Intel's Core 2 Extreme & Core 2 Duo: The Empire Strikes Back
by Anand Lal Shimpi on July 14, 2006 12:00 AM EST- Posted in
- CPUs
Memory Latency: No Integrated Memory Controller Necessary
One thing we pointed out in our earlier preview coverage of Intel's Core 2 Extreme is that the new processors have extremely low latency memory access, despite relying on the same memory controller as the previous generation of Intel CPUs.
Without an on-die memory controller Intel's Core 2 processor must use the memory controller in its chipset, which currently means the 128-bit DDR2 memory controller in either Intel's 965 or 975X chipset. The confusing thing is that although the Core 2 processors use the same memory controller as the old NetBurst processors, memory latency has been improved tremendously:
Intel's Core 2 processors now offer even quicker memory access than AMD's Athlon 64 X2, without resorting to an on-die memory controller. While Intel will eventually add one, the fact of the matter is that it's simply not necessary for competitive memory performance today thanks to Intel's revamped architecture. Update:As many astute readers have pointed out, Core 2's prefetchers are able to work their magic with ScienceMark 2.0, which results in the significant memory latency advantage over AMD's Athlon 64 FX-62. This advantage will not always exist; where it doesn't, AMD will continue to have lower latency memory access and where it does, Intel can gain performance advantages similar to what ScienceMark 2.0 shows.
| CPU | Everest |
CPU-Z 1.35 (8192KB, 128-byte stride) | CPU-Z 1.35 (8192KB, 64-byte stride) |
| AMD Athlon 64 FX-62 (2.8GHz) | 45.9 ns | 43.2 ns | 19.3 ns |
| Intel Core 2 Extreme X6800 (2.93GHz) | 59.8 ns | 52.8 ns | 10.9 ns |
Because of the Core 2's intelligent prefetchers doing too good of a job with ScienceMark memory latency test, we wanted to also showcase situations where Core 2 would feature slower memory accesses than the AMD platform with its integrated memory controller. Everest's results are more in line with what we'd expect to see, with the FX-62 offering over 23% faster memory accesses than the X6800. CPU-Z's latency tool also reported somewhat similar findings, with an 18% performance advantage due to AMD's integrated memory controller. CPU-Z also provided us with numbers that showcase how well Core 2 can perform if its prefetchers are able to "guess" correctly; at lower strides the Core 2 Extreme manages faster memory access than the FX-62. The 128-byte stride numbers are indicative of what will happen if the pre-fetchers are not able to get the Core 2 the data it needs, when it needs it, while the 64-byte numbers show you what can happen when things go well.
Facebook
Twitter
RSS
202 Comments
View All Comments
crystal clear - Saturday, July 15, 2006 - link
Just to remind all that-" Intel decides to release a B2 stepping of its Conroeprocessors.Also
BEWARE of Engineering samples & reviews based on Engineering samples inlcuded.
OcHungry - Saturday, July 15, 2006 - link
I don’t know if this Q been asked or answered (sorry no time reading 120 posts)Bu Mr. Anand, can I kindly ask you couple of concerns:
1) why don’t we see any reference to temp. under load? Temp. is a crucial factor in deciding wether or not I buy conroe since I live in a very hot climate.
2)A lot people make reference to 64bit and window vista and that conroe is a 32bit architecture and will not perform as good in 64bit. Is it true? can we have some 64bit benchmarks? It would be great to do this test in multitasking.
3) I have also noticed (from so many who have had pre-release ES @ XS, and other forums) that overclocked conroe does not correspond directly to performance, unlike A64.
What I mean is: If A64 is overclocked 20%, the performance increases ~20% 9more or less, in most cases), But have not seen this to hold w/ conroe. So I am wondering what would happen if we put a low end conroe, such as E6400 against A64 4000 x2, 2x1mb cache (same price range after price drop) and overclock them to their limit, using stock cooling, and do the benchmarks (64bit included). The reason I am interested in this type of review is because I am an average end user on budget and would like to know which would give me better price/performance. I think I am speaking for at least 90% of consumers. Not everyone buys $1000 cpu and consumers on budget is detrimental to survival of conroe or AM2 cpus. This alone should give you enough incentive to put together a review oriented around us, the mainstream computer users. We can make or break any chipmaker.
So please Mr. Anad, can we have another review along those lines described above?
We greatly appreciate it.
