Socket-AM2 Performance Preview
Without major architectural changes to the new AM2 CPUs, we wanted a quick and easy way to showcase the performance differences between AM2 and Socket-939. What we've got is a massive table below with all of our usual CPU benchmarks and their results for the same CPU in both Socket-939 and AM2 varieties, and the performance benefit offered by AM2:
| Benchmark | Socket-939 (DDR-400) | Socket-AM2 (DDR2-800) | % Advantage (Socket-AM2) |
| PC WorldBench 5 | 115 | 115 | 0% |
| Business Winstone 2004 | 23.3 | 23.2 | -0.4% |
| Multimedia Winstone 2004 | 38.4 | 38.9 | 1.3% |
| SYSMark 2004 | 220 | 224 | 1.8% |
| ICC SYSMark 2004 | 282 | 286 | 1.4% |
| OP SYSMark 2004 | 171 | 175 | 2.3% |
| 3dsmax 7 | 2.38 | 2.38 | 0% |
| Adobe Premier Pro 1.5 (Export w/ Adobe Media Encoder) | 130 s | 128 s | 1.5% |
| Adobe Photoshop CS2 | 210.6 s | 210.3 s | 0.1% |
| DivX 6.1 | 11.6 fps | 12.0 fps | 3.4% |
| WME9 | 35.2 fps | 35.6 fps | 1.1% |
| Quicktime 7.0.4 (H.264) | 3.63 min | 3.63 min | 0% |
| iTunes 6.0.1.4 (MP3) | 43 s | 43 s | 0% |
| Quake 4 - 10x7 (SMP) | 111.3 fps | 117.4 fps | 5.5% |
| Call of Duty 2 - 10x7 | 59.3 fps | 60.1 fps | 1.3% |
| F.E.A.R. - 10x7 | 92 fps | 94 fps | 2.1% |
| Multitasking Test (LAME + WME + Anti Virus + Zip) | 216.3 s | 213.4 s | 1.4% |
| ScienceMark 2.0 (Bandwidth) | 5007 MB/s | 6805 MB/s | 36% |
| ScienceMark 2.0 (Latency 512-byte stride) | 53.83 ns | 49.77 ns | 7.5% |
We'll start at the bottom of the table and go up from there. Rev F processors feature a 128-bit DDR2-800 memory controller, which works out to offer a peak theoretical bandwidth to/from memory of 12.8GB/s. As you can expect, that's twice the bandwidth of Rev E CPUs' 128-bit DDR-400 controller at 6.4GB/s. Thus to see a 36% increase in memory bandwidth according to ScienceMark is to be expected, albeit a bit on the low side. The old DDR-400 memory controller is able to deliver 5GB/s out of a maximum of 6.4GB/s, but now we're only seeing 6.8GB/s out of a maximum of 12.8GB/s with AM2. This however is a huge step for AMD, as it is the first spin of the Rev F silicon that we've been able to see such a significant advantage in theoretical memory bandwidth over previous DDR-400 cores.
What's even more important than the increase in memory bandwidth is that access latency has been reduced by 7.5% over the DDR-400 memory controller in the Rev E cores. Lower latency and more bandwidth means that, at bare minimum, performance won't go down. At least, not perceptibly: .4% slower in one test that has a 1-2% variability is nothing to worry about.
It also doesn't guaranee that performance will go up, as you can see from the results above. If we only count the overall SYSMark score and leave out the synthetic tests, the real world performance advantage averages out to a little under 1.3%. There are some special cases such as Quake 4 and DivX were performance goes up fairly reasonably, which can be expected since both of those tasks are fairly bandwidth intensive and make good use of both cores. However similar benchmarks, such as F.E.A.R. and Windows Media Encoder 9 show lower improvements, so it is very dependent on the specific application and workload.
It's important to note that until recently, AM2 samples were not able to produce scores even on par with Socket-939, so the fact that we're seeing a performance increase at all is a major step from where we were just a couple of months ago. The real question is, is this all we get?
Facebook
Twitter
RSS
107 Comments
View All Comments
Thor33 - Monday, April 10, 2006 - link
It seems the test was done with the fastest memory available in fact is a PC2-8000 1000mhz rated. OCZ EL DDR2 PC2-8000 / 1000 MHz / Platinum XTC / Extreme Editionhttp://www.anandtech.com/memory/showdoc.aspx?i=273...">http://www.anandtech.com/memory/showdoc.aspx?i=273...
Unless there is a typo on the test memory utilized below is PC8000 drop down to work as PC2-6400?
Memory:
PC2-6400 OCZ PC8000 DDR2-800 4-5-4-15 (1GB x 2)
OCZ DDR-400 2-3-2 (1GB x 2)
The test should be a PC2-6400 which is DDR2-800 mhz rated memory. If this test is performed with an average memory maker DDR2 PC2-6400 memory the score will be even less impressive. This is worth mentioning.
