AM2 or Conroe?

[AmdInside]

Does anyone know when a Conroe variant will come out for mobile? I am more interested in upgrading my laptop than desktop since my laptop has a Pentium M 1.6GHz and its starting to show its age in gaming and Photoshop.

 

[Oyeve]

Damn! And I just switched to AMD after over 20 years with Intel!

 

[Googer]

Originally posted by: DrMrLordX
Conroe, AM2 chips will be the same architecture as s939 chips with a new memory controller and DDR2. Expect maybe a 5% overall performance increase on AM2, or better yet, just read Anand's preview.

As things stand, it looks like similarly-priced Conroe parts will have a 17-20% performance advantage before overclocking is taken into account.

If Conroe forces a price crash on AM2 parts, then things will get interesting. It looks like Intel is low-balling on their Conroe prices. Will they follow through?

Correct, AM2 will share the same CPU design as the socket 939 part but like I said, AMD will revamp the processor for 2007; so we should be getting new/modified architecure in about a year.

 

[DrMrLordX]

Originally posted by: Googer

Correct, AM2 will share the same CPU design as the socket 939 part but like I said, AMD will revamp the processor for 2007; so we should be getting new/modified architecure in about a year.

We don't have any guarentee that K8L will launch on AM2. It could launch on AM3. The register has floated a rumor that AM3 is being fast-tracked for a 2007 release.

It's too far in the future for us to know what's going to happen.

 

[Topweasel]

Originally posted by: DrMrLordX

Originally posted by: Chris2wire
Actually I'd go AM2.

The first chips won't be that great, but the whole point is to introduce a new socket with more memory capabilities. DOwn the line the new M2 socket will house the greatest CPUs, much better than anything intel will offer

And you know this how? Currently, only rumors and speculation support the notion that K8L will launch on AM2. Other rumors indicate that AM3 will launch alongside K8L in 2007.

AM2's future is not set in stone, and it could be a dead-end street.

You do not know what future CPUs will be on AM2. You don't even know IF there will be any future CPUs on AM2. Nor does anyone else.

It makes more sense to get a Conroe-based system when Conroe launches and keep an eye on AM2's future. If powerful K8L-based procs are released that will actually work on old AM2 boards (also not guarenteed), it would then make sense to ditch one's Conroe CPU and motherboard for a cheap AM2 board + K8L chip. That way, you get to make your decision on that which is known rather than making guesses based on rumors.

You make it seem like the 2-3 year periods of Socket by AMD is relitively small period of time. Intel made 6 sockets for its Pentium, 3 sockets and a slot with 4 overlapping chipset requirements for its P6 line. for the P4 3 sockets and 4 chipsets inlcudding the 950 requirement just to support DCs.

AMD has had what 5 (which includes the server only 940 just for old FX buyers) sockets over the last 11 years. Anyone that really wants to buy a new CPU for their 754 they can get the Turian64. During the Athlon days the the only new chipset requirement was for the jump to 333 and 400 MHz FSB. For the 939 and the 940 nothing new was required for any upgrades to this point.

I fail to see how the AM2 socket can be seen as a dead end street. Chances are if AM3 becomes available it will be just adding support for a different memory type and that AM2 and AM3 devlopment will be developed in parity till they reach a point where sales drop and any further work on AM2 will show little reward (like the 3700+ for 754).

Don't throw out AMDs history just because you have ants in your pants.

 

[Googer]

Originally posted by: DrMrLordX

Originally posted by: Googer

Correct, AM2 will share the same CPU design as the socket 939 part but like I said, AMD will revamp the processor for 2007; so we should be getting new/modified architecure in about a year.

We don't have any guarentee that K8L will launch on AM2. It could launch on AM3. The register has floated a rumor that AM3 is being fast-tracked for a 2007 release.

It's too far in the future for us to know what's going to happen.

I have not read any reports from the register, I got my information from an AMD engineer who said it in an interview. (lost the link)

 

[thxdd]

I'm so happy I didn't purchase a Socket 939 dual core system like I had planned. My trusty Socket A system has proved it's worth and will easily tide me over until Conroe debuts. I'm glad to see Intel is making things interesting again.

As for the topic, it's a misleading question since you're comparing a Socket/Chipset to a Processor. Two very different things. I think most people are misunderstanding what Socket AM2 is all about. The current CPUs from AMD that fit in Socket AM2 are intended to be a transition product. We'll see processors coming down the line that take advantage of AM2 sooner than later. Currently the processors that are being tested by review sites like Anandtech don't need the increased memory bandwidth that AM2 offers.

