Pentium 4 3.46 Extreme Edition and 925XE: 1066MHz FSB Support is Here
by Anand Lal Shimpi on October 31, 2004 3:00 PM EST- Posted in
- CPUs
Does the 1066MHz FSB Improve Memory Performance?
Quite possibly the biggest feature of the 1066MHz FSB today is the fact that it runs at a clock multiple of DDR2-533's frequency. Why is that such a big feature? It's analogy time:
If two people are having a conversation and they can both talk and listen at the same rate, then the conversation will flow as smoothly as possible. If person A talks and listens slower than person B, then person B will always be waiting for person A instead of communicating as fast as possible - a frustrating situation for those that have been here before. Running a FSB and memory bus asynchronously is just as frustrating to the CPU; if the two frequencies aren't synchronous then there is an additional latency penalty incurred while transferring data between the two buses and because of that additional latency penalty, there is a reduction in usable bandwidth.
With the original 925X chipset we were a bit unhappy to see that the Pentium 4's 800MHz FSB was paired with DDR2-533, creating one of those frustrating asynchronous situations. But with a 1066MHz FSB (266MHz x 4), the 925XE can communicate synchronously with DDR2-533 (266MHz x 2), thus reducing memory latency and increasing memory bandwidth in theory. What do we see in practice? To answer this question we look to two trusted measures of memory bandwidth and latency: CacheMem and ScienceMark 2.0.
First looking at latency we see that with the 1066MHz FSB, memory latency with DDR2-533 looks like it's hardly improved. Running the FSB at 1066MHz manages to shave off a just a few clock cycles.
| DDR2-533 3-3-3-12 Latency Comparison - Cachemem | ||||
1066MHz
FSB
|
800MHz
FSB
|
Performance
Improvement
|
||
| 512-byte stride - 32MB block | 226
cycles
|
227
cycles
|
0.4%
|
|
| 1k stride - 32MB block | 239
cycles
|
241
cycles
|
0.8%
|
|
| 2k stride - 32MB block | 266
cycles
|
266
cycles
|
0%
|
|
| 4k stride - 32MB block | 311
cycles
|
311
cycles
|
0%
|
|
We look at ScienceMark and see the same basic situation but with slightly improved performance; looking at the absolute latency values in nanoseconds we see that the 1066MHz FSB manages to reduce memory latency by around 2 - 6%.
| DDR2-533 3-3-3-12 Latency Comparison - ScienceMark 2.0 | ||||
1066MHz
FSB
|
800MHz
FSB
|
Performance
Improvement
|
||
| 16-byte stride | 4.69 ns
|
5 ns
|
6.2%
|
|
| 64-byte stride | 17.5 ns
|
18.12 ns
|
3.4%
|
|
| 256-byte stride | 73.43 ns
|
75.93 ns
|
3.3%
|
|
| 512-byte stride | 75.93
ns
|
77.81
ns
|
2.4%
|
|
The reduction in latency isn't insignificant under ScienceMark, but what about its affects on memory bandwidth?
Looking at Cachemem once again we see an increase in memory bandwidth of just under 4%.
| DDR2-533 3-3-3-12 Bandwidth Comparison | ||||
1066MHz
FSB
|
800MHz
FSB
|
Performance
Improvement
|
||
| ScienceMark 2.0 | 4742.02 MB/s
|
4347.63 MB/s
|
9.1%
|
|
| CacheMem | 3455.3 MB/s
|
3324.7 MB/s
|
3.9%
|
|
ScienceMark appears more considerate of Intel's hard work and shows a 9% increase in memory bandwidth. The thing to keep in mind here is that the improvement in memory bandwidth will depend on the types of applications run, but the ScienceMark and Cachemem results should give you an indication of the range of improvements to be expected in applications that are memory bandwidth limited. In those applications that aren't currently bound by memory bandwidth, the impact will be much less.
What is important to keep in mind here is that DDR2 is still running at relatively high latencies. Even while running synchronously with the FSB, the 925XE and DDR2 combination still posts higher memory access latencies than 875/DDR400 platforms thanks to DDR2's high memory timings. We were able to run 3-3-3-12 timings on our DDR2 test platform by keeping memory voltage at a safe, but overclocked, 2.0V, but anything faster than that was too much for today's DDR2-533. It will take even lower latency DDR2 in order for even the 925XE platform to show some further performance advantages.
