/ne/p/2006/amdslide4_532x402.jpg). It's not the exact same as the QC core, but it's quite obvious we're looking at a totally refreshed K8 design including 4-way execution - just like on the QC - and an additional cache level. Then there's the part in the right edge of the die that exists also in the middle of the QC die right next to the IMC - Intracore/cache link?

The QC part clearly features a HUGE FPU unit, unlike the other core (the singlecore shot). The pictured QC die diagram also seems to suggest L2 capacity for the QC part is 512kB/core (2*4*4 cells) compared to 2*4*8 cells for the SC core.

Check your favorite AMD core roadmap and see that there's no new desktop cores between Brisbane and Greyhound. And since the latter is a QC part, it's not hard to guess what Brisbane really is...

[ozzimark]

irer.net/?article=31761"]showed off[/URL] few days ago (http://news.com.com//i/ne/p/2006/amdslide4_532x402.jpg). It's not the exact same as the QC core, but it's quite obvious we're looking at a totally refreshed K8 design including 4-way execution - just like on the QC - and an additional cache level. Then there's the part in the right edge of the die that exists also in the middle of the QC die right next to the IMC - Intracore/cache link?
it actually looks like they just doubled the FPU :p:
http://www.chip-architect.com/news/Quad_vs_Dual_.jpg

take a close look at the Fetch/Scan/Align part
here.. there's 3 big "blocks"
http://www.amd.com/us-en/assets/content_type/DigitalMedia/27382B-12_E.jpg
same applies to the picture of revE

but in the new die plot, there's 4 of them. (same for the picture at tweakers.net)
perhaps it really will be a 4 issue core?

[kl0012]

it actually looks like they just doubled the FPU :p:
http://www.chip-architect.com/news/Quad_vs_Dual_.jpg

take a close look at the Fetch/Scan/Align part
here.. there's 3 big "blocks"
http://www.amd.com/us-en/assets/content_type/DigitalMedia/27382B-12_E.jpg
same applies to the picture of revE

but in the new die plot, there's 4 of them. (same for the picture at tweakers.net)
perhaps it really will be a 4 issue core?

I doubt if AMD have any silicon of K8L. They shows some die shots which is looks like some photoshop work. Just compare this picture:
http://www.chip-architect.com/news/Quad_vs_Dual_.jpg
to this one:
http://news.com.com//i/ne/p/2006/amdslide4_532x402.jpg
and focus your attention on the orientation of the cores relatively to x-bar.

[largon]

kl0012,
I agree. The QC diagrams are very likely photochopped - or whatever imaging prog AMD happened to use, since it is a QC core diagram and not a real core shot.

There (http://crew.tweakers.net/Wouter/AMD65nm.jpg) you go. This pic isn't a photochopped diagram but a real SC K8L.

[Absolute_0]

If you want to be optimistic about K8L, assume they tweak 20% performance increase out of it -- putting it neck and neck with Conroe at the same clocks.

To compete in the WR benchmark arena, it would still need to bench well past 4 Ghz and that's highly unlikely, especially with the horrendous coldbug.

One of the main reasons Intel can take benches if because you can't put LN2 on an AMD 64.

AMD will make a comeback, but i think it's the "K9" not K8L. And let's not forget that Intel will be tweaking Conroe at the same time. I'm not worried about AMD, i know they can handle themselves, that much is certain. I think the next 15 months are going to be a lot of fun.

[fhpchris]

You need to remember that LN2 can be any temperature, and not just the coldest ones :)

[nn_step]

If you want to be optimistic about K8L, assume they tweak 20% performance increase out of it -- putting it neck and neck with Conroe at the same clocks.

To compete in the WR benchmark arena, it would still need to bench well past 4 Ghz and that's highly unlikely, especially with the horrendous coldbug.

One of the main reasons Intel can take benches if because you can't put LN2 on an AMD 64.

AMD will make a comeback, but i think it's the "K9" not K8L. And let's not forget that Intel will be tweaking Conroe at the same time. I'm not worried about AMD, i know they can handle themselves, that much is certain. I think the next 15 months are going to be a lot of fun.
I'm just going to come out and say 40% performance increase is going to be conservative for K8L

[onewingedangel]

k8l will be about 6 months behind conroe though, perhaps more, as we know k8l will be used for opterons to compete with woodcrest, but who knows when k8l will migrate to the desktop?

