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
AlexWade - Sunday, October 31, 2004 - link
Now if only I can afford and find one ...MMORPGOD - Sunday, October 31, 2004 - link
IntelUser2000 - Sunday, October 31, 2004 - link
DDR2 is not a stupid move, its the speed they are at that's stupid. Remember DDR? They first ones ran at 200MHz, which were 50% faster than PC133 and still way faster than the enthusiast 166MHz SDRAMs. DDR's latency were higher, but since their clock is much higher, it wasn't a big problem as DDR2 vs DDR. However, PC1600 DDR still was not a big improvement over PC133, it was when PC2100 came that DDR started to shine.Another thing:
Quote:"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."
I think 800MHz bus with DDR2-533 is actually VERY synchronous. First look it doesn't look like it. However since DDR2s latency is higher, it doesn't act like DDR533, it acts like DDR400. There was a Tomshardware review that was trying to predict the performance of 1066MHz bus.
First config was: 800MHz bus, DDR2-533
Second: 1066MHz bus, DDR2-533
Third: 1066MHz bus, DDR2-667
Guess which one had the biggest performance benefit? The third one, contrary to most people's belief. I think that tells that because of the DDR2's latency, you need DDR2-667 to perfectly match 1066MHz bus. Since Intel chose to stick with DDR2-533, they have created an asynchronous situation, making the performance not so much better. They should have went DDR2-667 with 1066MHz bus.
SLIM - Sunday, October 31, 2004 - link
One thing I didn't catch from anand's review is that the 3.46ee is rated at 110.7 watts according to [H]; just another reason to go AMD. Makes you wonder what the 3.73ee (which is supposed to launch this quarter) will have for a heatsink...Prometia for everyone:)
Tides - Sunday, October 31, 2004 - link
I remember reading a week or two ago about "AMD is going to have a tough time keeping up," from the lips of an Intel guy.Was this latest outing with the new P4EE's the proof? Perhaps I lack the foresight to understand what will happen in 6 months time, but in who's world is AMD going to have a hard time keeping up with? Cyrix's?
Tides - Sunday, October 31, 2004 - link
"ddr2 is a stupid move."Tides - Sunday, October 31, 2004 - link
not to mention, hi, ddr2 for is a stupid move. high latency, crap bandwidth, not just twice the price since you wouldn't have had to upgrade your ram otherwise if you already had solid ddr1.it reminds me of rambus. and beta max. and sony's discman. what else? ddr2 should have never come out imo. ddr3 is where it's at, hopefully amd will go straight to ddr3 and save it's customers and themselves the hastle of having to buy new ram, new mobos and so forth just to have to do it again with ddr3. i like faster everything as much as everyone else, but amd 64 proves ddr1 is alive and well, and ddr2 is what? exactly? perhaps in a year down the road, or two; it'll be worth something at the end of it's life cycle, just as ddr3 starts poking it's head about.
GhandiInstinct - Sunday, October 31, 2004 - link
Why don't they just screw any other core and focus on pumping out $1000 EEs? Everyones buying them, might as well. I really would like to know the stats for Intel's sales on their new cpus and chipsets, exact numbers.GhandiInstinct - Sunday, October 31, 2004 - link
#17 I was infering this world is off balance with that reality...Gnoad - Sunday, October 31, 2004 - link
wow. $1000 a pop for a CPU that gets destroyed by processors that cost a quarter as much. Totally asinine.