Thanks,
ochungry
aznskickass - Saturday, July 15, 2006 - link
Hey ochungry, I believe Xbitlabs did an overclocking comparison between a Conroe E6300 and an A64 X2 3800+, and while they didn't do any 64bit benchmarks, the conclusion is that at stock speeds the E6300 is slightly faster than X2 3800+, and when both are overclocked to the max, E6300's lead increases even more.Here is the review/comparison:
http://xbitlabs.com/articles/cpu/display/core2duo-...">http://xbitlabs.com/articles/cpu/display/core2duo-...
OcHungry - Saturday, July 15, 2006 - link
Those guys @ X-bit lab do not fool me. And I hope Anandtech conducts the same test w/ best suited memory module for “BOTH” platforms. We know (so as Xbitlab knows) that, because of IMC, A64 performs its best @ tightest memory timings. X-bit lab should have used http://www.newegg.com/Product/Product.asp?Item=N82...">This memory module if the test was going to be fair and square. Furthermore A64 3800 x2 @ 3ghz is 10x300, which means DDR2 667 1:1 could have been better to use. This http://www.newegg.com/Product/Product.asp?Item=N82...">DDR2 667 @ 3-3-3-10 would have given about 10% better performance than DDR2 4-4-4-12 that they used. X-bit does not mention anything about memory/cpu ratio. What divider was used? Was it 133/266? Or as close to 1:1 as possible? Sooner or later the truth will prevail when we end users try it for ourselves (oh BTW, not ES), and we will see if xbitlab and others were genuinely interested on behalf of consumers, or the interest destined to ill-fated purpose. Will not accuse anyone, but it all look very fishy.I am certain that Mr. Anad will clear all these conspicuous reviews , and hand us another that concerns the consumer’s majority- us average users.
IntelUser2000 - Saturday, July 15, 2006 - link
W-R-O-N-G!!! DDR2-800 at EVEN slower 5-5-5-15 timings is FASTER THAN DDR2-667 3-3-3-10: http://xbitlabs.com/articles/cpu/display/amd-socke...">http://xbitlabs.com/articles/cpu/display/amd-socke...
Prefer AT's results?: http://www.anandtech.com/cpuchipsets/showdoc.aspx?...">http://www.anandtech.com/cpuchipsets/showdoc.aspx?...
DDR2-800 is faster. The myth that it performs amazingly better(comparatively) with lower latency is just, a myth. I believe there was a thread in forums that tells exactly that.
OcHungry - Sunday, July 16, 2006 - link
Iguess you don’t understand much about AMD's memory latency, direct connect, and 1:1 ratio.It is not like Intel's illusionary, that faster than FSB is better. It is not and is useless. Anad proved it here. But this subject has a tendency to be dragged on for ever by those who don’t understand the concept of IMC. So it's better leave it alone.
But I still would like to see tightest timing and 1:1 ratio. It is now clear to me that those reviews in favor of Intel, artfully evade this argument/request, knowing it will give AMD advantage over Intel's FSB.
aznskickass - Sunday, July 16, 2006 - link
*sigh*How is the AMD disadvantaged if BOTH platforms are reviewed using the same RAM?
AMD needs ultra low latency DDR2 to attain best performance? Well, bad luck, Intel doesn't , there is no 'deliberate' conspiracy to put AMD in a bad light.
Look, if you just want to hang on to the notion that AMD has been cheated in the reviews, then go ahead and get your X2 4200+ and see how close you can get to Conroes numbers.
I'll be using my E6600 @ 3.5GHz+ and laughing at your stupidity.
aznskickass - Saturday, July 15, 2006 - link
I consider Xbitlabs to be one of the more trustworthy sites around, and do note that they are testing mainstream chips, and expensive CL3 DDR2 just doesn't make sense in a budget setup, which further puts Conroe in a good light, as it doesn't require expensive CL3 DDR2 to perform well.sum1 - Friday, July 14, 2006 - link
A good read. For the editor, I found 4 errors, search the full lines of text below at:http://www.anandtech.com/printarticle.html?i=2795">http://www.anandtech.com/printarticle.html?i=2795
That begin said,
and H.264 in coding
as soon as its available
at the fastest desktop processor we've ever tested
("and" would be more readable that "at" in this sentence)
JarredWalton - Friday, July 14, 2006 - link
Thanks. At least one of those (in coding) can be blamed on my use of Dragon NaturallySpeaking and not catching the error. The others... well, they can be blamed on me not catching them too. LOL. The last one is sort of a difference of opinion, and I've replaced the comma with a dash, as that was the intended reading. :)--Jarred