By which time conroe will have had a nice speedboost and dual die quad cores will be avaliable.

Given AMD gives everything on the early side of expectations, and have desktop k8l out towards the end of the year, or early next year, they will still get only a year tops before intel move to 45nm, get a speedboost, true quad cores, and possibly CSI.

K8L should stop conroe from decimating the A64, but conroes not really its competition, as k8L's liable to be priced a good deal higher than conroe, and targetted only a servers and FX's for a good while.

CSI 3 should really help in ever larger systems however, but I know I'm not going to be getting a 32 socket sytem, so Core2's weakness in scaling to large ammounts of sockets doesn't really bother me.

[Absolute_0]

I'm just going to come out and say 40% performance increase is going to be conservative for K8L

Alright -- but i am holding you to your word :)

[DilTech]

k8l will be about 6 months behind conroe though, perhaps more, as we know k8l will be used for opterons to compete with woodcrest, but who knows when k8l will migrate to the desktop?

By which time conroe will have had a nice speedboost and dual die quad cores will be avaliable.

Given AMD gives everything on the early side of expectations, and have desktop k8l out towards the end of the year, or early next year, they will still get only a year tops before intel move to 45nm, get a speedboost, true quad cores, and possibly CSI.

K8L should stop conroe from decimating the A64, but conroes not really its competition, as k8L's liable to be priced a good deal higher than conroe, and targetted only a servers and FX's for a good while.

CSI 3 should really help in ever larger systems however, but I know I'm not going to be getting a 32 socket sytem, so Core2's weakness in scaling to large ammounts of sockets doesn't really bother me.

CSI has been cancelled for quite some time.

[onewingedangel]

CSI is still there, its implementation on the old whitefield was cancelled, but the csi project continues, and tukwilla will use it

[DilTech]

http://www.theinquirer.net/?article=30710

Until Intel comes out with a coherent whole, basically a Hypertransport ecosystem for Intel parts, it will continue to have problems. It has cancelled at least half a dozen projects including CSI, its version of Hypertransport, in the last few years.

CSI wasn't just removed from whitefield, it's been canned as a whole. It wasn't ever even up to par with the original HTT found on socket 754.

[the]

yea k9 slipped beeath the waves a long time ago, and k8l is its replacedment, k10 will be the followup, but i;ve heard a few skeptics of its ability to be produced.

[onewingedangel]

Tukwilla still has CSI, as will the upcoming core and xeon processors with IMC - possibly coming with the 45nm nehelam architecture. Quite what the current implementation of CSI consists of is up for debate though. It would probably be better for intel to use HT3, or a variant thereof.

[nn_step]

Tukwilla still has CSI, as will the upcoming core and xeon processors with IMC - possibly coming with the 45nm nehelam architecture. Quite what the current implementation of CSI consists of is up for debate though. It would probably be better for intel to use HT3, or a variant thereof.
I'll believe it when I see it..

[perkam]

Well it seems we'll be waiting for the next IDF or something for better roadmaps. There is a LOT of info out there about Conroe, Woodcrest and Montecito but none at one place.

It wasn't a fruitless search though...I had no idea woodcrest was server version of conroe :p:

Perkam

[Fred_Pohl]

I don't understand why AMD fans are so defensive about the obvious fact that Intel will soon be overtaking AMD in almost all areas of x86 cpu performance with the only possible exception being 4+ socket servers. It was bound to happen eventually. AMD had a good run with K8 vs Prescott but Core 2 Duo is simply a far superior new x86 architecture to the venerable old K8.

I know there's a lot of hope that K8L will arrive way ahead of schedule and miraculously turn the tables, just like there was a lot of hope that AM2 would be enough to fend off Core 2 Duo. I seriously doubt that K8L will arrive in Q1-07 with Brisbane and even if it does, the most it is likely to do is even things up.

Have you guys noticed how much AMD is charging for dual cores today? How about Intel's projected Core 2 Duo prices? I fail to see why there is so much hope that AMD can remain ahead of Intel, possibly enabling them to keep their prices exhorbitantly high. Do you really want to pay $1236 for a FX62? I don't. I'll take a $316 E6600 instead and consider the FX62 when it offers the same performance for the same price.

