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
Kougar - Friday, July 14, 2006 - link
I wasn't sure of the B2 Stepping 6 info because it has been impossible to find CPU-Z shots of this chip, let alone anyone reviewing them!!Mostly I just want to ensure Intel doesn't pull any surprises or OC limiting with them, at this point if the B2 Stepping 6 with that last round of bugfixing performs even the same as a Stepping 5 I'd be extremely happy!
redpriest_ - Friday, July 14, 2006 - link
Thanks Wesley - my Core 2 Duo Extreme X6800 is definitely a Stepping 6, Revision B2 Conroe. It has a packaging date of July 5, 2006. I have all the multipliers unlocked, and I am stuck at around 3.466 ghz using multipliers only for stability. I am running on the Intel 975X BadAxe rev 304 with the latest BIOS.I haven't tried upping the bus from there, yet - I am using 1.4 volts; default voltage isn't dual prime 95 stable at 3.466 ghz, but is at 3.2 ghz.
I haven't tried any intermediate voltages between there to test, so it's possible my Conroe is stable at 3.466 ghz at less voltage. I'll give that a try.
Kougar - Friday, July 14, 2006 - link
What motherboard is this? Have you double-checked your RAM settings and memory ratio? I'd imagine setting a 1:1 ratio, using the most relaxed timings, and slowly ramping up the FSB would get you further. And do try less voltages... ;)That's my plan anyway, DDR2-800 memory speeds on a 1:1 ratio thanks to a 400FSB, 10x multiplier... :)
Kougar - Friday, July 14, 2006 - link
Anandtech used Engineering Sample CPUs, so they would be B0 Stepping 4 or B1 Stepping 5 only. Only retail models will be B2 stepping 6...Did you change your cpu voltage any? Sounds like you didn't ;)
redpriest_ - Friday, July 14, 2006 - link
I had to - 1.4 (1.35 real) to get to 3.466, and 1.5 to 3.733.Wesley Fink - Friday, July 14, 2006 - link
Did you OC by adjusting multipliers first? We could run the X6800 at 3.46 at 13 multiplier using stock voltage, but using 315 bus times the stock 11 multiplier required a small voltage increase. The easiest OCs used higher multipliers and modest FSB increases. The E Conroes are locked, however, and can only be overclocked by increasing the FSB.redpriest_ - Friday, July 14, 2006 - link
Yes, I tried using multipliers only, I haven't fiddled with bus speed yet.Kougar - Friday, July 14, 2006 - link
First, the X6800 is selling at NewEgg, and they are price gouging it for all it's worth. Still listed in stock after having been up for 3+ hours...What a great way to start the day... Thank you Anandtech for this great article! Supurb! Now I have to agree with the poster above me, some info on the E6300 overclocking would about round out my curiosity!
I have a question though, do y'all or anyone else have some CPUZ shots of a retail shipping Conroe? I believe this are supposed to be revision B2 stepping 6??
As for the expected skeptics, my saying the numbers appear to jive won't mean anything... so I'll just say there are over 25 reviews out now of the Core 2 Duo released today. :)
mAdMaLuDaWg - Friday, July 14, 2006 - link
I'm curious on if you tried OCing the E6300. What was the highest stable speed you were able to get.xsilver - Monday, July 17, 2006 - link
for those who dont want to read the xbit labs articlethe overclocking of the e6300 is limited by the motherboard - eg. cant get the board to boot at any higher than 420mhz fsb
we may have to wait for nvidia's core2 solution to see higher fsb's