Bandwidth and Memory Scaling
One of the surprises in comparing DDR2 performance on AM2 and Core 2 Duo was the much better memory bandwidth found on the AM2 platform courtesy of the on-chip memory controller. Unfortunately, this did not translate into significant performance improvements compared to a similar AMD processor running DDR. At that point we concluded that Core 2 Duo was not particularly bandwidth sensitive, since it made very good use of the memory bandwidth available.
In our earlier review we were really comparing the DDR2 memory controller on AM2 to the 975X chipset memory controller, since Intel continues to place the memory controller in the chipset. We have speculated since then whether an improved memory controller in a socket 775 chipset would bring with it improved performance.
P965 brought very minor changes, mainly in the straps and overclocking ability of the memory. The NVIDIA 680i/670/650 actually shows decreased buffered bandwidth, but unbuffered bandwidth is about the same as P965. This reinforced the notion that memory bandwidth didn't matter much with Core 2.
To begin our investigation into DDR3 performance, we compared Standard or Buffered bandwidth on the P965 running DDR2, the new P35 running DDR2, and the new P35 running DDR3. As you can see the results are very interesting.
While the purpose of this review was to compare DDR3 and DDR2 performance, something completely different emerged from the memory bandwidth tests. Namely, the memory controller on the P35 is definitely an improvement over the P965 memory controller. This is evident whether the P35 is running DDR2 or DDR3 memory.
In those cases where we can run timings the same or close to the same, as in 800 memory speed performance, DDR2 and DDR3 results are virtually identical. By 1067 the current slow DDR2-1067 timings of 7-7-7-20 are performing just as well as DDR2 running at 6-6-6-15. The superior timings of DDR2-1067 at 4-4-3 still provides the best bandwidth at that speed. Of course, DDR3 is currently alone at the 1333 memory speed, but even with the current slow 9-9-9-25 timings it performs nearly as well as DDR2-1067 at 4-4-3 timings.
We normally also test memory with buffering schemes like MMX, SSE, SSE2, SSE3, etc. turned off. While these features do provide apparent improved bandwidth, we have found the unbuffered bandwidth to correlate better with real-world application performance. Unbuffered performance does not always follow the patterns of buffered memory performance.
Unbuffered results show the same basic pattern as buffered results in this case. Here DDR3 is clearly the best performer at the same slow timings at DDR2-800, with DDR2 on the P35 behind about 3% and DDR2 on P965 about 12% lower. DDR2 is still faster at the better timings available with current DDR2 memory.
In Standard/Buffered memory bandwidth, the P35 (Bearlake) chipset is providing a 16% to 18% improvement in memory bandwidth compared to the P965. This is a significant improvement. The Unbuffered improvement is smaller, in the range of 4% to 8%. These bandwidth improvements may or may not translate into improved system performance. We will examine that in the SuperPi and Gaming benchmarks.
One of the surprises in comparing DDR2 performance on AM2 and Core 2 Duo was the much better memory bandwidth found on the AM2 platform courtesy of the on-chip memory controller. Unfortunately, this did not translate into significant performance improvements compared to a similar AMD processor running DDR. At that point we concluded that Core 2 Duo was not particularly bandwidth sensitive, since it made very good use of the memory bandwidth available.
In our earlier review we were really comparing the DDR2 memory controller on AM2 to the 975X chipset memory controller, since Intel continues to place the memory controller in the chipset. We have speculated since then whether an improved memory controller in a socket 775 chipset would bring with it improved performance.
P965 brought very minor changes, mainly in the straps and overclocking ability of the memory. The NVIDIA 680i/670/650 actually shows decreased buffered bandwidth, but unbuffered bandwidth is about the same as P965. This reinforced the notion that memory bandwidth didn't matter much with Core 2.
To begin our investigation into DDR3 performance, we compared Standard or Buffered bandwidth on the P965 running DDR2, the new P35 running DDR2, and the new P35 running DDR3. As you can see the results are very interesting.
| Standard (Buffered) Sandra XI.SP2 Memory Bandwidth - 2.66GHz | |||
| Memory Speed | P965 ASUS P5B Dlx |
P35 DDR2 ASUS P5K Dlx |
P35 DDR3 ASUS P5K3 Dlx |
| DDR2-800 3-3-3-9 | 5531 | 6456 | - |
| DDR2-800 5/6-6-6-15 DDR3-800 6-6-6-15 |
5207 | 6143 | 6156 |
| DDR2-1067 4-4-3-11 | 5782 | 6811 | - |
| DDR2-1067 5/6-6-6-15 | 5712 | 6621 | - |
| DDR3-1067 7-7-7-20 | - | - | 6613 |
| DDR3-1333 9-9-9-25 | - | - | 6757 |
While the purpose of this review was to compare DDR3 and DDR2 performance, something completely different emerged from the memory bandwidth tests. Namely, the memory controller on the P35 is definitely an improvement over the P965 memory controller. This is evident whether the P35 is running DDR2 or DDR3 memory.
In those cases where we can run timings the same or close to the same, as in 800 memory speed performance, DDR2 and DDR3 results are virtually identical. By 1067 the current slow DDR2-1067 timings of 7-7-7-20 are performing just as well as DDR2 running at 6-6-6-15. The superior timings of DDR2-1067 at 4-4-3 still provides the best bandwidth at that speed. Of course, DDR3 is currently alone at the 1333 memory speed, but even with the current slow 9-9-9-25 timings it performs nearly as well as DDR2-1067 at 4-4-3 timings.