[nn_step]

I don't understand why AMD fans are so defensive about the obvious fact that Intel will soon be overtaking AMD in almost all areas of x86 cpu performance with the only possible exception being 4+ socket servers. It was bound to happen eventually. AMD had a good run with K8 vs Prescott but Core 2 Duo is simply a far superior new x86 architecture to the venerable old K8.

I know there's a lot of hope that K8L will arrive way ahead of schedule and miraculously turn the tables, just like there was a lot of hope that AM2 would be enough to fend off Core 2 Duo. I seriously doubt that K8L will arrive in Q1-07 with Brisbane and even if it does, the most it is likely to do is even things up.

Have you guys noticed how much AMD is charging for dual cores today? How about Intel's projected Core 2 Duo prices? I fail to see why there is so much hope that AMD can remain ahead of Intel, possibly enabling them to keep their prices exhorbitantly high. Do you really want to pay $1236 for a FX62? I don't. I'll take a $316 E6600 instead and consider the FX62 when it offers the same performance for the same price.
Do you honestly expect for AMD to keep the prices high and not drop them and for Intel to keep the prices low if Intel does take the performance crown back :fact: . It isn't going to happen.. Intel prices will soar and AMD prices will plummet, and the cost advantage that you think Intel has will vanish faster than a donut at a weight watcher's convention

[perkam]

Do you honestly expect for AMD to keep the prices high and not drop them and for Intel to keep the prices low if Intel does take the performance crown back :fact: . It isn't going to happen.. Intel prices will soar and AMD prices will plummet, and the cost advantage that you think Intel has will vanish faster than a donut at a weight watcher's convention Remember that Intel can play the pricing game better than AMD. Intel's the manufacturing and financial powerhouse here, not AMD, so let's put that cost advantage theory to rest.

Perkam

[Fred_Pohl]

K8L has been release ready for a little while now.. but they are tweaking it while they wait to see what Intel is going to release.

LOL! Got link?

[nn_step]

Double poster :eek:

[Fred_Pohl]

AMD will make a comeback, but i think it's the "K9" not K8L. And let's not forget that Intel will be tweaking Conroe at the same time. I'm not worried about AMD, i know they can handle themselves, that much is certain. I think the next 15 months are going to be a lot of fun.

AMD cancelled the K9 long ago and now the K10 appears to be either cancelled or indefinitely delayed.

http://www.theinquirer.net/?article=27421

[Fred_Pohl]

I'm just going to come out and say 40% performance increase is going to be conservative for K8L

40% eh? You're joking, right?

[perkam]

Fred Pohl you are hereby warned for double posting...two consecutive times...which makes it double double posting !? Look how you make me waste my precious brain cells !!! To the 3-day ban with you !!! **Click!**

:D

Perkam

[Fred_Pohl]

Do you honestly expect for AMD to keep the prices high and not drop them and for Intel to keep the prices low if Intel does take the performance crown back :fact: . It isn't going to happen.. Intel prices will soar and AMD prices will plummet, and the cost advantage that you think Intel has will vanish faster than a donut at a weight watcher's convention

No, I expect AMD will have no choice but to slash prices. I just don't see how they can cut FX62 from $1236 to match a $316 E6600. What makes you think that Intel's prices will soar? Scarcity due to overwhelming demand?

[perkam]

No, I expect AMD will have no choice but to slash prices. I just don't see how they can cut FX62 from $1236 to match a $316 E6600. What makes you think that Intel's prices will soar? Scarcity due to overwhelming demand?That's true...its going to have ample production capacity...

When are they releasing the Conroe again ?? :confused:

Perkam

[Fred_Pohl]

Fred Pohl you are hereby warned for double posting...two consecutive times...which makes it double double posting !? Look how you make me waste my precious brain cells !!! To the 3-day ban with you !!! **Click!**

:D

Perkam

No warning? No chance to review the TOS? I didn't even know that replying to 2 separate posts in succession was considered double posting.

When are they releasing the Conroe again ??

Not until July. :)

[perkam]

No warning? No chance to review the TOS? I didn't even know that replying to 2 separate posts in succession was considered double posting. Just kidding...though hopefully the K8L does startle Intel for competition's sake.

Perkam

[Fred_Pohl]

Just kidding...though hopefully the K8L does startle Intel for competition's sake.

