Socket 939 Roundup: Battle at the Top
by Wesley Fink on July 30, 2004 8:00 AM EST- Posted in
- Motherboards
Gigabyte K8NSNXP-939: Overclocking and Stress Testing
FSB Overclocking Results
| Front Side Bus Overclocking Testbed | ||
| Processor: | Athlon 64 FX53 Socket 939 2.4GHz |
|
| CPU Voltage: | 1.5V (default) | |
| Cooling: | Thermaltake Silent Boost K8 | |
| Power Supply: | Antec TruePower 430W | |
| Maximum OC: (Standard Ratios) |
204FSB x13 2652MHz (+10.5%) |
|
| Maximum FSB: (Lower Ratio) |
267FSB x 9 at 1:1 Memory | |
Our FX53 easily achieved the next speed level of 2.6GHz on the Gigabyte. In fact, the two nVidia nF3 boards tied for the highest overclock of this FX53 at 2652MHz. This is a bit faster than the 2.6GHz, which will likely be the speed rating of the FX55 to appear later this year. On recent Gigabyte boards, the BIOS overclocks have been more modest, with the rest of the overclock only doable with EZTune in Windows. It is good to see Gigabyte allowing a higher OC in BIOS, since we could achieve 267 in BIOS at lower multipliers. However, any higher value requires EZTune after booting into the OS. We still did not reach spectacular levels with EZTune, topping out about 280, but these levels should be more than enough to satisfy most users.
Memory Stress Test Results:
Since you need a score card for Dual Channel setup on 939, DIMMs 1 and 2 are the first DC bank on the Gigabyte, and yes, this can be very confusing. Our advice is to check memory and your start-up screen carefully to make sure that you are really in Dual-Channel mode when setting up a 939. Everyone seems to be doing it differently.The memory stress test is very basic, as it simply tests the ability of the Gigabyte K8NSNXP-939 to operate at its officially supported memory frequency (400MHz DDR), at the best performing memory timings that our Mushkin PC3500 Level 2 or OCZ PC3500 Platinum Ltd Modules will support. Memory stress testing was conducted by running RAM at 400MHz with 2 DIMM slots in a Dual-Channel configuration.
| Stable DDR400 Timings - 2 DIMMs (2/4 DIMMs - 1 Dual-Channel Bank) |
|
| Clock Speed: | 200MHz |
| Timing Mode: | N/A |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 10T* |
| Precharge Delay: | 2T |
| Command Rate: | 1T |
The Gigabyte 939 was completely stable with two DIMMs in Dual-Channel mode at the best 1T Command Rate settings.
Filling all four available memory slots is more strenuous on the memory subsystem than testing 2 DIMMs on a motherboard.
| Stable DDR400 Timings - 4 DIMMs (4/4 DIMMs - 2 Dual-Channel Banks) |
|
| Clock Speed: | 200MHz |
| Timing Mode: | N/A |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 10T* |
| Precharge Delay: | 2T |
| Command Rate: | 2T |
The K8NSNXP-939 was able to run with all 4 DIMM slots at the same aggressive 2-2-2-10 settings used for 2 DIMMs. However, as we have seen on the other 939 boards running all 4 DIMMs, Command Rate must be reduced to 2T when filling both Dual Channels.
Facebook
Twitter
RSS
83 Comments
View All Comments
Klaasman - Wednesday, July 21, 2004 - link
#32 Socket 940 boards require ECC memory. 939 don't but they might run it. Go to AMD's website and see.FactorOfTwo - Wednesday, July 21, 2004 - link
Do any Socket 939 boards support ECC memory? I am having a hard time finding a definitive answer to this question.TheLiquidH20 - Wednesday, July 21, 2004 - link
Quote - The general performance of the VIA and nVidia-based 939 boards was virtually the same in DirectX 9 games, with one notable exception. Microsoft's Halo performs almost 15% better on the nVidia nForce3 Ultra .Could this have something to do with Halo being a direct port of the xbox ? Seeing that the Xbox is basically one , big intergrated nForce ? . Would explain microsoft adding some optimazations for niVdia hardware .
Klaasman - Wednesday, July 21, 2004 - link
Ive got an ABIT AV8 and I wouold like to get a copy of that 1.3 bios you claim you had.rjm55 - Wednesday, July 21, 2004 - link
This is the first time I've seen AnandTech use color in the benchmark graphs. It really makes reading the data a LOT easier. Thanks!Wesley Fink - Tuesday, July 20, 2004 - link
#5 and #26 - I received a retail K8T Neo2 about 10 days ago and the board is now entering retail. I received the retail K8N Neo2 about 4 days ago and I am told it should enter retail in the next week to 10 days. That is the best information I have, but the date has already slipped from early July to the end of July. MSI should have the most reliable information on when the board will finally hit the retail channel.kmmatney - Tuesday, July 20, 2004 - link
Thanks for using the AutoGK encoding benchmark! I prefer the XVid codec, over DivX, but I very much appreciate the benchmark. AutoGK is by far the best freeware encoding suite I've come across.kd4yum - Tuesday, July 20, 2004 - link
See #5ibid
Where is the MSI 939 board?!
Wesley, I asked same question in another Comments section. I can't get answers from MSI (phone) or Newegg (phone). Can you?
Wesley Fink - Tuesday, July 20, 2004 - link
#23 - I am now working on a 925X roundup and had switched mental gears. It looks like I need to check my mind set before correcting reviews :-) Now fixed.kd4yum - Tuesday, July 20, 2004 - link