Bandwidth and Memory Scaling
AMD includes the memory controller on their processors, but Intel continues to make the memory controller a part of the motherboard chipset. While the on-chip memory controller is theoretically superior, Intel manages to continue to improve memory bandwidth on their new chipsets. You have already seen in DDR3 vs. DDR2 and Intel P35 Memory Performance: A Closer Look that the new P35 improves memory bandwidth 16% to 18% compared to the same speed and memory timings on P965.
Kingston KHX11000D3LLK2 is the first to offer lower latency in DDR3 and this should further improve memory bandwidth. We compared Standard or Buffered bandwidth on the P965 running DDR2, the new P35 running DDR2, and the new P35 running Kingston DDR3-1375.
At DDR3-800 the Kingston manages stable 5-4-3-10 timings at 1.75V. Based on current motherboards and JEDEC standards the fastest available DDR3 timings are 5-3-3, so the Kingston DDR3-1375 is very close to the theoretical limit at DDR3-800. With the improved timings DDR3 is about 3% faster than DDR3 at 6-6-6 timings. Fast DDR2 on the P35 supporting DDR2 is the widest buffered bandwidth, but the difference between P35 DDR2 3-3-3 and P35 DDR3 5-4-3 is less than 2%. All P35 results, even the lower 6-6-6 timings, exhibit higher bandwidth than P965 at DDR2 3-3-3.
At both 800 and 1066, Kingston DDR3-1375 bandwidth is very close to the bandwidth of fast DDR2 on the P35. Again, all P35 results, even slower DDR3, are faster than P965 with fast DDR2 memory. At 1333 the 7-7-6-15 timings improve bandwidth by around 2.5%, and using lower latency allows the Kingston to run as fast as DDR3-1520 at 8-8-8-22 timings. However, the best bandwidth was achieved at faster timings and slightly slower speed. As shown above, the Kingston managed 1500 speed at 7-7-7-15 timings where standard buffered bandwidth is almost 7500 MB/s.
We also test memory with buffering schemes like MMX, SSE, SSE2, SSE3, etc, turned off. While these features do provide apparent improved bandwidth, the unbuffered bandwidth tends to correlate better with actual gaming and 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, although the domination of P35 in bandwidth performance is not as pervasive. At 800 and 1066 speeds, best bandwidth is with fast DDR2 on the P35 chipset, next is this Kingston DDR3-1375, then Fast DDR2 on P965, and last is slower DDR3. Unbuffered bandwidth is a good mirror of real-world performance, and this is what we expect in gaming tests. It is interesting that the lower latency Kingston has now passed DDR2 on the P965 and is nearly the equal in unbuffered bandwidth to fast DDR2 on P35.
Of course DDR2 could not do the 1333 speed, so the higher speeds of 1333 and 1500+ are the domain of DDR3 and here the Kingston memory shows its true capabilities. Lower latency DDR3 appears to be able to close any gaps that might exist in the overlap speeds of 800 and 1066.
AMD includes the memory controller on their processors, but Intel continues to make the memory controller a part of the motherboard chipset. While the on-chip memory controller is theoretically superior, Intel manages to continue to improve memory bandwidth on their new chipsets. You have already seen in DDR3 vs. DDR2 and Intel P35 Memory Performance: A Closer Look that the new P35 improves memory bandwidth 16% to 18% compared to the same speed and memory timings on P965.
Kingston KHX11000D3LLK2 is the first to offer lower latency in DDR3 and this should further improve memory bandwidth. We compared Standard or Buffered bandwidth on the P965 running DDR2, the new P35 running DDR2, and the new P35 running Kingston DDR3-1375.