Perkam

Knowing how censorship happy this place is, you had me worried there! :D

Although I find it impossible to believe, as others have speculated, that AMD is cleverly hiding a ready-to-launch 65nm K8L architecture behind a disappointing AM2 release, or that adding some of Core 2 Duo's features to K8 will give it a 40%+ performance boost, I don't believe that AMD will remain farther behind Intel than they were in the K6-2 days for very long. I look for AMD to make a comeback in 2008 with a new architecture.

[Pinnacle]

Knowing how censorship happy this place is

Least censorship place on the WWW

[Thorry]

How is this news?

[timpanogos]

m/b companies must just hate this ... all the way to the bank

[fhpchris]

Lets not get into pissing contests about what is "wrong" about A64. A64 has had the performance crown here for ~3 years, and many people here have been buying intel -- not because AMD sucks, but because intel makes good products! (and CHIPSETS!)

The Core architecture's L1 cache delivers about twice as much bandwidth (Measured by ScienceMark), while it's L2-cache is about 2.5 times faster than the Athlon 64/Opteron one.

Core has a kickass l1 and L2 cache.

So how do Intel's Core and AMD's Hammer compare when it comes to decoding? It is hard to say at the moment without access to Intel's optimization manuals. However, we can get a pretty good idea. In almost every situation, the Core architecture has the advantage. It can decode 4 x86 instructions per cycle, and sometimes 5 thanks to x86 fusion. AMD's Hammer can do only 3.

Core can decode x86 better/faster than Hammer can.

On K8, 128-bit SSE instructions are decoded into two separate 64-bit instructions. Each Athlon 64 SSE unit can only do one 64-bit instruction at a time, so the Core architecture has essentially at least 2 times the processing power here. With 64-bit FP, Core can do 4 Double Precision FP calculations per cycle, while the Athlon 64 can do 3.

When it comes to integer execution resources, the Core architecture is an improvement over the Pentium 4 and Dothan CPUs, and is at the same level (if we only look at the number of execution units) as the Athlon 64. The Athlon 64 seems to have a small advantage when it comes to calculating addresses: it has 3 AGU compared to Core's 2. This could give the Athlon 64 an advantage in some less common integer workloads such as decrypting algorithms. The deeper, more flexible (Memory disambiguation, see further) out of order buffers and bigger, faster L2-cache of the Core should negate this small advantage in most integer workloads.


Core does SSE faster than Hammer.

The P6 and P-M could already reorder Loads pretty good. They could move one Load before other Loads, as well as before Stores which have no unknown addresses or addresses which do not reference the same address as the load. In contrast, the Athlon 64 can only move loads before independent ALU operations (ADD etc.). Loads cannot be moved ahead much at all to minimize the effect of a cache miss, and other loads cannot be used to keep the CPU busy if a load has to wait for a store to finish. This means that the Athlon 64 processor is severely limited when it comes to reorder code.

This is probably one of the most important reasons why the Athlon 64 does not outperform the P-M in gaming and integer workloads despite having a lower latency memory system and more integer execution sources. Integer workloads tend to jump around in memory, and have many unknown addresses which must be calculated first. It is less important for FP intensive loads, which is also one of the reasons why the Athlon 64 had no problem with Dothan in this kind of workload. FP workloads access the memory in a much more regular fashion.

Once Loads and Stores are in the queues of Load/Store units, the Athlon's L/S unit allows Loads to bypass Stores, except of course when the load would bypass a store to the same address. Unfortunately, by then the Loads are already out of the ICU and cannot be used to fill the holes that dependencies and cache misses make. You could say that the Athlon (64) has some Load/Store reordering but it's much later in the pipeline and is less flexible than the P6, P-M, and Core architectures.


Memmory disambiguation helps the cpu keep working and not get into a cache miss.

Core has some very distinct advantages, but they are things that do not make A64 instantly obsolete. Intel has taken the largest disadvantages of netburst (FSB bandwith fighting memmory bandwith along with the inefficiencies and thermal leakage) and used new technology to reduce the impact of these on performance.

They did a kick ass job.

Now, amd needs to take the ball. AMD knows the problems and downsides to A64 and fix them or reduce the impact of them on performance -- some of the problems with it mentioned in the article even were in the K7!

The A64 scales well in the server enviroment, Core does not -- it is a compromise of server and desktop architectures. K8L will run over Core in the server arena, but what improvements will AMD finally give the AMD desktop guys? One completly misleading thing, the first post -- is just wrong. It was wrong before, and it is wrong now.

