nForce3-250 - Part 2: Taking Athlon 64 to the Next Level
by Wesley Fink on March 29, 2004 11:00 AM EST- Posted in
- CPUs
nForce3-250Gb: Stress Testing
We performed stress tests on the nForce3-250Gb Reference Board in these areas and configurations:- Chipset and motherboard stress testing was conducted by running the FSB at 249MHz at a multiplier of 9.5.
- Memory stress testing was conducted by running RAM at 400MHz with a DIMM slot filled and at 400MHz with 2 DIMM slots filled at the lowest memory timings possible.
Front Side Bus Stress Test Results:
As normally done in our testing of production motherboards, we ran a full range of stress tests and benchmarks on the nForce3-250Gb Reference Board to test stability at an overclocked speed. This included Prime95 torture tests, and the addition of other tasks - data compression, various DX8 and DX9 games, and apps like Word and Excel - while Prime95 was running in the background. Finally, we ran our benchmark suite, which includes ZD Winstone suite, Unreal Tournament 2003, SPECviewperf 7.0, and Gun Metal Benchmark 2. At default voltage, 249MHz was the highest overclock that we were able to achieve with the nForce3-250Gb and the HyperTransport setting reduced to 4X or 800 (Actual 996 at 249FSB) while running these tests.Unlike our experiences with some of the overclocks run on other A64 boards, the nForce3-250Gb was completely stable when overclocked. We did have stability problems at the highest 250 setting, which remains a puzzle, but 249 was completely stable. It would be impossible to run at 249FSB with SATA drives and our picky ATI 9800 PRO without a working PCI/AGP lock.
Memory Stress Test Results:
This memory stress test simply tests the ability of the nForce3-250Gb Reference Board to operate at its officially supported memory frequency (400MHz DDR) at the lowest supported memory timings that our OCZ PC3500 Platinum Ltd memory will support:| Stable DDR400 Timings - 1 DIMM (1/2 DIMM populated) |
|
| Clock Speed: | 200MHz |
| Timing Mode: | N/A |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 5T |
| Precharge Delay: | 2T |
| Command Rate: | N/A |
We had no problem running one DIMM of our standard OCZ PC3500 Platinum Ltd at the highest memory timings in the nForce3-250Gb Reference Board. We were also able to run the memory test suite with complete stability at 2-2-2-5 timings.
Filling all available memory banks is more strenuous on the memory subsystem than testing 1 DIMM, but 2 DS DIMMs worked just fine on the nForce3-250Gb. With 2 DIMMs, we could run the same aggressive timings used for one DIMM, which is excellent performance for an Athlon 64 motherboard.
| Stable DDR400 Timings - 2 DIMMs (2/2 DIMMs populated) | |
| Clock Speed: | 200MHz |
| Timing Mode: | N/A |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 5T |
| Precharge Delay: | 2T |
| Command Rate: | N/A |
We tested the memory timings with both banks filled using several stress tests and general applications to guarantee stability. Prime95 torture tests were run successfully at the timings listed in the above charts. We also ran ScienceMark (memory tests only) and Super Pi. None of the three stress tests created any stability problems for the nForce3-250Gb Reference Board at these memory timings.
We suspect production nForce3-250Gb motherboards will have 3 or 4 DIMM slots, but as we have reported in the past, we have had some difficulties with every Athlon 64 motherboard so far when we tried to use more than 2 DIMMs. We will test memory slots again when production motherboards appear on the market in a few weeks.
Certainly, we will see 4 DIMMs, or two dual-channel pairs, on the nForce3-250Gb Ultra version for Socket 939.
Facebook
Twitter
RSS
46 Comments
View All Comments
PrinceGaz - Friday, April 2, 2004 - link
I was also staggered by how bothered some people were with SoundStorm -- it was never state-of-the art sound, just like how the onboard GeForce4 MX-like graphics the nForce2 had was something no serious gamer would be happy with. If you want top good quality sound you buy something specifically designed for it as you'll never get it on a motherboard (unless you live in an AOpen fantasy world and think putting valves on a motherboard counts).The CPU utilisation while using RAID on the nVidia chipset, and also on a straight IDE drive as well, did come as a bit of a surprise to me. I'm into capturing and encoding high-quality video so excessive CPU load from disk-access while doing so would be a real issue to me as it demands both CPU power as well as disk-throughput.
vlor - Friday, April 2, 2004 - link
hummm,I only saw a few things that bugged me but what amazes me is out of all the people who posted, hafc why sound storm wasn't on the 250... do a little research and you will find that the codec's cause interference, and we as take up significantly more signal processing than most people care to think about. I'd love to see what Nvidia could do with a riser card once the PCI buss isn't so saturated. Oh and the is a board with an realtek850 linked to an nforce2, and DFI doesn't even mention sound storm but the setup is one of the better soundstorms out there including my asus deluxe.
Reflex - Wednesday, March 31, 2004 - link
Wow, the hate level here is extreme, especially since it was a pretty good review overall...Phiro: Go elsewhere. Your not being helpful at all, and your complaints are simply rediculous in many situations. If you really want the level of technical detail you seem to be demanding(but do not appear to have the background to comprehend), go to Ace's Hardware. More knowledge than you'll ever need to know.