| Standard (Buffered) Sandra XI.SP2 Memory Bandwidth - 2.66GHz | ||||
| Memory | 800 | 1066 | 1333 | 1520 (380x7) |
| Kingston DDR3-1333 KHX11000D3LLK2 |
6341 5-4-3-10 1.75V |
6736 6-5-5-12 1.7V |
6928 7-7-6-15 1.7V |
7329 8-8-8-22 1.8V |
| Corsair DDR3-1066 CM3X1024-1066C7 |
6156 6-6-6-15 1.5V |
6613 7-7-7-20 1.5V |
6757 9-9-9-25 1.5V |
- |
| DDR2 - P35 Corsair Dominator |
6456 3-3-3-9 2.25V |
6811 4-4-3-11 2.3V |
- | - |
| DDR2 - P965 (10x266) Corsair Dominator |
5531 3-3-3-9 2.25V |
5782 4-4-3-11 2.3V |
- | - |
At DDR3-800 the Kingston manages stable 5-4-3-10 timings at 1.75V. Based on current motherboards and JEDEC standards the fastest available DDR3 timings are 5-3-3, so the Kingston DDR3-1375 is very close to the theoretical limit at DDR3-800. With the improved timings DDR3 is about 3% faster than DDR3 at 6-6-6 timings. Fast DDR2 on the P35 supporting DDR2 is the widest buffered bandwidth, but the difference between P35 DDR2 3-3-3 and P35 DDR3 5-4-3 is less than 2%. All P35 results, even the lower 6-6-6 timings, exhibit higher bandwidth than P965 at DDR2 3-3-3.
At both 800 and 1066, Kingston DDR3-1375 bandwidth is very close to the bandwidth of fast DDR2 on the P35. Again, all P35 results, even slower DDR3, are faster than P965 with fast DDR2 memory. At 1333 the 7-7-6-15 timings improve bandwidth by around 2.5%, and using lower latency allows the Kingston to run as fast as DDR3-1520 at 8-8-8-22 timings. However, the best bandwidth was achieved at faster timings and slightly slower speed. As shown above, the Kingston managed 1500 speed at 7-7-7-15 timings where standard buffered bandwidth is almost 7500 MB/s.
We also test memory with buffering schemes like MMX, SSE, SSE2, SSE3, etc, turned off. While these features do provide apparent improved bandwidth, the unbuffered bandwidth tends to correlate better with actual gaming and application performance. Unbuffered performance does not always follow the patterns of buffered memory performance.
| Unbuffered Sandra XI.SP2 Memory Bandwidth - 2.66GHz | ||||
| Memory | 800 | 1066 | 1333 | 1520 (380x7) |
| Kingston DDR3-1333 KHX11000D3LLK2 |
4411 5-4-3-10 1.75V |
4761 6-5-5-12 1.7V |
4936 7-7-6-15 1.7V |
5172 8-8-8-22 1.8V |
| Corsair DDR3-1066 CM3X1024-1066C7 |
4098 6-6-6-15 1.5V |
4547 7-7-7-20 1.5V |
4702 9-9-9-25 1.5V |
- |
| DDR2 - P35 Corsair Dominator |
4536 3-3-3-9 2.25V |
4926 4-4-3-11 2.3V |
- | - |
| DDR2 - P965 (10x266) Corsair Dominator |
4226 3-3-3-9 2.25V |
4608 4-4-3-11 2.3V |
- | - |
Unbuffered results show the same basic pattern as buffered results in this case, although the domination of P35 in bandwidth performance is not as pervasive. At 800 and 1066 speeds, best bandwidth is with fast DDR2 on the P35 chipset, next is this Kingston DDR3-1375, then Fast DDR2 on P965, and last is slower DDR3. Unbuffered bandwidth is a good mirror of real-world performance, and this is what we expect in gaming tests. It is interesting that the lower latency Kingston has now passed DDR2 on the P965 and is nearly the equal in unbuffered bandwidth to fast DDR2 on P35.
Of course DDR2 could not do the 1333 speed, so the higher speeds of 1333 and 1500+ are the domain of DDR3 and here the Kingston memory shows its true capabilities. Lower latency DDR3 appears to be able to close any gaps that might exist in the overlap speeds of 800 and 1066.
Facebook
Twitter
RSS
42 Comments
View All Comments
Kozusnik - Thursday, December 6, 2007 - link
Kingstone ram is some of the best ram you can put into your computer by asking me i use it in every computer i build!begsh - Saturday, June 23, 2007 - link
are you really achieved this??i have same modules and they cant get even 1400 at 7-7-7, with mobo asus p5k3 and 0403 bios.
any tips?
