Oh yes, Core 2 is pretty damn amazing, but...

[racerboywonder]

That said...here's a thought I'd like to submit to the community. For those who don't need the latest and greatest immediately, wouldn't this be a good time to get an AM2 system? I say this in light of all the processing advancements AMD is gearing up for, in response (both directly and indirectly) to Intel...65nm process, K8L, K10 or whatever else they got under their sleeves.

I'm not an AMD fanboy...but yes, I feel some affinity towards AMD because I built my first system with an A64 CPU. I've been planning an X2 3800+ AM2 build for August, but I have to say looking at the E6300 OC'd benchmarks (I'm not a rich man...leave the $1000 CPUs to the super-enthusiasts) at Xbitlabs (http://www.xbitlabs.com/articles/cpu/display/core2duo-e6300.html) has got me seriously thinking about switching camps. It handily trumped an X2 3800 OC'd to 3.0 Ghz...that's damn impressive. I think we're pretty much going to find those results across most of the review sites.

As I implied however, AMD is not just sitting on its haunches...with that kind of thrashing from Intel's lowest offering, it can't afford to lay stagnant too long, price cuts or not (yes it makes more money in servers and supercomputers, but you can't ignore the general consumer, either). As we know, the "reign of superiority" ebbs and flows...so would building an AM2 based system now be a form of "future proofing" (AM3 compatible, remember) for the tech they're going to trump Intel with? That is, instead of buying Core 2 now, and then having to spend more money to change to an AMD mobo and[/i] use its all-conquering processors.

That's what I'm wondering...thoughts?

 

[Action_Man]

AMD Roadmap. Unimpressive indeed.

 

[Skidd]

Nice points there jack, Although AM2 would become a cheaper option which is a good thing, it doesnt mean its going to be the best option. Remember you can get a C2D clocked 400mhz slower than a Athlon 64 chip and still be on top.

 

[jap0nes]

i've read that AMD is planning on reducing their prices to a "competitive price-performance ratio". that means amd chips cannot perform as good as core 2, but they will be cheaper (hope). Then we'll get back to the k6, k6-2 era, when amd wasnt soo good, but was cheap.
good for us, now we have competition

- 65 nm, don't get your hopes up for several reasons

65 nm reduces price, which i think it will be what amd will compete from now on. They cant offer better performance, then they'll offer an ok performance for a low price until they come up with a better performer

 

[chunkymonster]

IMO, it's times like these that, for those of us who don't need the "latest and greatest", to make our upgrades and purchases...the "price wars" are when we can really get the best bang for the buck...let the price drops sink in and the market respond, then scour the sites for the best bargains on "last gen" hardware...for example, a Skt939 Opteron 175 with fast PC3200 and a PCIe gpu is just as viable and as high performing a solution as most systems today...no reason not to cash in on that or pass it up for the latest and greatest that ultimately, really only offer incremental gains...

 

[BaronMatrix]

Yeah, I am going to be upfront and give my interpretation of the situation with AMD.... I will likely get blasted here by the BM's, 9-inch and such but judging from their debating skills I can hold my own ....


You argument is ok Core is great (and it is impressive from the data we have seen tonight and over the last month or so).... but wait until AMD has an answer. Let's take this point by point as you list above:

- 65 nm, don't get your hopes up for several reasons. First, the raw parametrics for transitor performance shown to date is not that impressive for AMD's 65 nm process. That is, they are not achieving the performance gains necessary to drive clocks up very high. From the data we see now, AMD will need to get to 3.6 to 3.8 GHz on the rev G release to surpass, and this simply will not happen. Even their roadmap shows the fastest 65 nm part will only be 2.8 GHz by Q3 2007 (a year from now). Second, 65 nm is a process technology and the affect of process technology on raw performance in computational performance is by clock speed -- simply building a 2.4 GHz K8 on 65 nm gives you no extra performance over a 2.4 GHz K8 built on 90 nm. Power maybe, performance no. Finally, volume shipment, if they keep to schedule, will not commence until at least Jan or Feb 07, 6-8 months from now.

- Wait for K8L -- this is your most compelling argument, however, if you truly look at the 'enhancement list' for K8L there are more bandwidth items on the plate and fewer IPC items -- and it is IPC that will determine it. Nonetheless, AMD would be better off not falling into the same Intel trap that plagued Netburst (i.e. release a subferior product than your competitor and try to run with it). I am of the opinion that AMD would rather be late on K8L than release a product that does not take back the performance crown, so I am confident that K8L will be spectacular. The question really boils down to time -- and that is where it get's ambiguous. The Inquirer is certain it is coming in Q1 07, others say mid-2007, and rumors in Taiwan say 2008. If Q1, then you only need to wait 6-8 months, if mid 2007 you have about a year, and if 2008 then 1 1/2 years before the crown changes back. Question is, are you willing to wait that long? It is a gamble sorta -- that AMD will deliver timely. They do not have a good track record on timely delivery, even after this spectacular market share grab --- K8 was actually late, 65 nm is very late, and there are rumors that K8L will be late. My take is that K8L will not see the light of day until 2008.