Yes. Yonah (it is retail now) spanks the 90nm A64 on the desktop segment. Conroe makes this gap larger. But you folks need not confuse bad architecture with architecture that has been proven. The A64 has drawbacks compared to Conroe, but it isnt Outdated or Old or Crappy.

Intel has stepped up to the plate. It should have happened YEARS ago, and it is something that will benifit everyone here -- AMD or Intel Fan alike.

[keiths]

For those curious for an analys of K8L's IPC improvements, check out real world tech and aceshardware forums. K8L's improvements are less extensive than that of core, like its memory ambiguation implementation; it's a simpler design. 40% or more is without foundation.

[Shadowmage]

Core will still be ahead in terms of IPC, since K8L matches Core in MOST areas (128bit SSE units, OoO loads, etc) but still lacks others (memory disambiguation, 256bit cache bus, super branch predictor).

However, K8L built with a higher-tech process (65nm SiGe DSL SMT SOI) which may mean that K8L will clock higher than Conroe. Let's not forget K8L still has the on-die memory controller.

K8L will beat Conroe if it clocks to 3.6ghz-ish, i'm guessing (assuming 3.33ghz Conroe)

[Teflon]

K8L will beat Conroe if it clocks to 3.6ghz-ish, i'm guessing (assuming 3.33ghz Conroe)
based on? :confused:

[MaxxxRacer]

no offense perk, but where's the news?

[Flexkill]

no offense perk, but where's the news?
lol, dang that sux:p: :slapass:

[Fred_Pohl]

K8L will run over Core in the server arena...

Great post but I think you should have qualified this statement to specify 4S+ server arena. I don't agree that K8L will run over Woodcrest in the 1 and 2S server segment. K8 certainly can't do it and it's a bit early to make such assumptions about K8L.

[Fred_Pohl]

Least censorship place on the WWW

Hardly. I know of a hardware forum where there has never been a single post deleted (except by the poster), a thread closed, a member banned or a word censored. It is completely 100% uncensored and there aren't even any mods.

Beat that. :D

[Flexkill]

Hardly. I know of a hardware forum where there has never been a single post deleted (except by the poster), a thread closed, a member banned or a word censored. It is completely 100% uncensored and there aren't even any mods.

Beat that. :D
Sounds like chaos...you can keep it.;)

[perkam]

no offense perk, but where's the news?I'll remove the thread once Conroe launches.

Right now we're in an interesting position where only a select few have access to Conroe compatible components while most people have just seen the performance.

Perkam

[situman]

Do you honestly expect for AMD to keep the prices high and not drop them and for Intel to keep the prices low if Intel does take the performance crown back :fact: . It isn't going to happen.. Intel prices will soar and AMD prices will plummet, and the cost advantage that you think Intel has will vanish faster than a donut at a weight watcher's convention

Everyone in the world knows AMD has craploads of debt. Everyone in the world knows AMD is strapped for cash a lot of times. Everyone in the world know AMD does not plan ahead because of poor management. Intel can basically drive AMD into the ground with substantially lower prices and a superior product. It costs AMD more to produce a cpu than Intel and AMD needs to keep a certain price point to make a profit. Intel wouldn't be so stupid as to jack up the prices to allow AMD to raise its prices.

Now imagine Intel cripples a few Conroes by 20% or so and calls it a Celeron and sells it for 40% cheaper than a full fledged conroe. WHere would AMD be in the next six months? Scary times for AMD, especially if K8L is delayed (which most likely will happen due to their past track records and especially so with a new manufacturing process and technologies).

[zakelwe]

Damn thread full of fence sitters .. :(

:D

I seem to recall a slide showing that the increase in performance from the new chemistry of the K8L chip was 15% so you'd assume that the bare minimum would be this and more likely a fair bit higher than that. Maybe the problem will be more to do with how high they can be stretched upwards along their 85nm lifetime of course ?

On the other hand Intel has another problem, a great new chip but probably cannot cope with demand unless they price it over the odds ( which is not good for the buyer ) and so that limits them and stops them making as much ground as they could before K8L.

I would say Conroe is a no brainer for the latter half of 2006, not sure about Woodcrest though, and come 2007 it would be a very brave person to bet indeed who will be top dog product wise rather than profits.

