DFI UT P35-T2R: Tweakers Rejoice!
by Rajinder Gill on October 18, 2007 2:00 PM EST- Posted in
- Motherboards
All Other Sub-Timings-
Using AUTO for all sub-timings is generally acceptable until we start pushing for every last MHz of bandwidth and latency. The BIOS defaults are generally tight enough for very good performance on 24/7 PCs. The BIOS profiles we are including in this article have been tuned for maximum performance while retaining full system stability. We used Memset 3.3 while running Microsoft XP to check where the AUTO DRAM timing defaults set the memory parameters. The timings shown by Memset were noted by us and entered manually into the relevant BIOS functions. We then began to tighten these values one at a time, noting the effects of any changes on bandwidth, latency, and stability.
For benchmarking we tend to use Memset once the XP operating system has loaded rather than entering overly tight settings directly in the BIOS. We can then begin to tighten the sub-timings far beyond 24/7 settings. We have attached Memset screenshots for both our 24/7 stable and sub-zero benchmarking so that users can see how far we were able to push for maximum scores.
Additional Setting Ranges:
MCH ODT Latency: ODT (On Die Termination) is used to match the MCH output impedance to the termination resistance of the RAM. Use of 1-2 is fine in most cases. Even for 4GB overclocking we did not find that more than a setting of 1 was required.
Write to PRE Delay (tWR): Range is from 10-13. For high RAM speed, use 12 and above for stability; for benchmarking 10-12 will be faster. Lower is faster.
Rank Write to Read (tWTR): Range is 9-11. Use 11 for stability and 10 or 9 for benchmarking. Lower is faster.
ACT to ACT Delay (tRRD): Range is 2-5. 3 is a good setting that allows high enough FSBs for most users. Higher numbers result in less memory read performance.
Read to Write Delay (tRDWR): Range is 8-9. 8 is ideal for most users; higher values are slower.
Ranks Write to Write (tWRWR): AUTO
Ranks Read to Read (tRDRD): AUTO
Ranks Write to Read (tWRRD): AUTO
The ranges for all these settings are quite well manipulated by SPD and BIOS. For those wishing to experiment, 4 is the lowest and most aggressive setting. 5-4-5 may work for SuperPi type benching. There is a slight write/copy speed increase with tighter settings, but this is only useful for the extracting the last ounce of performance. For stability these are best left on AUTO
Read CAS# Precharge (tRTP): 2-3 works well on this board for most users. 2 is used for benchmarking while 3 and upwards are best for stability.
ALL PRE to Refresh: Range is 4-5. 4 works fine with quad-core CPUs all the way to 475 FSB. A setting of 4 is fine for most benchmarking, unless shooting high over 500FSB.
Using AUTO for all sub-timings is generally acceptable until we start pushing for every last MHz of bandwidth and latency. The BIOS defaults are generally tight enough for very good performance on 24/7 PCs. The BIOS profiles we are including in this article have been tuned for maximum performance while retaining full system stability. We used Memset 3.3 while running Microsoft XP to check where the AUTO DRAM timing defaults set the memory parameters. The timings shown by Memset were noted by us and entered manually into the relevant BIOS functions. We then began to tighten these values one at a time, noting the effects of any changes on bandwidth, latency, and stability.
For benchmarking we tend to use Memset once the XP operating system has loaded rather than entering overly tight settings directly in the BIOS. We can then begin to tighten the sub-timings far beyond 24/7 settings. We have attached Memset screenshots for both our 24/7 stable and sub-zero benchmarking so that users can see how far we were able to push for maximum scores.
Additional Setting Ranges:
MCH ODT Latency: ODT (On Die Termination) is used to match the MCH output impedance to the termination resistance of the RAM. Use of 1-2 is fine in most cases. Even for 4GB overclocking we did not find that more than a setting of 1 was required.
Write to PRE Delay (tWR): Range is from 10-13. For high RAM speed, use 12 and above for stability; for benchmarking 10-12 will be faster. Lower is faster.
Rank Write to Read (tWTR): Range is 9-11. Use 11 for stability and 10 or 9 for benchmarking. Lower is faster.
ACT to ACT Delay (tRRD): Range is 2-5. 3 is a good setting that allows high enough FSBs for most users. Higher numbers result in less memory read performance.
Read to Write Delay (tRDWR): Range is 8-9. 8 is ideal for most users; higher values are slower.
