nForce4 SLI Roundup: Painful and Rewarding
by Wesley Fink on February 28, 2005 7:00 AM EST- Posted in
- Motherboards
Ethernet Performance
We have often talked about the throughput advantage of PCI Express compared to PCI and its impact on Gigabit LAN performance. To show you this, the new motherboard test suite needed to include LAN performance measurements.The Windows 2000 Driver Development Kit (DDK) includes a useful LAN testing utility called NTttcp. We used the NTttcp tool to test Ethernet throughput and the CPU utilization of the various Ethernet Controllers used on the nForce4 SLI motherboards.
We set up one machine as the server; in this case, an Intel box with an Intel CSA Gigabit LAN connection. Intel CSA has a reputation for providing fast throughput and this seemed a reasonable choice to serve our Gigabit LAN clients. At the server side, we used the following Command Line as suggested by the VIA whitepaper on LAN testing:
Ntttcps -m 4,0,‹client IP› -a 4 -l 256000 -n 30000On the client side (the motherboard under test), we used the following Command Line:
Ntttcpr -m 4,0,‹server IP› -a 4 -l 256000 -n 30000At the conclusion of the test, we captured the throughput and CPU utilization figures from the client screen.
While Marvel PCI has a bit lower overhead, the throughput for Gigabit LAN on PCI is almost 30% lower than Gigabit on PCI Express. The nVidia on-chip PCIe LAN also exhibits lower CPU utilization than the full Gigabit LAN on PCIe. It has also been reported that enabling nVidia Active Armor lowers CPU overhead even further.
All Ethernet tests were performed with standard frames, but Gigabit Ethernet supports Jumbo frames as well. Jumbo frames will theoretically provide a further reduction in CPU overhead. We have seen test results that show the combination of Active Armor and Jumbo Frames, which can reduce CPU utilization below 10%, a very respectable performance for on-chip gigabit LAN.
108 Comments
View All Comments
ajmiles - Wednesday, March 2, 2005 - link
If i was a doubting type I would suggest that Nvidia spent as much time tuning their drivers for benchmarks as they do games.Nice to see support for some unreleased games such as Battlefield 2 on the list though.
Wesley, you get my email btw? (sorry for bugging you)?
Wesley Fink - Wednesday, March 2, 2005 - link
nVidia has just advised the release of Beta 71.84 drivers now supporting 70 games in SLI. The drivers can be downloaded at http://www.nzone.com/object/nzone_downloads_rel70b... Below is a list of suypported games and benchmarks.Age of Mythology
AquaNox 2: Revelation
Armed & Dangerous
Battlefield 1942
Battlefield 2
Battlefield Vietnam
Breed
City of Heroes
Colin McRae Rally 2005
Colin McRae Rally 4
Conan
Dark Age of Camelot: Atlantis
Desert Rats vs. Afrika Korps
Dirt Track Racing 2
Doom 3
EverQuest
EverQuest II
Far Cry
Flat Out
Ground Control II : Operation Exodus
Half-Life 2
Halo
Hitman 2
IL-2 Sturmovik: Forgotten Battles
Joint Operations: Typhoon Rising
Kohan II: Kings of War
Leisure Suit Larry
Lineage II
Lock On
Lord of the Rings, Battle for Middle-earth
Madden NFL 2005
Max Payne 2
Medal of Honor
NBA Live 2005
Need for Speed: Underground 2
Painkiller
Perimeter
Prince of Persia: The Sands of Time
Quake III
Return to Castle Wolfenstein: Enemy Territory
Return to Castle Wolfenstien
Rome: Total War
Serious Sam: The Second Encounter
Sid Meier's Pirates!
