Conroe vs. AM2: Memory & Performance

[Kougar]

Core 2 Duo (Conroe) launched about twelve days ago with a lot of fanfare. With the largest boost in real performance the industry has seen in almost a decade it is easy to understand the big splash Core 2 Duo has made in a very short time. AnandTech delivered an in-depth analysis of CPU performance in Intel's Core 2 Extreme & Core 2 Duo: The Empire Strikes Back (http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2795). With so much new and exciting information about Conroe's performance, it is easy to assume that since Core 2 Duo uses DDR2, just like NetBurst, then memory performance must therefore be very similar to the DDR2-based Intel NetBurst architecture.

Actually, nothing could be further from the truth. While the chipsets still include 975X and the new P965 and the CPU is still Socket T, the shorter pipes, 4 MB unified cache, intelligent look-ahead, and more work per clock cycle all contribute to Conroe exhibiting very different DDR2 memory behavior. It would be easy to say that Core 2 Duo is more like the AMD AM2, launched May 23rd, which now supports DDR2 memory as well. That would be a stretch, however, since AM2 uses an efficient on-processor memory controller, and the launch review found Core 2 Duo faster at the same clock speed than the current AM2. This is another way of saying Conroe is capable of doing more work per cycle - something we had been saying for several years about Athlon64 compared to NetBurst.

With the world now united behind DDR2, it is time to take a closer look at how DDR2 behaves on both the new Intel Core 2 Duo and the AMD AM2 platforms. The performance of both new DDR2 platforms will also be compared to NetBurst DDR2 performance, since the DDR2 NetBurst Architecture has been around for a couple of years and is familiar. We specifically want to know the measured latency of each new platform, how they compare in memory bandwidth, and the scaling of both Core 2 Duo and AM2 as we increase memory speed to DDR2-1067 and beyond. With this information and tests of the same memory on each platform, we hope to be able to answer whether memory test results on Conroe, for instance, will tell us how the memory will perform on AM2.

http://www.anandtech.com/memory/showdoc.aspx?i=2800 (http://www.anandtech.com/memory/showdoc.aspx?i=2800)

[vfrex]

The article touches upon AMD's switch to 65nm, but it doesn't really say much about what that will do. In theory, it should allow more room to grow in terms of clock speed, right?

[werty316]

Yep should help clock speeds and reduce power required and thermal rating (Thermal Design Power or Thermal Dissipation Power TDP).

[Kougar]

The article touches upon AMD's switch to 65nm, but it doesn't really say much about what that will do. In theory, it should allow more room to grow in terms of clock speed, right?

That is outside the scope of this article's purpose, which was memory latency and memory performance for both AM2 and Core 2 Duo platforms. ;)

But yes, smaller die will give AMD some headroom again to play with, although if you find a few good reviews they already extrapolate the clocks AMD would need to use to level the playing field against Conroe. AMD won't be able to play the old Intel game of ramping up clock speed, they need a new design before they can surpass Intel again in performance.

[vfrex]

I thought that their mentioning the switch to 65nm was out of place in the article without going into more detail. They say that AMD will make the transition from the bottom of their product line to the top. Does that mean Sempron first? It seems like it would make more sense to start at the top end first, so that they could push out higher clocked FX's to come closer in performance to the top end Duo Extremes.

I was just hoping that anandtech would have gone into more detail about the switch to 65nm, because I really haven't seen any discussion of it anywhere.

[Kougar]

Their mention of AMD's 65nm was to provide a little backdrop for the specific review, and then in the conclusion in how it would affect the results. Anandtech has written more about AMD's 65nm process in their other Core 2 Duo/AMD articles already as well.

And yes, it is likely they will first produce single core Athlons and Semprons, both of which qualify as the low end now I would think. When either Intel or AMD first start ramping a new fab process working CPU yields will be poor. Many of the chips won't even work or will work poorly so their only ability to sell these sub-par CPUs is to disable the faulty cache (Which turns them into Semprons) and clock them at lower speeds.

It's not a matter of them choosing to start out at the low end and work up, but just necessity of how the fabrication process works. Once they improve the 65nm fab process they can produce better quality yields and ramp up production.

[GIBSON]

While DDR2 Memory does not exhibit the same bandwidth or performance on the AM2 and Conroe platforms, they do perform at the same timings and voltages when going from one platform to another. This was clearly demonstrated in benchmarking tests performed on AM2 and Conroe platforms. This means readers can examine test results performed on a Core 2 Duo test bed with XYZ memory, and reasonably expect that XYZ memory to perform at the same speeds and the same memory timings and voltages on an AM2 platform - provided those settings are available.
^they mention this in the conclusion. Isn't this quite obvious. Why would memory rated at DDR800 perform at a different speed in an other system. Or need an other voltage to operate.

[Kougar]

Gibson, weren't you the guy that has memory that is rated for better timings on a Intel system as opposed to a AMD system? I know someone here mentioned that... ;)

To answer your question, the eact same memory would perform different depending on the quality of the memory controller. Now AMD integrates their memory controller onto their CPUs, and Intel lets the chipset makers build their own memory controllers into their respective chipsets. Not all memory controllers are created equal.

AMD only upgrades and improves the memory controller in stages because it is part of the CPU. Currently the Intel platform can already support DDR2 memory with CL2 timings. AMD's memory controller is not yet built for that low a latency. Also Intel systems are officially rated for DDR2-1067mhz speeds, but AMD's memory controller is again not yet designed to use DDR2-1067 memory... yet.

[GIBSON]

Gibson, weren't you the guy that has memory that is rated for better timings on a Intel system as opposed to a AMD system? I know someone here mentioned that... ;)

To answer your question, the eact same memory would perform different depending on the quality of the memory controller. Now AMD integrates their memory controller onto their CPUs, and Intel lets the chipset makers build their own memory controllers into their respective chipsets. Not all memory controllers are created equal.

AMD only upgrades and improves the memory controller in stages because it is part of the CPU. Currently the Intel platform can already support DDR2 memory with CL2 timings. AMD's memory controller is not yet built for that low a latency. Also Intel systems are officially rated for DDR2-1067mhz speeds, but AMD's memory controller is again not yet designed to use DDR2-1067 memory... yet.
Well, yes, I was the one. I never quite understood why though. And the current sticks that corsair has to offer don't seem to have that info any longer. Maybe an other revision. Mines are rev1.2 if i'm not mistaken. Thanx for the explaination though Kougar, that really clears things up for me ;)