We normally also test memory with buffering schemes like MMX, SSE, SSE2, SSE3, etc. turned off. While these features do provide apparent improved bandwidth, we have found the unbuffered bandwidth to correlate better with real-world application performance. Unbuffered performance does not always follow the patterns of buffered memory performance.
| Unbuffered Sandra XI.SP2 Memory Bandwidth - 2.66GHz | |||
| Memory Speed | P965 ASUS P5B Dlx |
P35 DDR2 ASUS P5K Dlx |
P35 DDR3 ASUS P5K3 Dlx |
| DDR2-800 3-3-3-9 | 4226 | 4536 | - |
| DDR2-800 5/6-6-6-15 DDR3-800 6-6-6-15 |
3668 | 3975 | 4098 |
| DDR2-1067 4-4-3-11 | 4608 | 4926 | - |
| DDR2-1067 5/6-6-6-15 | 4389 | 4557 | - |
| DDR3-1067 7-7-7-20 | - | - | 4547 |
| DDR3-1333 9-9-9-25 | - | - | 4702 |
Unbuffered results show the same basic pattern as buffered results in this case. Here DDR3 is clearly the best performer at the same slow timings at DDR2-800, with DDR2 on the P35 behind about 3% and DDR2 on P965 about 12% lower. DDR2 is still faster at the better timings available with current DDR2 memory.
In Standard/Buffered memory bandwidth, the P35 (Bearlake) chipset is providing a 16% to 18% improvement in memory bandwidth compared to the P965. This is a significant improvement. The Unbuffered improvement is smaller, in the range of 4% to 8%. These bandwidth improvements may or may not translate into improved system performance. We will examine that in the SuperPi and Gaming benchmarks.
Facebook
Twitter
RSS
45 Comments
View All Comments
vaystrem - Tuesday, May 15, 2007 - link
Thank you for your response :)cool - Tuesday, May 15, 2007 - link
What am I missing? How can you have DDR2 800 results for the DDR3 Bearlake mobo?Look at the 3rd row, 4 column of the Sandra benchmarks results on page 5.
TA152H - Tuesday, May 15, 2007 - link
Their chart is a little bit off. It is how they did their best to speed normalize DDR2 and DDR3, so for the DDR2 column for the Bearlake it is DDR2 memory they tested, for the DDR3 version, it is DDR3 memory.It's probably the most useful thing in there, in my opinion. It shows you get extra speed even at the exact same timings, with memory using lower voltage. It's pretty impressive.
cmdrdredd - Tuesday, May 15, 2007 - link
not very impressive to me considering the prices you'll be paying. That's $300 difference between a good DDR2 kit and a DDR3 kit (probably not even high bin or binned at all).TA152H - Tuesday, May 15, 2007 - link
Well, agreed, currently you're not getting much bang for the buck, but that's always the case when new memory comes out. Look at the price of RDRAM, it never got better :P.But, obviously, the prices today mean essentially nothing for the future of the technology, it's just representative of it being a very new technology right now. I'd expect it should be a tiny bit more expensive everything else being equal because some of the additions to it, but it should be insignificant. It's a much cleaner design, and you get some performance on top of it, so I think it's a good technology. I agree, for now, it's difficult to validate the price of it. It's not even like you could say that you're better off getting a DDR3 based design now so you can reuse the RAM. The latencies are so poor on the current stuff you'd probably be aghast at using it a year from now, and you'd probably be able to buy 1 GB of DDR2 today, and 1 GB of DDR3 in a year for less than 1 GB today. So, right now, I don't think it makes much sense to anyone.
Good thing for Intel pushing it though. Anyone that dislikes them needs to recognize how important they are for moving technology forward even at their own cost. Sometimes they've failed though, such as with RDRAM and, so far, EPIC.
Wesley Fink - Tuesday, May 15, 2007 - link
DDR3 will run at DDR3-800 6-6-6-15 timings. We will add this to the DDR-800 5/6-6-6- line to clarify. The point was to run all 3 boards and the two memories at the exact same speed and timings. 800 was the only speed that allowed this.cool - Tuesday, May 15, 2007 - link
Thanks for the explanation, Wesley.I was confused thinking the P5K3 was maybe one of those motherboards that can take 2 types of different RAM, namely DDR2 and DDR3. But that is not the case.
TA152H - Tuesday, May 15, 2007 - link
I see timings of 5/6-6-6-15 and such. What does 5/6 mean? The P965 is running at five and the others at six?Wesley Fink - Tuesday, May 15, 2007 - link
We added this clarification to Page 4 - Memory Test Configuration:"While memory timings were matched to the same memory speed wherever possible, there were a few settings where the chipsets did not allow a direct comparison. DDR3-800 runs at 6-6-6-15 timings. The P965 has options to set 6-6-6-15 timings but the board would not boot under any settings or voltage we fed it at 6-6-6 timings. The closest timings that would work on the P965 at 800 speed were 5-6-6-15. The P5K DDR2 board, based on the P35 chipset, would allow setiing and running 6-6-6-15 timings. This is reflected in our charts with the line ID of 5/6-6-6- for timings. We also tested DDR2 at the fastest timings it could achieve with complete stability on both the P5B Deluxe and P5K Deluxe. This was 3-3-3-9 at DDR2-800 and 4-4-3-11 at DDR2-1066."
TA152H - Tuesday, May 15, 2007 - link
That's clear, I think that will remove any questions relating to the charts.