Ranks Write to Write (tWRWR): AUTO
Ranks Read to Read (tRDRD): AUTO
Ranks Write to Read (tWRRD): AUTO
The ranges for all these settings are quite well manipulated by SPD and BIOS. For those wishing to experiment, 4 is the lowest and most aggressive setting. 5-4-5 may work for SuperPi type benching. There is a slight write/copy speed increase with tighter settings, but this is only useful for the extracting the last ounce of performance. For stability these are best left on AUTO
Read CAS# Precharge (tRTP): 2-3 works well on this board for most users. 2 is used for benchmarking while 3 and upwards are best for stability.
ALL PRE to Refresh: Range is 4-5. 4 works fine with quad-core CPUs all the way to 475 FSB. A setting of 4 is fine for most benchmarking, unless shooting high over 500FSB.
Facebook
Twitter
RSS
30 Comments
View All Comments
Rocket321 - Tuesday, October 23, 2007 - link
I would like to see the overclocking results put into a graph or chart of some kind.I guess tweakers might like the screenshots as "proof" that the overclock ran, but personally I trust you and would rather just have one place to look rather than clicking to enlarge multiple screenshots sequentially.
It was a great review though, I look forward to the future tweaker guides & reviews.
Raja Gill - Wednesday, October 24, 2007 - link
There will be a more typical Anandtech look to future articles, with the DFI board revisited for a round up on a suite of benchmarks, this will be used to cross compare with other boards, clocked to equivalent with board maximums in the range and voltage. In terms of the screenshots, it is nice to have 1 persons trust, but there are many we have to please and not everyone is always as convinced..thanks for the suggestions..
Next up is the Asus Maximus Formula..
regards
Raja
Jodiuh - Thursday, October 25, 2007 - link
Specifically Windvd conversions from divx/xvid to DVD would be wonderful as I've found this app benefits from a solid OC.beoba - Friday, October 19, 2007 - link
It'd be great if this came with a glossary."Strap"?
retrospooty - Saturday, October 20, 2007 - link
strap is a term used for memory clocking. for example, at 266mhz bus, memory can be "strapped" to one of the following.266x(stap2)=533 or DDR 1066
266x(strap1.5)=400 or DDR 800
266x(strap1.25)=333 or DDR 666
If you are running at stock 266 there is no way to have DDR 950 because it has to be strapped to one of the above settings.
I use the 1/1 strap so my bus speed is 500x(strap1)=500 or DDR 1000, in most cases 1/1 is the most efficient, if you can utilize it with your particular hardware, do it.
Avalon - Thursday, October 18, 2007 - link
Unfortunately, DFI's asking price of admission continues to rise for each new board they release. I was mildly annoyed when they started selling boards for $200+ that had little to no tangible benefit over $100-$150 boards, but now they're at the $300 mark? No thanks.This board is for someone who likes to spend his time tweaking and not actually using his computer.
retrospooty - Saturday, October 20, 2007 - link
"This board is for someone who likes to spend his time tweaking and not actually using his computer."The article title is called "Tweakers Rejoice" after all. The idea is not to tweak forever . I did spend alot of time over the first few weeks, but now that its tweaked, I just use it as is.
Avalon - Sunday, October 21, 2007 - link
Yes, I am quite capable of reading the article title. My whole point is that you are working for diminishing returns that I feel could be better spent using your system. If you're doing it to set a record, fantastic. I support that.retrospooty - Sunday, October 21, 2007 - link
Understood... This is obviously not the motherboard for you. I personally love the BIOS options and CMOS reloaded functionality. That alone makes the extra cost well worth it to me. Asus BIOS just sucks, and I have had too many quality problems with them in the past, and Gigabyte just underperforms. I like to know I will not be held back by my motherboard for the next couple of CPU's I buy (will likely get a dual core Penryn on release for under $200, then a high end quad core Penryn a year or so later when it is under $200).I do see your point, but in spite of this article's stock speed comparison (totally pointless for a OC geared mobo), and similar results with one particular CPU, that looks as if it has an FSB limit equal on all 3 boards (meaning the CPU is holding it back) This board overclocks and performs better than any ASUS, or Gigabyte, or any other board out there. If AT tested the max FSB limits on a dozen or so CPU's, or if they had a "golden sample" that had a high FSB limit, you would see the difference. Also if they had time to test many diff RAM stocks and to tweak the memory settings you would also see the difference. Its a good article, but no reviewer has time to really dig into this mobo and all it has to offer. I do feel the article did a good job at explaining that.
JNo - Sunday, October 21, 2007 - link
"I like to know I will not be held back by my motherboard for the next couple of CPU's I buy"I don't know much about overclocking and I hold your views valid Retrospooty but surely this is still a lot of money that will still be needed to replaced in the short/medium-term if a) you want to start using DDR3 once prices come down b) if GPUs come out that take advantage of PCI-E 2 standard (as used on X38). So all that money is only paying for great OC'ing potential for *now* only... no?