Night201 - Friday, May 25, 2007 - link
http://www.memory4less.com/m4l_itemdetail.asp?rid=...">Seems pretty Expensive: ~ $500MadBoris - Saturday, May 26, 2007 - link
P.S. Some of the recent reviews almost seem a bit minimalistic. Hope it's not a trend of things to come. Not to be critical, but I would like to see anandtech provide fresh content, perspectives and methodologies like I've grown accustomed to.Wesley Fink - Monday, May 28, 2007 - link
We would appreciate it if you could share specifics of what would constitute a non-minimalist memory review. What tests and procedures would you add?MadBoris - Saturday, May 26, 2007 - link
You know the more i think about it, the results aren't that tangible. Sure Sandra shows benefits. But if I am running a game at 40 fps, is DDR3 going to give me 41, 42?It won't be noticeable.
Spend less on reliable decent RAM, get a faster CPU or GPU, seriously.
Same with the P35, just not too tangible with speed tests. Mobo's should be about reliability, features sets, testing devices(USB, SATA, RAID) on them and how well they work.
Speed testing with RAM or Mobo's isn't tangible enough. When a new chipset or RAM increases things 15 - 20% then I'll be interested. I'm not really interested in shaving .5 seconds off a compile or an encode.
Wesley Fink - Friday, May 25, 2007 - link
We have updated ALL charts in the review to make it easier to compare performance. Bandwidth Standard, Bandwidth Buffered, Super Pi, and the 3 games now included 1520 (380x7) results in the last column. This means all four rows are now running at 2.66GHz, with just a change in the memory bandwidth. *00, 1066, and 1333 are running 8x333, and 1520 is 7x380 - all 2.66GHz as stated at the top of the chart.We have added an Overclocking Chart to p.5 that includes 7x380 (2.66GHz)- 8-8-8-22, 8x380 (same multiplier as 800/1066/1333 but pushed ot highest OC at 3.04GHz)- 8-8-8-20, and 8x275 (3.0GHz - highest speed at 7-7-7 timings) - 7-7-7-15. so you cna see the impact of timings at the very top overclocks. It should be no surprise that 1500 7-7-7-15 results are the fastest.
With these changes we think we have addressed your suggestions on making the performance charts more useful for readers.
Wesley Fink - Friday, May 25, 2007 - link
For those who requested them, these are the results for the higheest memory speed at slower timings. After further testing, we managed 1520 8-8-8-22 timings at 1.8V.The first result is 7x380, which is the same 2.66GHz run at all other memory speeds, and the second is 8x380, which is the same ratio but the highest OC we could reach from the base memory setting of 1333. The sequence is test, 7x380 (2.66), 8x380 (3,04):
Sandra XI-Standard Buffered - 7329, 7462
Sandra XI-Standard UNBuffered - 5172, 5263
Super Pi 1.5 - 45.31, 40.40
Far Cry River - 107.46, 117.82
Quake 4 - 116.0, 123.5
Half Life 2-Lost Coast - 109.5, 111.5
We will add a chart with these results to the bottom of the overclocking section later today.
Googer - Friday, May 25, 2007 - link
2GB of DDR3 will cost you close to $385!http://www.google.com/products?q=KHX11000D3LLK2&am...">http://www.google.com/products?q=KHX11000D3LLK2&am...
TA152H - Thursday, May 24, 2007 - link
A lot is being made of DDR3 latency and such, and performance, but hasn't anyone considered the impact of voltages? I see these grossly inflated voltages for DDR2 memory, and I can't help but wonder if they would have so much better performance, even clock speed normalized, if they were both run at stock voltages.A lot of places aren't stupid enough to run DDR2 at 2.2 or 2.3 volts, it creates a lot of heat and lowers the lifespan of the device. Sure, the kiddies will, but the business world isn't that crazy about running things out of spec. Now we have the jackasses at Kingston already producing 1.7v DDR3. Why even bother having a spec if no one pays attention to it???? The memory is just out, and they can't stay to spec.
But anyway, it might be interesting to compare memory at spec, which, last I remember, was 1.5v for DDR3 and 1.8v for DDR2. Or even at the same voltage, to see what is intrinsic to DDR3 and DDR2. It might be the voltage difference accounts for a lot of the higher timings, and not the standard. Not that I'm advocating running DDR3 at 1.8v, but for testing, it would be informative. Certainly if these nitwits are running DDR2 at 2.2-2.3v, DDR3 at 1.8v can't be too far behind. Good grief.