As for when Conroe debuts, I don't have a doubt in my mind it will be the processor to buy assuming Intel prices it right. Hopefully this will push AMD into selling cheaper processors again since they will be the ones playing catch up. I see it as win-win for the consumer.

I'm already impressed with Intel's new chips, the Core Duo @ 1.66 in my new Mac Mini is an absolute screamer, especially considering it's intended for the mobile/SFF markets. I can't wait to see Conroe in action.

 

[Fullmetal Chocobo]

I'll be doing the research to see how much of a performance boost I would get after converting to a dual Woodcrest machine (currently dual Xeon 2.66). And it would likely be an entire platform upgrade, so it would be expensive. No idea when they come out, but it won't be for a while, so I'm not worried.

 

[DrMrLordX]

Originally posted by: Topweasel

Don't throw out AMDs history just because you have ants in your pants.

I'm not. What makes AM2 special is that it's the first time in AMD's history that they're stuck with a memory technology that does not play along well with their CPUs. Or, at least, not their current CPUs (yes, I read Anand's review; DDR2-533 and DDR2-400 on AM2 isn't impressive). Intel had a similar problem with the marriage of the Pentium III and RDRAM. It worked okay, but not well enough to justify the expense or abandonment of PC100 and PC133. They had to release an entirely different sort of processor (the Pentium IV) to justify their use of RDRAM.

If AMD plans to stick with AM2 for an extended period of time, that certainly is not good news unless K8L prives to be an "entirely different sort of processor" that can and will play along well with DDR2. That's something I can't predict. What I can predict, or at least will predict anyway, is that K8L will probably do better on socket AM3 using DDR3.

I see no reason to invest in a socket that AMD would likely be insane not to abandon ASAP. If K8L proves me wrong and can thrive on AM2, more power to AMD, but I'm not holding my breath waiting for it.

 

[DrMrLordX]

Originally posted by: Googer

I have not read any reports from the register, I got my information from an AMD engineer who said it in an interview. (lost the link)

I believe you. Turns out it wasn't the Register but the Inq, my mistake.

Here's the link I was talking about:

http://www.theinquirer.net/?article=30617

Take it for what it's worth. An AMD engineer would, or at least should, have better information unless the Inq is also getting their information from AMD engineers.

 

[Topweasel]

Originally posted by: DrMrLordX

Originally posted by: Topweasel

Don't throw out AMDs history just because you have ants in your pants.

I'm not. What makes AM2 special is that it's the first time in AMD's history that they're stuck with a memory technology that does not play along well with their CPUs. Or, at least, not their current CPUs (yes, I read Anand's review; DDR2-533 and DDR2-400 on AM2 isn't impressive). Intel had a similar problem with the marriage of the Pentium III and RDRAM. It worked okay, but not well enough to justify the expense or abandonment of PC100 and PC133. They had to release an entirely different sort of processor (the Pentium IV) to justify their use of RDRAM.

If AMD plans to stick with AM2 for an extended period of time, that certainly is not good news unless K8L prives to be an "entirely different sort of processor" that can and will play along well with DDR2. That's something I can't predict. What I can predict, or at least will predict anyway, is that K8L will probably do better on socket AM3 using DDR3.

I see no reason to invest in a socket that AMD would likely be insane not to abandon ASAP. If K8L proves me wrong and can thrive on AM2, more power to AMD, but I'm not holding my breath waiting for it.

For starters Intel only switched to Rambus on the P3 because they wanted the P4 to be only available with Rambus, therefore they need to make sure poduction was ramped up prior to the release of the P4. Also Rambus P3 was the fastest Intel performer at the time and only the price of rambus and the limitation to 2 slots held it back.

AMD on the other hand has made switches to memory when its prudent and would only do this if it was needed. Right now DDRII is cheaper and with DDR 633 and up the latencies are down (to almost DDR levels) at higher speeds of more bandwidth and the responce time is nearly the same if not more in favor of DDRII. This is why AMD is making the switch.

Think about it the Downside to the On die memory controller is that you basically are forcing someone to a certain type of memory, if that memory is costly then you basically are pushing yourself into a corner. The last thing AMD would do when they are trying to grab market share is eliminate their prospective clients by forcing them into a more expensive memory type. Think of how long they have been using DDR, almost 6 years now. Do you really think they will jump the gun on DDR3 or any thing else before Intel has time to test the waters. AM2 is going to be the Socket of choice till damn near 2008 at the minimum, and I firmly believe that it will be late 2008 before they stop devloping new DC and SC cpus for that socket.

