nForce4 SLI Roundup: Painful and Rewarding
by Wesley Fink on February 28, 2005 7:00 AM EST- Posted in
- Motherboards
The New Motherboard Test Suite
One of the ongoing concerns at AnandTech has been the tight clustering of performance results in our recent motherboard tests. In general, it is rare to see really wide variations in stock performance with motherboards these days. This has been made even clearer by the AMD Athlon 64 CPU, which has the memory controller on the CPU itself, removing another variable from the chipset equation. This is not bad news for buyers, since more consistent performance at stock speeds makes choosing a motherboard an easier task. Readers have pointed out that we need to do more tests, which would really differentiate boards, and we have been working on updates to our tests.First and foremost, we have been including overclocking tests and memory stress testing for some time - simply because motherboards can vary a great deal in these capabilities. This tells you which motherboards overclock well and which ones are poor, and even if you don't ever plan to overclock, the ability of a motherboard to run at much higher than stock speeds tells you something about the quality of components used in a motherboard. Good overclockers generally use better components and regulate power on the board better, so the good overclocking boards often make sense to buy even if you will never overclock. You can reasonably expect better stability and a longer service life.
Features are increasingly important in motherboards these days as well. With USB, Firewire, IDE, SATA controllers, RAID, LAN, and audio commonly found on top-line motherboards, you are buying much more than sockets for a processor and memory. There are potentially great variations in performance of these features, which could be very important for certain uses of the board. AnandTech has done a good job of detailing these features in past motherboard reviews, but we confess that we have not done very well in actually testing and comparing performance of these features. Our new motherboard tests are designed to correct that.
Starting with this roundup, we are adding iPeak storage tests first used by Anand in his storage reviews. iPeak will be used to establish baseline performance for nForce4 on-chip IDE and SATA performance; iPeak will also be used to test the throughput of the various additional SATA controllers on these SLI motherboards. USB 2.0 and Firewire 400/800 throughput will be measured with a new test developed for motherboard testing. Basically, we create a RAM disk in Windows XP, write a standard test file to the RAM disk, and then copy the file from the RAM disk to a USB 2.0, Firewire 400 or Firewire 800 connected hard drive. We record the time to copy from RAM disk to the connected drive with a timer program developed by our IT Manager.
Ethernet testing uses the Windows 2000 DDK to connect two computers with a CAT 6 crossover cable. We then use a standard host computer as the server and measure the transmission rate and CPU overhead at the client side, which is our test motherboard. We have talked about the advantages of PCIe over PCI gigabit Ethernet in the past, and in this roundup, you will be able to see the actual difference in the performance of Gigabyte LAN over these two busses.
Audio is an area that is still under development and we will be adding tests of audio quality, as well as do subjective listening in future testing. For this roundup, we have included results from the industry standard RightMark benchmark suite for CPU utilization or overhead. There is only room for so much in a roundup, but in the future, we will definitely be including additional audio benchmarks to our motherboard tests.
Last, we have added some new benchmarks, like the popular video synthetic benchmarks from FutureMark - 3DMark 2005 and 3DMark 2003. These tests are particularly useful for testing SLI, since current nVidia drivers support SLI mode in both benchmarks. We are continuing Winstones 2004 for Business and Multimedia, PCMark04, and AutoGK for media encoding. Games are now more heavily weighted toward the most current games with Half Life 2, Far Cry, Doom 3, and Unreal Tournament 2004. Aquamark 3, which is better known as a benchmark than the game on which it is based, is also continued. Return to Castle Wolfenstein-Enemy Territory and Quake 3 have been retained primarily because of their sensitivity to memory performance. It is also a useful reference to include Open GL-based games with so many new game offerings based on Direct X or sporting DX9 front ends.
We are still experimenting with methods of presenting this new data to you, so please let us know if you have suggestions for the future.
Changes in Memory tRAS Recommendations
In past reviews, memory bandwidth tests established that a tRAS of 10 was optimal for the nForce3 chipset and a tRAS setting of 11 or 12 was generally best for nForce2. In the first memory stress test of a production nForce4 board, tRAS timings were first tested with memtest86, a free diagnostic program with its own boot OS that will boot from either a floppy disk or optical disk. Bandwidth of OCZ PC3200 Platinum Rev. 2, based on Samsung TCCD chips, was measured from tRas 5 to tRAS 11 to determine the best setting.| Memtest86 Bandwidth DFI nForce4 with Athlon 64 4000+ |
|
| 5 tRAS | 2191 |
| 6 tRAS | 2242 |
| 7 tRAS | 2242 |
| 8 tRAS | 2242 |
| 9 tRAS | 2141 |
| 10 tRAS | 2141 |
| 11 tRAS | 2092 |
The best bandwidth was achieved with this combination of nForce4/4000+/TCCD in the 6 to 8 range, so a mid-value tRAS of 7 was chosen for all tests. It appears that optimal tRAS timings may also be memory dependent on the nForce4, so we recommend a quick series of memtest86 to establish the optimum tRAS timings for other memories.
Facebook
Twitter
RSS
108 Comments
View All Comments
Reflex - Tuesday, March 1, 2005 - link
#83: Not sure what to say to you about that but AMD has made it clear that ECC is a Opteron but not a A64 feature. The memory controller still has the capability I am certain, but the pinout on the CPU's is different and does not support it as far as I am aware.There are no boards that support it either as far as I know.