Regards
Andy

[DoubleZero]

AMD cancelled the K9 long ago and now the K10 appears to be either cancelled or indefinitely delayed.

http://www.theinquirer.net/?article=27421

I guess you've stop having problems with rumormill speculation... cheers for double standarts.

Everyone in the world knows AMD has craploads of debt. Everyone in the world knows AMD is strapped for cash a lot of times. Everyone in the world know AMD does not plan ahead because of poor management. Intel can basically drive AMD into the ground with substantially lower prices and a superior product. It costs AMD more to produce a cpu than Intel and AMD needs to keep a certain price point to make a profit. Intel wouldn't be so stupid as to jack up the prices to allow AMD to raise its prices.

/ironic mode on
I knew there was a reason for the Intel stock being so low and continue to drop and Amd stock being high and continue to rise.
Dumb investors for investing in crippled companies.
/ironic mode off

I'll remove the thread once Conroe launches.

I knew there was a reason for this thread

[dogsx2]

Least censorship place on the WWW

You got that right, how do you think this happened?


nn_step
DDR3 Guru

Join Date: Aug 2005
Location: Lansing, MI
Posts: 7,368

[perkam]

You got that right, how do you think this happened?


nn_step
DDR3 Guru

Join Date: Aug 2005
Location: Lansing, MI
Posts: 7,368 LOL :rofl:

Good One :p: But let's not start picking up on ol' nn again...he was right about k8L and beyond being ddr3 compatible ;)

Perkam

[Thorry]

Why is this topic in the News section?

[onewingedangel]

LOL :rofl:

Good One :p: But let's not start picking up on ol' nn again...he was right about k8L and beyond being ddr3 compatible ;)

Perkam

can we start calling me ddr4 guru, because I reckoned the athlon xp would use ddr4. As it turned out i was 100% wrong, yet seeing as it will eventually be used on some system, I still claim to be right....

[don_vercetti]

can we start calling me ddr4 guru, because I reckoned the athlon xp would use ddr4. As it turned out i was 100% wrong, yet seeing as it will eventually be used on some system, I still claim to be right....

In that case, i take DDR5 Guru. I just know it will be used sometime in the future on the K10..but AMD might change its name, just so y'all know now.


Sorry, enough bashing NN_step.

personally, i'm going to call Conroe over K8L. Just have that feeling in my gut..

[Pinnacle]

You got that right, how do you think this happened?


nn_step
DDR3 Guru

Join Date: Aug 2005
Location: Lansing, MI
Posts: 7,368

:D :p:

[nn_step]

LOL :rofl:

Good One :p: But let's not start picking up on ol' nn again...he was right about k8L and beyond being ddr3 compatible ;)

Perkam
Well it happens every day.. I'm always used to it.. But then again that is why I'm a Dimer but yet I don't post over in the Forum for it..

--Post edited for "observations"

[perkam]

[QUOTE=nn_step]Well it happens every day.. I'm always used to it.. But then again that is why I'm a Dimer but yet I don't post over in the Forum for it.. [QUOTE]Meow...lets leave the observations out of the thread nn...just a small detour there :)

Perkam

[nn_step]

Meow...lets leave the observations out of the thread nn...just a small detour there :)

Perkam
You forgot to remove the observations in the quote man ;)

[[XC]melymel]

Well it happens every day.. I'm always used to it.. But then again that is why I'm a Dimer but yet I don't post over in the Forum for it..


Oi carefull!! I think you got off on the wrong foot with Jeff over in the DIMES section which you dont wanna do :slap: quick edit :clap:

BTW Why are people such as nn saying that K8l will be the best without any hard evidence when in previous threads you have posted how people shouldn't judge conroe until it is released...

:toast:

[Absolute_0]

K8L will beat Conroe if it clocks to 3.6ghz-ish, i'm guessing (assuming 3.33ghz Conroe)

the 3.6 K8L will beat the 3.33 Conroe, but what about when that 3.33 Conroe is clocked to 4.0 on air? I need to see a whole lot more about K8L before i'm convinced that it will exceed Conroe.