AMD has built up their success off of playing it smart, not trying to force any new techs, and letting the market unfold a bit before making their move. It will be a while before they have enough market share to force the issue and untill they kick Intel down a couple more pegs it would be retarded of them to play the Market Gorilla on anybody. Sorry but your Ideas don't really seem to flow with me really well.

 

[One43637]

from early impressions, i'm going with Cornrow. good thing i built a decent S939 system 2 months ago instead of waiting for AM2.

 

[Shadow Conception]

Poll added. Go ahead and vote, if not, go away.

 

[DrMrLordX]

Originally posted by: Topweasel

For starters Intel only switched to Rambus on the P3 because they wanted the P4 to be only available with Rambus, therefore they need to make sure poduction was ramped up prior to the release of the P4. Also Rambus P3 was the fastest Intel performer at the time and only the price of rambus and the limitation to 2 slots held it back.

Intel switched to RDRAM as a legal-yet-anticompetitive maneuver to use their overwhelming market share to compel RAM manufacturers to switch to RDRAM. Since Intel had brokered an exclusive agreement with RAMBUS, Inc. that would have prohibited AMD from producing any memory controllers that would handle RDRAM, it would have put AMD in quite a bind had Intel's little game worked. Too bad for Intel that most RAM manufacturers did not like the idea of mass-producing non-JEDEC standard RAM.

AMD on the other hand has made switches to memory when its prudent and would only do this if it was needed. Right now DDRII is cheaper and with DDR 633 and up the latencies are down (to almost DDR levels) at higher speeds of more bandwidth and the responce time is nearly the same if not more in favor of DDRII. This is why AMD is making the switch.

The only thing prudent about AMD's switch to DDR2 is the reality that RAM manufacturers are abandoning DDR. DDR2 is cheaper to produce, and since Intel already switched to it, AND it's a JEDEC-standard RAM, it only made sense for RAM manufacturers to back off from the aging DDR standard.

I suspect that AMD would have been better off never using DDR2 at all, but getting reliable supplies of cheap DDR550 or DDR600 to keep them going until DDR3 became mainstream simply wasn't going to happen.

Think about it the Downside to the On die memory controller is that you basically are forcing someone to a certain type of memory, if that memory is costly then you basically are pushing yourself into a corner. The last thing AMD would do when they are trying to grab market share is eliminate their prospective clients by forcing them into a more expensive memory type. Think of how long they have been using DDR, almost 6 years now. Do you really think they will jump the gun on DDR3 or any thing else before Intel has time to test the waters. AM2 is going to be the Socket of choice till damn near 2008 at the minimum, and I firmly believe that it will be late 2008 before they stop devloping new DC and SC cpus for that socket.

Sadly, if you want to get memory for AM2 that's going to make it faster than a s939 platform using relatively common DDR400, you need DDR2-667 with tight timings. That's enthusiast RAM, and that costs money. So by your logic, AMD is forcing people to buy a certain type of memory (enthusiast DDR2-800 most likely) and thereby pushing people into a corner with AM2. Though I guess buyers could get the currently prevelant DDR2-533 instead and suffer performance inferior to older s939 products.

AMD has built up their success off of playing it smart, not trying to force any new techs, and letting the market unfold a bit before making their move. It will be a while before they have enough market share to force the issue and untill they kick Intel down a couple more pegs it would be retarded of them to play the Market Gorilla on anybody. Sorry but your Ideas don't really seem to flow with me really well.

Whether or not AMD is forcing tech on others isn't the point. The point is that DDR2 just doesn't make sense for their product lineup. They MIGHT keep AM2 on for awhile as a second-tier chipset, but in order for them to regain the performance crown when K8L launches, they will need the best chip (K8L will be that, or at least AMD hopes it will) paired with the best platform. AM2 with DDR2 is not the best possible platform for K8L. AM3 with DDR3 will likely be the platform-of-choice for K8L, and AMD will want/need to get that combination on the market ASAP for early adopters and high-end buyers. If they want the performance crown back (and they do), that's what it will take.

If they do release K8L for AM2, I suspect it'll be in the form of lower-end chips.

 

[vtohthree]

I would want to wait for AM(3)(AMD's next socket to support DDR3)...but even then I would still prefer AM2, even with the performance benchmarks favoring Conroe, the onboard memory controllers with the Hyper Transport still rips. It's just my opinion but I'm not looking forward to either one, I'm more interested in the what's next.