Lakku - Tuesday, March 1, 2005 - link
#77 Point well taken. However, this last generation of video cards was a HUGE leap in performance compared to previous generations. Probably not since the move to T&L chips have we seen such a performance increase. I have doubts that the next generation of cards will be able to double performance of current cards. SLI, on average, increases performance from 20% to almost 100% (mostly in synthetic tests on the latter). On newer games that are not CPU/system bound (or games at resolutions above 1024x768), and rely more on the GPU, the increase is greater (i.e. Doom 3 and HL2). This generation doubled the pipelines and increased memory throughput by quite a bit. I don't forsee the next generation being able to do this without substantial hurdles that would need to be overcome or at substantial costs. The fact remains that I don't think a single next generation card (7xxx series lets assume) will handily beat an SLI system, let alone the fact that if supply problems continue, next generation will be 6 or 700 dollars while the 6xxx series will be down to a decent price. The next Unreal engine can run on current SLI setups, as Epic demo'd the engine on 6800 Ultra's SLId. All of this gives support to my idea of mainstream SLI making sense. Two 6600GT's out now give the power (more so in some tests, less when FSAA or super high reso's get involved) of a 6800GT for a lot less cost. Next generation you can upgrade to two 7600GT's that will probably match one future high end card and more then likely beat out any high end card on the market now (and possibly match current high end in SLI, but this is wishful thinking). The concluding points are that it is a good upgrade path. 1) You can get two mainstream cards now that perform great under SLI, 2) Next gen high end will be expensive and probably in short supply, meaning if you did splurge on one high end card now, you can get another then at a reduced cost (probably with your two cards totalling less then one new high end), and 3) The two SLI cards form this generation should match a single card from the future, meaning you can then wait until the NEXT next generation when features will have changed enough to warrant a new generation of cards to play the new generation of games.bob661 - Tuesday, March 1, 2005 - link
#82Are you sure about that? How do you know that the 5.1 speaker test hasn't been encoded in DD?
endrebjorsvik - Tuesday, March 1, 2005 - link
In the 5th paragraph in the Gigabyte K8NXP-SLi-review you have written: "Four ports are 3Gb/s ports provided by the nForce*_3_* chip".That can't be correct, huh?
1955mm - Tuesday, March 1, 2005 - link
#71, Do your homework. http://www.amd.com/us-en/assets/content_type/white... Both the socket 754 and 939 processors support ECC.EODetroit - Tuesday, March 1, 2005 - link
Giz02, I look forward to hearing the results of your audio testing on the MSI. That will probably determine whether I get the DFI or the MSI.The easiest way to test, (Wesley or anyone with the MSI could do this too) is to just run the speaker test in the nvidia control panel with some 5.1 speakers connected to the toslink or digital coax out. If you can only hear front left and front right, its only outputting stereo, just like all other soundblasters. If you can hear all the positions, like I can on my NForce2 Soundstorm audio, then its doing what I hope it does :) .
giz02 - Tuesday, March 1, 2005 - link
Wesley #80Didn't work for me. I tried that with two Ultra PC4000's in the green slots and no post :( The manual has it documented well enough, and I was just replying to what #48 had an issue with.
In my post (79) what do you know what would be better (performance wise) of the two (well three) options I listed?
Wesley Fink - Tuesday, March 1, 2005 - link
#79 -I mentioned in the review of the earlier MSI Neo2 that 4 dimms on auto sets timings to DDR333 on almost every board. This is an AMD Athlon 64 spec for 4 DS dimms. However, on the MSI and all boards in this roundup I found you could force the memory speed to DDR400, lower the Command Rate to 2T - with all other settings the same as 2 DS dimms - and boot just fine. This has been stated clearly in every review and should not be new information.
I have not yet found any board that will run 4 DS dimms with both DDR400 and 1T. We are told the upcoming Rev.E Athlon 64 chips have an updated memory controller that will support 4 DS dimms at DDR400 and 1T. We are looking forward to reviewing Rev. E when it appears in Athlon 64, since it is appearing already in a few Opteron chips.
giz02 - Tuesday, March 1, 2005 - link
#68, I hope to prove otherwise tonight. (ran outta time yesterday!) I purchased the board and installed winxp. Will install the drivers tonight, and all indications thus far indicate that it is a modified chip (or drivers) that supports intel HD. There is also a quote from MSI that indicates it is the only licensed Dolby motherboard on the market (whatever that means). If I am wrong, and there is no Dolby Live Support (PCSTATS said outright that DICE was supported), then I'll be selling this board and hitting DFI!Asus vs MSI. One thing I will say is that the documentation from MSI is pathetic. Broken english, an vague references! I'm not even sure what I was supposed to have in my box! It said the User Manual was optional! Funny thing is, they printed that IN THE USER MANUAL!
Asus has been solid for documentation for some time now, and while thier boards haven't traditionally been the best for performance/OC'ing/features, they are pretty reliable!
giz02 - Tuesday, March 1, 2005 - link
#48,I just read your post, and I had the same problem with the MSI board.
I have an Ultra PC4000 Dual Channel Kit (512x2), and booted up , SPD set the memory to 167 (333) I was blown away, so I manually set it to 200 (400) BOOM. no post. It wasn't until later that I noticed in teh manual that with this board (MSI Neo4SLI) you CANNOT use Dual sided memory in DUAL CHANNEL mode ABOVE 167 (333). In order to go above 333, I'll have to run the chips in slots 1 and 2 (green and purple), thus negating any benifits of DC.
Anyone know what would be better:
1. Running Dual channel at 333
2. Running single channel at 400+ (400-500)
3. Offloading this cheap ass memory to someone else and buying a single sided DC kit :D
giz