[ozzimark]

the 3.6 K8L will beat the 3.33 Conroe, but what about when that 3.33 Conroe is clocked to 4.0 on air? I need to see a whole lot more about K8L before i'm convinced that it will exceed Conroe.
i'll be mighty impressed if the 3.33ghz conroes have that much headroom on air, considering that fugger had to resort to phase and disabling one core to bench at 3.9ghz ;)

[don_vercetti]

I'll be mighty impressed to see a 3.6ghz K8L on air. Currently the max on air is about 3.2ghz for 90nm K8, and that's for high end air and good chips. I'll be suprised if the shrink to 65nm enables them to clock a decent amount of chips to 3.6ghz, using stock air.

[DilTech]

I'll be mighty impressed to see a 3.6ghz K8L on air. Currently the max on air is about 3.2ghz for 90nm K8, and that's for high end air and good chips. I'll be suprised if the shrink to 65nm enables them to clock a decent amount of chips to 3.6ghz, using stock air.

I wouldn't...Especially if their heatsink is the same or better than their current opteron/high-end x2 cooler.

Look at it like this.
130nm AMD, 2.9 24/7 stable was phase territory
90nm(single core) AMD, 2.9-3.1 24/7 stable is obtainable on Air, 3.5 is now phase level
90nm(dual core) AMD, 2.6-3.0 air, 3.4-3.6 phase

Now, looking back, AMD generally seems to hit phase level when they do a die-shrink on the next chip. So I wouldn't be surprised to see 3.6, especially with their new silicon method inwhich they claim will either bring a 40% power drop, or 40% faster transistor switching.

[don_vercetti]

130nm A64's only did 2.9ghz on phase???

That's nonsense. Ricky and Pedro made 4ghz on clawhammer, hell athlon XP's used to make 3.4ghz with luck. Clawhammer generally did 3.6ghz or so on phase, and 3 on air if you were lucky. Now we get 3.2 on air if we're lucky with san diegos, and don't forget, most people don't even manage that. So i think the max we're likely to see from these things is 3.2ghz or so at first.

Besides, even if we assume that 40% faster transistor switching translates to 40% higher clockspeeds, we only get to 3.5ghz (assuming 2.8ghz is the current fastest).


Now, looking back, AMD generally seems to hit phase level when they do a die-shrink on the next chip. So I wouldn't be surprised to see 3.6, especially with their new silicon method inwhich they claim will either bring a 40% power drop, or 40% faster transistor switching.

ANY evidence for this at all? The only evidence for that is your "130nm = 2.9ghz" which is incorrect.

Let's see....Winchesters. They were generally lucky to get to 3ghz phased.

And the Thoroughbreds (first AMD 130nm chips). The T'Bred A was a terrible clocker, wasn't until the T'bred B that we started seeing high clocks. T'Bred A barely clocked past 2ghz did it, which the palomino managed as well. So if anything, the past demonstrates that the initial print of 65nm CPU's will be poor clockers.

[DilTech]

130nm A64's only did 2.9ghz on phase???

That's nonsense. Ricky and Pedro made 4ghz on clawhammer, hell athlon XP's used to make 3.4ghz with luck. Clawhammer generally did 3.6ghz or so on phase, and 3 on air if you were lucky. Now we get 3.2 on air if we're lucky with san diegos, and don't forget, most people don't even manage that. So i think the max we're likely to see from these things is 3.2ghz or so at first.

Besides, even if we assume that 40% faster transistor switching translates to 40% higher clockspeeds, we only get to 3.5ghz (assuming 2.8ghz is the current fastest).



ANY evidence for this at all? The only evidence for that is your "130nm = 2.9ghz" which is incorrect.

Let's see....Winchesters. They were generally lucky to get to 3ghz phased.

And the Thoroughbreds (first AMD 130nm chips). The T'Bred A was a terrible clocker, wasn't until the T'bred B that we started seeing high clocks. T'Bred A barely clocked past 2ghz did it, which the palomino managed as well. So if anything, the past demonstrates that the initial print of 65nm CPU's will be poor clockers.

I was talking about standard a64 130nm, not the FX-55 Clawhammer, which used SOI/DSL silicon methods compared to the initial winchester 90nm chips(which btw, 4 ghz was a screenshot on LN2 or cascade, I'm talking 24/7 P-95 stable). You were considered lucky to get to 3 ghz stable on phase.

90nm didn't recieve SOI/DSL until venice/san diego.

So yes, I do have my point. Again, remember, I don't care about bench stable, I'm talking 100% 24 hour prime 95 stable.