- K10 -- isn't even on the roadmap as a name, this should not be within your scope at this point in my opinion.

- What else under their sleeve -- well, this is the second most compelling arugment to go down the AMD road -- there is the 4x4 for those with money to burn that will not beat Core 2 in many cases, but will be cool. The second is opening up the cHT license, again you will be spending extra money because that tech does not exists, so if you do AM2 now you will be bying AM2+ (i.e. AM2 + new socket) MBs anyway, you have really not saved or positioned an upgrade for anything.

- RHT is a myth. Though conceptually, theoretically doable just not soley in hardware -- do not expect, in my opinion, a magical bios flash to do RHT and get an instant 30% bump.

Ok, there you go.... my thoughts.

Jack

Blasting people with differeing opinions is the job of the Intel fanboys. I may disagree or present the opposite case but you guys AM, Spud, AND sometimes you, only want to hear how great Intel is.

Anyone who says otherwise is spreading FUD. Even though everytime I say something voila, it's in the news.

 

[Action_Man]

LMAO! You're such a dipsh!t.

Even though everytime I say something voila, it's in the news.

Thats classic. Probably your best yet.

 

[TechnologyCoordinator]

You can get AM2 on the hopes and dreams that something compatible and better will come out. Or you can get a technology that's (almost) currently out and has proven itself in preliminary benchmarks.

If I was building a system and had the choice between AM2 and Conroe, it'd be Conroe.

 

[Skidd]

LOL!!

This is BM Today
"You damn INTEL Fanboys blah blah blah (Obviously blinded by AMD)

This is BM 1 and half years from now (after finaly seeing the light but too late because K8L is soon to be released and kicks some major arse)
"You damn AMD Fanboys blah blah blah.

BaronMatrix
When you join the real world, let us know. Untill then all I see coming from you is Blah blah blah Im a big fat IDIOT...

 

[miecz]

I agree with the first poster. Just support AMD. Intel is too over priced. Plus, you support military adventures when you buy it - at least tacitly. Recent news...my sig...you know...

 

[Skidd]

I agree with the first poster. Just support AMD. Intel is too over priced. Plus, you support military adventures when you buy it - at least tacitly. Recent news...my sig...you know...

Where the F**k have you been the past couple of months?

Do some research before brain farting all over the place here.

In the words of Action Man

DIE NEWBIE DIE!!!!

 

[uk_gangsta]

I agree with the first poster. Just support AMD. Intel is too over priced. Plus, you support military adventures when you buy it - at least tacitly. Recent news...my sig...you know...

Intel is Al'Quieda's main infiltration in the US, LOL i thought everyone knew that..............

 

[The_Abyss]

Freak.

 

[MasterLee]

Things are going just fine. There is always a winner and a loser.
AMD has been on top, now it's Intels' turn, it took a while but it happened. AMD never says what it's doing, never has, and no one will know till it's out.
These are interesting times indeed for both companies.

 

[mjp1618]

RHT is a myth. Though conceptually, theoretically doable just not soley in hardware -- do not expect, in my opinion, a magical bios flash to do RHT and get an instant 30% bump.

Jack

Regarding RHT, interesting article by Jon Hannibal Stokes at ArsTechnica - Reverse hyperthreading? Um, no.

First off, there's no way this would work the way the author seems to think it would. How would the cores' pipelines support this in any phase of execution? In the fetch phase, there would have to be some arbitration mechanism whereby the two cores fetched alternate instruction blocks from the I-cache, thus distributing the instruction stream across two processors.

Then, once the instruction stream is fragmented inside the two cores, how are the register files kept in sync? If an add in one line of code writes its result to a register in one core, then how could a test instruction in the other core read that distant register to see if it needs to branch? Or how would out-of-order execution work across two cores? Would the instruction schedulers have their own separate bus to communicate over?

Anyway, it's not worth going into too much detail here, because it's kind of like asking how Superman could lift an entire continent up into space without it breaking apart, or how he manages to fly in the first place, and so on.

Stick a fork in this one, because it's done.

 

[mjp1618]

RHT is a myth. Though conceptually, theoretically doable just not soley in hardware -- do not expect, in my opinion, a magical bios flash to do RHT and get an instant 30% bump.

Jack

Regarding RHT, interesting article by Jon Hannibal Stokes at ArsTechnica - Reverse hyperthreading? Um, no.