I'd like to see Conroe with a different bus system...and I'd like to see AMD make more engineering improvements on their cpu's(if in fact conroe benches hold true).

I'm planning to skip the upcoming gen chips, build one last dual core S939 overclocked to a wishful 2.9ghz prime stable with 2gigs of ram OC'd, and pass through for 3 years(upgrading my video card along the way of course)

 

[TheArabian]

... AMD should just nuke Intel, it can be WWI of cpu manufacturers.. and after that they can have a silicon war. And AMD becomes the new super power

 

[JackBurton]

Originally posted by: Topweasel

Originally posted by: DrMrLordX

Originally posted by: Topweasel

Don't throw out AMDs history just because you have ants in your pants.

I'm not. What makes AM2 special is that it's the first time in AMD's history that they're stuck with a memory technology that does not play along well with their CPUs. Or, at least, not their current CPUs (yes, I read Anand's review; DDR2-533 and DDR2-400 on AM2 isn't impressive). Intel had a similar problem with the marriage of the Pentium III and RDRAM. It worked okay, but not well enough to justify the expense or abandonment of PC100 and PC133. They had to release an entirely different sort of processor (the Pentium IV) to justify their use of RDRAM.

If AMD plans to stick with AM2 for an extended period of time, that certainly is not good news unless K8L prives to be an "entirely different sort of processor" that can and will play along well with DDR2. That's something I can't predict. What I can predict, or at least will predict anyway, is that K8L will probably do better on socket AM3 using DDR3.

I see no reason to invest in a socket that AMD would likely be insane not to abandon ASAP. If K8L proves me wrong and can thrive on AM2, more power to AMD, but I'm not holding my breath waiting for it.

For starters Intel only switched to Rambus on the P3 because they wanted the P4 to be only available with Rambus, therefore they need to make sure poduction was ramped up prior to the release of the P4. Also Rambus P3 was the fastest Intel performer at the time and only the price of rambus and the limitation to 2 slots held it back.

There were no speed gains with RDRAM and the PIII. SDRAM/440BX vs RDRAM/i820 were vertually identical in performance but RDRAM PC800 went for $1000 per 128MB stick at that time making the 440BX a no brainer choice for a PIII platform. That started the whole anti-RDRAM movement. RDRAM was still very expensive compared to DDR when the P4 was released but it did offer a performance advantage, although some would argue that the performance did not justify the high cost.

That being said, Rambus has some REALLY nice technology ahead and I hope to see it soon.

 

[l Thomas l]

I doubt I'll get either, I'm just hoping prices drop on AMD chips, especially the Opteron.

 

[AlgaeEater]

As much as I would like to, I'm just really sitting on the fence on the whole issue. I've never been a "Get the latest and greatest" guy, I've always been the "Get the most optimal price/performance ratio" guy. I think Conroe looks like a real big winner here, but what's important to me (and maybe others) is that hopefully prices will drop significantly on older technology.

A whole year from now people may consider a Athlon 64 X2 chip ancient technology, but it's most likely what I'll be running.

 

[imported_UncivilizedAMD]

Originally posted by: AmdInside
Does anyone know when a Conroe variant will come out for mobile? I am more interested in upgrading my laptop than desktop since my laptop has a Pentium M 1.6GHz and its starting to show its age in gaming and Photoshop.

Mobile Conroe = Merom
It's supposed to be released in August.
http://www.reghardware.co.uk/2006/05/08/intel_core_2_brand/

There's reviews of Merom performance over at xtremesystems.org forums already.

 

[bluemax]

[sitting on hands] I will not order now... I WILL wait two more months.... I will not order now... I WILL wait two more months.... [rocking on haunches]

[/crazy]

 

[fireontheway]

I think i will go with Intel next round.

 

[gxsaurav]

My next upgrade will be after vista & it's going to be either an AMD Athlon64 3800+ energy efficient CPU or the Conroe E6400 (2.13 GHz), this should be enough for my needs, which ever is cheap, combined with 1 GB DDR2 RAM, i don't find a reason to argue, just buy whatever is cheap & efficient, best bang for the buck

 

[meksta]

I agree whoever has the best bang for the dollar. I highly suspect that the 2mb cache conroe's will be much slower than the 4mb ones (ie. E6xxx).

I also suspect that the amd X2 prices will come down somewhat from their current levels when conroe launches, because the X2 prices haven't moved much at all since they first came to market.

I think that ddr2-800 + nf5 <= ddr2-667 + i975x in price....

 

[soydios]