Given the relatively small increases in memory bandwidth and decreases in latency, the extra 66MHz of the 3.46EE will have to go a long way in order to gain any more ground for Intel. Let's see how things shape up in some real world tests.
Facebook
Twitter
RSS
63 Comments
View All Comments
IntelUser2000 - Monday, November 1, 2004 - link
Intel is not doing bad. They are doing terrible. So terrible that you might as well call them dead. Probably will last till 2009 before they fill bankruptcy.To those people who say people in forums don't know anything and that there are other people stupid enough to buy Intel chips(I mean all Intel chips): Uhh, yeah, get your head straight, since AMD is closing with Intel very rapidly in marketshare, in server, desktop, and laptop, and that means that gamers actually do make a difference(albeit slowly) making other people buy computers. You think other people will buy P4's because of high clock speed? That's BS, since people who is stupid enough to buy Intel chips don't even know what clock speeds does. There are only a very few that knows computers JUST enough to say clock speed is good.
Tides - Monday, November 1, 2004 - link
believe me, everyone wants the best they can get for the least cash.FinalFantasy - Monday, November 1, 2004 - link
Intel is not doing good right now. I know a lot of people rested their hopes for Intel to strike back with the release of this chip. But alas, we are still seeing the same problems with this chip as we've seen with it's predecessors.1. Way overpriced
2. Still getting whooped by AMD's 64-bit chips
The choice is clear here, buy an average AMD chip for a fraction of the price and still be able to outperform your friend's Intel EE based machine that he spent $2,500-3,000 on to build, while you spent about $1,500 (factor in price a person pays for an Intel EE chip and a couple of sticks of DDR2).
Pandaren - Monday, November 1, 2004 - link
What the community here forgets is that the common person doesn't care for 10 extra FPS in a game or 3 seconds faster on that photoshop filter. They want a reliable, dependable computer with good support at a reasonable price.Dell provides that with Intel chips. People honestly don't care if the Intel chip is not faster.
Those of us who do care about performance and price/performance will build our own. I replaced my Dell with a homebuilt AMD box for that reason, but I don't expect everyone to do the same.
SLIM - Monday, November 1, 2004 - link
"SLIMYou're forgetting DDR2 price which this needs in your so called Intel is "cheaper" comparison. If you want the same price setup you can get a FX-55 and really bring the wood."
ZEBO
Perhaps you didn't read my post correctly or didn't read the review, but the first paragraph in post #14 is a direct quote from the review (that's why there are QUOTATION MARKS around it). The two comments below the QUOTE were my views on Anand's conclusion of the 560 v 3800 comparison at the end of the review. Your comment actually agrees with what I said.
Tides - Monday, November 1, 2004 - link
they would if they had itswatX - Monday, November 1, 2004 - link
why doesnt intel just release a dual core platform or a 64bit chip already..gezz its like they are acting like ATI "we will release 64-vit chips only when apps start to use it" ...NotMrT - Monday, November 1, 2004 - link
Last time i remember AMD domminating this much was in the time of the thunderbirdsChapbass - Monday, November 1, 2004 - link
#43:One thing that i can vouch for, being a college student: Almost every college student not "in the know" with building systems buys a dell....because theyre sold through the college. At least my college they are. DEFINITELY the most common systems around here (and it makes me sick) : P.
It seems like households are more into HP/compaq, where schools, both k-12 and college, are totally dell.
Just something ive noticed around where im from. YMMV.
-Chap
justly - Sunday, October 31, 2004 - link
I find it interesting that the page compareing an A64 3800+ to the P4 560 shows the P4 winning the multitasking content creation. Well that isn't the interesting part, but the fact that Intel only won this because it took all three SYSmark tests is (the SYSmark tests wher the only thing Intel won in this catagory). I guess what really amuses me is remembering a comment in a article a while back (I think it was Anand himself who made the comment) implying that the SYSmark scores did not reflect the rest of their testing and that it seemed to favor Intel.