First off, there's no way this would work the way the author seems to think it would. How would the cores' pipelines support this in any phase of execution? In the fetch phase, there would have to be some arbitration mechanism whereby the two cores fetched alternate instruction blocks from the I-cache, thus distributing the instruction stream across two processors.

Then, once the instruction stream is fragmented inside the two cores, how are the register files kept in sync? If an add in one line of code writes its result to a register in one core, then how could a test instruction in the other core read that distant register to see if it needs to branch? Or how would out-of-order execution work across two cores? Would the instruction schedulers have their own separate bus to communicate over?

Anyway, it's not worth going into too much detail here, because it's kind of like asking how Superman could lift an entire continent up into space without it breaking apart, or how he manages to fly in the first place, and so on.

Stick a fork in this one, because it's done.

Go here -- RHT in the concept.... long discussion, not worth arguing.

Mitosis has been demonstrated. Joset can fill you in.

Yes, I have read such discussions people like you and Joset have had... Just thought I would just post the link... Jon Hannibal Stokes seems to be a very knowledgeable person, so...

 

[jaguarskx]

I've been planning an X2 3800+ AM2 build for August,...

As we know, the "reign of superiority" ebbs and flows...so would building an AM2 based system now be a form of "future proofing" (AM3 compatible, remember) for the tech they're going to trump Intel with? That is, instead of buying Core 2 now, and then having to spend more money to change to an AMD mobo and[/i] use its all-conquering processors.

That's what I'm wondering...thoughts?

Assuming your current system is socket 939 and you have a PCI-e slot, then if I were you I would just upgrade to a S939 Athlon 64 X2 3800+. There is hardly any performance difference between socket AM2 and S939 of the same model. Plus you will save a lot of money. You can even opt for a faster X2 like to 4200+ or 4600+. Unless you want the X2 5000+ or FX-62 it is really pointless to switch to socket AM2.

On the other hand, if your current mobo only has AGP, then okay, upgrading to AM2 to get a PCI-e slot is an acceptable reason. I’m not sure how much the X2 3800+ will cost after AMD’s price cuts, but if the Conroe E6300 will only be $50 more, then it may be worth it to simply go with Conroe based on benchmark performance. With the exception of the motherboard, everything else you buy will be the same.

Getting back to AMD, by sticking with socket 939 you can save money (as I mentioned before), but you can also wait for AM2 to become mature enough to the point where motherboards are more or less stable. By the time K8L rolls out motherboards and RAM will be cheaper as well. By that time socket AM3 maybe ready, and perhaps the performance difference between DDR3 and DDR2 will be better than between DDR2 and DDR.

Based on what others have said the server version of K8L will be coming in Q1 2007, however that will be socket F not AM2. AMD may also decide to price them relatively high to discourage a lot of people from buying a CPU that is meant for servers. AMD has done so with the Opteron series since so many consumers were buying them. Also, server motherboards tend to be a bit expensive for most people’s tastes. You may need to wait until Q4 2007 to buy a desktop K8L.

Lastly, you are speculating on the performance capabilities of the K8L since it still exists only on paper. By next year Intel will be using the 45nm die process instead of 65nm. If the desktop K8L will not be released until 2008, then Intel's next CPU architecure, Nehalem, will be out. But that depends on if Intel can stick to it's aggressive roll-out schedule.

 

[racerboywonder]

jaguarskx, I'm actually running a 754 3000+...and for a long time I considered getting a 939 system so that I could save a little money (keeping my RAM), but I finally decided to go with AM2 for the better upgrade path. That choice had only been enforced by the looming price cuts.

Intial reports based on AMD documents said that the X2 3800+ will be priced at $169...not bad, until you realize:

a) I would probably only OC it to 2.5 or 2.6 Ghz max (sticking with air-cooling, liquid too much hassle);
b) Core 2 E6300 (1.86 GHz) will be priced at $183...and stock speeds, it beats a factory clocked X2 3800+ by a significant margin;
c) Xbitlabs overclocked the E6300 to 2.94 GHz pretty easily (in fact it was only limited to that by the motherboard), and that absolutely trounced an X2 3800+ at 3.0 Ghz by a very significant amount;
d) I play a lot of sims, which tend to depend on a faster CPU.

So unless AMD cuts the price even more, Core 2 is the price/performance leader hands down...and its TDP is 65W or less (for most, not the high end). I just keep having this nagging feeling that AMD has got something they're not showing just yet. Hell, I could be utterly wrong, but they've suprised us before. As I said though, if things don't change, I'm getting a lot more enthusiastic about getting an Intel for the next build.