Rated at DDR2-1000, this OCZ announces right off the bat that it is serious new memory. Past DDR2 has struggled to reach DDR2-1000, so rating this new OCZ at DDR2-1000 is a "look-at-me" announcement. The rated timings of 4-5-4-15 at DDR2-1000 are also exceptionally fast for DDR2 memory at that speed. By the way, for those of you who have trouble translating PC speed ratings to Memory Speeds, all you have to do is divide by 8. Thus PC2-8000 is DDR2-1000 speed, PC-5400 is DDR2-667 (results are approximate), PC2-6400 is DDR2-800, etc.
waldo - Monday, April 10, 2006 - link
I haven't heard anything about changes in regards to the opteron platform, or are they staying the same. I would assume they would change to use the ddr2 memory as well.I actually can't see how purchasing a new AM2 system is good for anyone, whether a new system or upgrading. With prices being higher, almost double, and parts being barely similar in performance with premium boards, memory, etc., it doesn't seem like a good idea for anyone to hop on early to the AM2 platform. Just seems like a waste of a lot of money. In this case, the latest is not the greatest.
bob661 - Monday, April 10, 2006 - link
Where are the double prices? The ram is more money but everything else looks to be the same. According to http://tinyurl.com/koapt">this, the prices for the CPU's are the same as current one's. I doubt the motherboards will cost more than the typical new release prices.petz - Monday, April 10, 2006 - link
What was the date code on the CPU? Why no screenshot showing that the CPU was at least recognized? The BIOS memory settings are meaningless if the CPU is not properly recognized. I would have done a disk benchmark, because the fact that the poorest result was from Business Winstone tells me that something was wrong in that department.Anand Lal Shimpi - Monday, April 10, 2006 - link
The shots of the chip itself were kept out of the review to protect our source for the chips obviously. A difference of 0.1 points in Business Winstone is within the normal margin of error (< 3%) for that benchmark, the difference itself is basically 0. The CPU was also recognized properly by the BIOS.Take care,
Anand
nicolasb - Monday, April 10, 2006 - link
It seems to me that socket AM2 is likely to be just a stepping stone. Current AMD processors don't need DDR2 - but the next generation will.Conroe seems likely to stamp all over AMD for a few months. The point when things will get interesting again is when we see quad-core processors begin to ship from both Intel and AMD.
AMD's architecture has been tuned for multi-processor systems right from the start. Post-Conroe quad-core chips may be a bit clunky - effectively two dual-core chips glued together, in the same kind of way that Pentium D is two single-core chips glued together. AMD may therefore regain the lead in the quad-core arena - if there are no other limiting factors, such as memory bandwidth.
If a dual-core chip is using barely half of the available memory bandwidth of DDR2-800, what that means is that a quad core chip will not be significantly limited by memory bandwidth. I think this is what AMD is planning: they will need DDR2 bandwidth for quad-core, and they want to switch over to it now to get all the bugs ironed out in time for K8L. So socket AM2 has no significance for the current generation of chips, and was never meant to have: it's merely preparation for 2007.
If this is true, then the fact that current chips use only half the available bandwidth may actually be deliberate, and something that won't change.
sprockkets - Monday, April 10, 2006 - link
I like your post.Remember the KT266 fiasco? Then KT266A, then KT333, then KT400?
Yeah, the KT266A killed everything, when the KT266 showed no performance gains and just like then, said memory bandwidth is useless for AMD Athlons.
Except the problem here is, to get an updated "chipset" or memory controller, you have to upgrade the processor and perhaps the board, and AMD has already done 3 rev of the processor.
MrKaz - Monday, April 10, 2006 - link
PC133 SDRAM > DDR at early stagePC133 SDRAM > RDRAM at early stage
It will improve on next revision I bet, I was expecting 0% improve.
This is just a RAM type exchange.
You AMD systems users must show appreciation to AMD for the excellent platform they developed.
With Intel each time there is a memory speed bump, you "have" to upgrade the motherboard because of the memory controller on the chipset, even if Intel decides to keep the same socket. How many times since DDR266 have Intel “forced” people to upgrade motherboard/chipset in order to catch up every time there is a memory speed bump?
Even at the same RAM speed Intel “offers” (example) 865 and 875 chipsets where one is faster than the other with the same RAM?
Since AMD integrates the memory controller, you never get problems like that. AMD integrated memory controller is compatible with DDR 266 to 400 (unofficial 433,466,500).
With DDR2, you will get 533,667,800 (unofficial 933, 1066).
You always keep the same board and RAM. The processor is the only that changes.
If AMD didn’t get socket 754, and only 940 or 939 we where looking at the perfect computer platform.
Calin - Monday, April 10, 2006 - link
RDRAM was a very bad choice for the processors to which was first mated. Pentium !!!, like the Athlons of today, fare better on low latency lower bandwidth memory than on high latency high bandwidth memory (as RDRAM was). However, RDRAM has its place in small devices - where you can not fit all of the 168 pins of SDR memory, or all of the 180+ of the DDR memory.Pentium 4 was doing very well on RDRAM, dual RDRAM was quite a bit faster than dual SDR or single DDR. There was a time when intel 850 with dual RDRAM was the high mark in the chipsets
flyck - Monday, April 10, 2006 - link
RDRam had a LOWER latency then DDR.Mainly because of the superior chipset design (which had less internal latency).