As for the nit-pickers: A bit of terminology was off, however overall they were accurate and I think some of these complaints are not concerning any mistake Anandtech made, but simply that they did not include information that you personally felt should be in there(never mind the fact that since you noticed it missing, it apparantly is information you already knew). I also will point out that much of what you are complaining about is already covered in previous articles on the various technologies, and are available for anyone who wants to know exactly what HyperTransport means, etc. Nothing wrong with clicking a few links to find it. I would not mind seeing AT add a 'related articles' link somewhere obvious with all related technologies listed(and possibly even some off site links, Ace's and Ars Technica often have great technology writeups). But its not that big a deal overall...
To the guys about SoundStorm: You are really in the minority, anyone interested in quality sound uses and add-on card already, and as for basic sound the solution they are including is good enough for them. Your essentially begging for a mid-range to low end solution on an already feature rich setup, I'd say spend the $40 on something like that if you really want to, or just drop $20 more and you can have a Santa Cruz and blow it out of the water.... It is not a huge loss, its the first thing I disabled on my NF2 board.
Anandtech: Keep up the good work. I can't wait to see the first PCI Express capable chipset review..
Da3dalus - Tuesday, March 30, 2004 - link
There should be tests of the 5950U on a non-nVidia chipset as well, to see how much it gains when coupled with the nf3-250Gb. It just seems a bit lacking in that right now.agent2099 - Tuesday, March 30, 2004 - link
“Apparently you COMPLETLY overlooked the fact that most (all, AFAIK) SoundStorm boards offered digital output which bypassed those codec's completly. Which just happens to be EXACTLY what I use, and was a major selling point, allowing me to send out my DD/DTS signal to my amp for DVD's as I use for regular computer use. VERY handy”I agree, I think the lack of soundstorm is a huge oversight. I currently use the digital (optical out) from my nforce2 board and couldn’t live without it. This is especially important for SFF PCs.
“I do not find it surprising at all that most customers don't care about the sound storm stuff. After all, the vast majority of PC users (outside of the geek market of course) have their PC hooked up to $9.99 basic speakers -- sound storm is not going to make a difference at all, and is simply not a selling point.”
Well of course we are talking about the “geek market.” If we are talking about the average PC user that has $9.99 speakers, they don’t need ½ of the other things on the chipset, including SATA RAID and Gigabit Ethernet.
Foxbat121 - Tuesday, March 30, 2004 - link
I'm a little disappointed and alarmed by the high CPU utilization (up to 35.1%) on this nVidia chipset(or maybe all A64 platform?) compared to Intel's 3.8% CPU utilization. Which means your CPU could spend 1/3 of its time processing disk access during disk intensive operations.Nighteye2 - Tuesday, March 30, 2004 - link
the single 10k benchmark is a good addition, but there still isn't any comparison of SATA RAID between 2 Athlon 64 systems - can you add the SATA RAID performance of the integrated RAID of the ASUS K8V MB? The review just isn't complete without a good comparison of RAID performance to a similar platform. (with the same processor)Wesley Fink - Tuesday, March 30, 2004 - link
#38 -This is not completely accurate. While there is not and will not be a specific dual-channel version of the chipset, as you point out, Ultra is not the same chipset. Ultra for nVidia represents the 1000 HyperTransport bus, while the regular designation of nForce3-250Gb represents 800 HyperTransport. The chipset in the nVidia Reference Board we tested, as we clearly spelled out in the review, is the nForce3-250Gb Ultra used in a socket 754 configuration. Since it is used in Socket 754 it is a single-channel memory configuration, but it does offer 1000 HyperTransport.
It is true than any of the nVidia chipsets can be used in any of the Athlon 64 Socket configurations, but the available HT and features do vary, whether by design or binning. An nForce3-150 on Socket 940 boards was still 600HT.
As you point out, the names will likely be used in confusing ways, and features can be limited by the board manufacturer's design, but nForce3-250 Ultra is 1000HT, nForce3-250 is 800HT, and nForce3-150 is 600HT. The Gb represents the versions with on-chip Gigabit LAN and Firewall.
There are early 800HT versions of the 250 chip floating around, but like Intel's 875/865 the HT frequency will likely result from binning - those that fail 1000HT become 800HT versions.
Cygni - Tuesday, March 30, 2004 - link
"", on the dual-channel nForce3-250Gb Ultra version of this chipset."I want to emphasis that THERE IS NO and WILL BE NO "dual channel CHIPSET version" for A64, S754, S939, S940 or whatever. If Nvidia wants to make us to believe "buy newer and more costly nf3-250 ULTRA boards for S939" this does not mean that we (and mobo makers) should believe them and not use S939 CPUs with non-Ultra nf3s."
If you are using a Nforce3-250 chipset on a S939 board, it automatically becomes an "Ultra". Its not a different chip. AMD has announced that the A64's for Socket 939 will be dual channel, making the Nforce3-250 Ultra the member of this family that will see dual channel. In that sense, the NF3-250-U is tbe dual channel version.
wassup4u2 - Monday, March 29, 2004 - link
To clear up a little A64 confusion... Hyper Transport vs. Front Side Bus. The A64, as we all know, has an integrated memory controller. The FSB always has been, and always will be, the link between the processor and the memory. In the K8, I believe, all frequencies are determined based on multipliers of the HTT frequency. Unlike in P4/AXP systems, in which the memory controller is on a separate chip and all frequencies are multipliers of the FSB. So by increasing the FSB, you can increase memory performance. In a 2U or more config, HTT is the bus connecting the CPUs and the chipset. That's why there's not much of a performance difference between NF3-150 (600MHz HTT) and K8T800 (800MHz HTT).And enough with all the hating! This article did have its flaws, but by the way ppl are talking, you'd think it was in Chinese originally and translated by an online translator.