Silent Storm
Sims 2
SpellForce
Splinter Cell
Splinter Cell: Chaos Theory
Star Wars Battlefront
Star Wars Jedi Knight: Jedi Academy
Star Wars Jedi Knight II: Jedi Outcast
Star Wars: Knight of the Old Republic
SWAT 4
The Chronicles of Riddick
Thief: Deadly Shadows
Tiger Woods PGA Tour 2005
ToCA Race Driver 2
Tomb Raider: The Angel of Darkness
Tony Hawk's Underground
Tribes Vengeance
Tron 2.0
Unreal
Unreal 2
Unreal Tournament 2003
Unreal Tournament 2004
Vampire: Bloodlines
Warhammer 40,000: Dawn of War
World of Warcraft
X2: the Threat
Xpand Rally
In addition to these top games, NVIDIA SLI supports the following applications:
3DMark01
3DMark03
3DMark05
AquaMark 3
Code Creatures
D3DRightMark
HDRLighting
NVIDIA Clear Sailing Demo
NVIDIA Dawn Demo
NVIDIA Nalu Demo
NVIDIA Timbury Demo
PCMark04
Shadermark 2.1
Trees of Pangaea
giz02 - Wednesday, March 2, 2005 - link
I've done some further testing, and still no luck. The PCStats review also indicated 192Khz output, but I can't find that either. I'm still hoping for something, and will let you guys know if anything comes up.Other points:
+ The onboard Creative can attenuate the digital outputs just like a regular live could. Most onboard solutions that i have used could not
- Cannot have Analog and Digital outs enabled at the same time (at least I haven't found that yet). All other onboard solutions that I have tried were able to do this. An example of what I'd like to do (ideally) is have my Zalman Real Surround headphones plugged in to the analog ports, and the z5500's plugged into the digital (coax/optical). When the GF complains, I could turn the z5500's off, and put on the headphones. With creative you do this BUT you also have to uncheck the digital out only box. If they can bot h be enabled at the same time, let me know (Y)
- Either the Z5500's can't accept 96/24 on the optical in, or the creative isn't outputting 96/24 on the optical out.
- Only Coax or Optical work at one time (with the Z5500's)
- DD and DTS passthroughs work with Videolan and DVD's/.ts's.
I'll try this board for a while longer, but if encode will not work, I'll be heading to DFI. It's a bit more expensive, but you get the Lan Tote (woohhoo!) and the extra PCIx slots. Anyone have any comments on DFI's onboard sound?
EODetroit - Wednesday, March 2, 2005 - link
Thanks giz02. One of these days some manufacturer is going to realize there's demand for this and meet it... I just hope that day comes sooner than later.SLK75 - Tuesday, March 1, 2005 - link
I bought the GA-K8NXP-9 because fo the rave reviews on its OC abilities which Anandtech also proved in their pre-production sample reviewed towards the end of last year...All of a sudden now Gigabyte's production version of the board does not seem to clock high as was expected and proved previously WHYYYY ???? and Anandtech make it really clear to Gigabyte that people went and bought their board not only first its great features but also for its OC capabilities...I hope Gigabyte can address this with a new BIOSgiz02 - Tuesday, March 1, 2005 - link
I feel DUPED by that PCSTATS review. I've asked the reviewer for why he indicated that the board can DICE (either how to do it, or who told him it was possible).The board is .... OK...
I'da rather had the DFI (if DICE is not possible!)
Aquila76 - Tuesday, March 1, 2005 - link
#89 - looks like we'll both still wait for the next SoundStorm. Maybe the next gen of PCI-E sound cards will have DDL?bob661 - Tuesday, March 1, 2005 - link
#89Thanks for the test.
giz02 - Tuesday, March 1, 2005 - link
So far, It's a negative on the DICE :(I have a single coax cable connected from the onboard card to my Z-5500's and they are not recieving Dolby on the speaker tests. Left and Right channels come through but that is is... 96-24 is working as well, but zilch on the 5.1 :(
1955mm - Tuesday, March 1, 2005 - link
#87: AMD has NOT made it clear that ECC is an Opteron only feature. Read the document from the link I posted. As for ASUS not supporting ECC, download the manual and look at pages 4-21 and 4-22. In the screenshot for DRAM configuration there is an item for ECC enablement. The ASUS K8N-E deluxe (socket 754) also supports ECC. If you still have doubts that the Athlon 64 supports ECC, go to crucial.com and see what memory is supported by the ASUS K8N-E deluxe and A8N-SLI deluxe. I think that you might be confusing registered memory with ECC. If you write code work with critical data ECC is worth having. I have had bad memory in the past that corrupted data without crashing the machine. Considering misinformation that is sometimes provided by motherboard manufacturers and your obvious confusion about the Athlon 64, I think that ECC deserves some mention by motherboard reviewers. I myself would like to understand why the A8N-SLI apparently supports 4 256MB ECC memory modules but not 2 512MB ECC memory modules (page 2-12 in manual), Wesley?