 

[HYST3R]

That said...here's a thought I'd like to submit to the community. For those who don't need the latest and greatest immediately, wouldn't this be a good time to get an AM2 system? I say this in light of all the processing advancements AMD is gearing up for, in response (both directly and indirectly) to Intel...65nm process, K8L, K10 or whatever else they got under their sleeves.

I'm not an AMD fanboy...but yes, I feel some affinity towards AMD because I built my first system with an A64 CPU. I've been planning an X2 3800+ AM2 build for August, but I have to say looking at the E6300 OC'd benchmarks (I'm not a rich man...leave the $1000 CPUs to the super-enthusiasts) at Xbitlabs (http://www.xbitlabs.com/articles/cpu/display/core2duo-e6300.html) has got me seriously thinking about switching camps. It handily trumped an X2 3800 OC'd to 3.0 Ghz...that's damn impressive. I think we're pretty much going to find those results across most of the review sites.

As I implied however, AMD is not just sitting on its haunches...with that kind of thrashing from Intel's lowest offering, it can't afford to lay stagnant too long, price cuts or not (yes it makes more money in servers and supercomputers, but you can't ignore the general consumer, either). As we know, the "reign of superiority" ebbs and flows...so would building an AM2 based system now be a form of "future proofing" (AM3 compatible, remember) for the tech they're going to trump Intel with? That is, instead of buying Core 2 now, and then having to spend more money to change to an AMD mobo and[/i] use its all-conquering processors.

That's what I'm wondering...thoughts?

your right. its just a matter of fact of, how long do you want to wait. some people want a new setup now, and some will be patient enough to wait and see what happens.

 

[LordPope]

amd zone is claiming the 4x4 is seeing a 80% jump in performance in AMD 's labs right now.

your take is prett close to mines

4x4 is a wldcard..... it will obliterate CORE duo in a few things.... but the pirce will be nuts and again it will only be great in a few things....

however VISTA is much better at routing for miltiple cores than XP....

so where 4x4 may shine is in VISTA in again..certain apps

if they open the board up to cheap x2's ...and I have not seen a reason why they cant.... will the 4x4 not operate if it doesnt see a FX chip?

i will have to see that to believe that...

 

[LordPope]

45um will not be easy

 

[ivan_lee05]

45um will not be easy

i disagree...


intel has already shown its working 45nm sram chips.


and also, intel has already shown to us how fast they can do a transition from 90nm to 65nm..

and intel is on thier track.. "there is new architecture for every 2years."

 

[bixplus]

Blasting people with differeing opinions is the job of the Intel fanboys. I may disagree or present the opposite case but you guys AM, Spud, AND sometimes you, only want to hear how great Intel is.

Anyone who says otherwise is spreading FUD. Even though everytime I say something voila, it's in the news.

Go cry somewhere else BarrenBrainwaives. Everyone here knows how much of a loser you are.

 

[bixplus]

Plus, you support military adventures when you buy it(Intel) - at least tacitly

I would buy Intel if for this reason alone!!!

Hooorah!!!

 

[apache_lives]

heh your titles all wrong - you forgot the end bit, of is it thg did, it supposed to be "Oh yes, Core 2 is pretty damn amazing, but IM A FANBOY"

and BTW - u remember AMD's first s940, then 939 and the whole 754 mess? heh dont tell me future proof.

That said...here's a thought I'd like to submit to the community. For those who don't need the latest and greatest immediately, wouldn't this be a good time to get an AM2 system? I say this in light of all the processing advancements AMD is gearing up for, in response (both directly and indirectly) to Intel...65nm process, K8L, K10 or whatever else they got under their sleeves.

I'm not an AMD fanboy...but yes, I feel some affinity towards AMD because I built my first system with an A64 CPU. I've been planning an X2 3800+ AM2 build for August, but I have to say looking at the E6300 OC'd benchmarks (I'm not a rich man...leave the $1000 CPUs to the super-enthusiasts) at Xbitlabs (http://www.xbitlabs.com/articles/cpu/display/core2duo-e6300.html) has got me seriously thinking about switching camps. It handily trumped an X2 3800 OC'd to 3.0 Ghz...that's damn impressive. I think we're pretty much going to find those results across most of the review sites.

As I implied however, AMD is not just sitting on its haunches...with that kind of thrashing from Intel's lowest offering, it can't afford to lay stagnant too long, price cuts or not (yes it makes more money in servers and supercomputers, but you can't ignore the general consumer, either). As we know, the "reign of superiority" ebbs and flows...so would building an AM2 based system now be a form of "future proofing" (AM3 compatible, remember) for the tech they're going to trump Intel with? That is, instead of buying Core 2 now, and then having to spend more money to change to an AMD mobo and[/i] use its all-conquering processors.

That's what I'm wondering...thoughts?