nForce2 - High Tras = Better?

**Robbeh** · 07-09-2003, 05:43 AM

Thats cool. I get higher mem scores, can overclock further, and can now also boot with an ATI card on my k7n2 without disabling fast boot. Goes to show how strange the nforce mem controller is, bigger latency = faster??? good work!

**SlyJavi** · 07-10-2003, 09:24 PM

WOW!!!!! Great work on this thread. I can't believe how long this has been overlooked by practically every Overclocking forum. This thread is one hell of a find for anyone OC'ing.

Big props to TheOtherDude and everyone else that posted their benchmark results.

I didn't get Prime95 happy, but I did test with 3DMark2001se.

11.5x200 w/ 222,5 timing = 17,577

11.5x200 w/ 222,11 timing = 17,676

Two thumbs up for the formula here.

**RaptorX** · 07-11-2003, 06:38 AM

great new forum

very interesting thread, so i did some tests. i ran each three times, each from a clean boot. posting the middle score, but there was only 1 or 2 points difference.

msi k7n2 delta l, xp2600+ @166 (333)
2 x 256 crucial cl2.5 non parity @166 (333)

sandra max results:

5,2,2,2 2349/2279
6,2,2,2 2353/2279
7,2,2,2 2375/2295
8,2,2,2 2373/2297
9,2,2,2 2377/2291
10,2,2,2 2376/2295
11,2,2,2 2364/2291

11's not better for me

had it on 5 before

think i'll set to the default 7.

hope this helps

**mypolar** · 07-11-2003, 01:08 PM

Originally posted by zephyr666
great new forum

very interesting thread, so i did some tests. i ran each three times, each from a clean boot. posting the middle score, but there was only 1 or 2 points difference.

msi k7n2 delta l, xp2600+ @166 (333)
2 x 256 crucial cl2.5 non parity @166 (333)

sandra max results:

5,2,2,2 2349/2279
6,2,2,2 2353/2279
7,2,2,2 2375/2295
8,2,2,2 2373/2297
9,2,2,2 2377/2291
10,2,2,2 2376/2295
11,2,2,2 2364/2291

11's not better for me had it on 5 before
think i'll set to the default 7.

hope this helps

Performance differnce caused by Tras is related to fsb as well, there was a really great analogy which is basicaly the ram is a reader and if tras is too short then it hasn't finshed the page before it is shut causing corruption. basically the higher the fsb the higher the tras for better performance.

**Drisler** · 07-20-2003, 05:47 PM

A bump and a request for a sticky for a good thread!

**sanford1** · 07-20-2003, 10:54 PM

Bump. Agreed, this one deserves a stickey.

**cjh6386** · 07-20-2003, 11:11 PM

i know if I were a mod I would sticky this.

Great work.

I can't wait to get home and try this out.

**TheOtherDude** · 07-22-2003, 01:44 AM

Originally posted by zephyr666
great new forum

very interesting thread, so i did some tests. i ran each three times, each from a clean boot. posting the middle score, but there was only 1 or 2 points difference.

msi k7n2 delta l, xp2600+ @166 (333)
2 x 256 crucial cl2.5 non parity @166 (333)

sandra max results:

5,2,2,2 2349/2279
6,2,2,2 2353/2279
7,2,2,2 2375/2295
8,2,2,2 2373/2297
9,2,2,2 2377/2291
10,2,2,2 2376/2295
11,2,2,2 2364/2291

11's not better for me had it on 5 before
think i'll set to the default 7.

hope this helps

A quick question:

Are you sure you are running in dual channel mode?

I am unfamilar with your specific baord, but most Nforce2 Ultra 400 baords can run in single channel with two sticks inserted.

TIA

**Gigantoid** · 07-23-2003, 01:54 PM

6-3-3-2 3019/2837

7-3-3-2 3023/2837

11-3-3-2 3030/2848

on 198mhz FSB with Corsair XMS PC3200

Not to great a difference but one nonetheless.

**danny5** · 07-23-2003, 03:26 PM

This trick did not work on my system either. It works best left at default ram timings.

**Fareastgq** · 07-23-2003, 05:17 PM

on mine between 8, 9, 10, 11, there is very very little difference. But I leave it at 11, it still seems like the best performance. (could be in my head, hehe). Anyway it works for me too

Keep in mind everyone's system is different, you just have to test it for yourself

**Drisler** · 07-23-2003, 05:52 PM

Ok All tests have been done!

Anyone know why the scores are better with higher TRAS.

sorry if its already been answered in this thread already. Don't feel like looking.

**TheOtherDude** · 07-23-2003, 06:10 PM

sorry if its already been answered in this thread already. Don't feel like looking.

Unlike most modern chipsets, the Nforce2 doesn't seem to make internal adjustments when you change the tRAS setting in the BIOS. These "internal" (not really sure if thats the right word) settings seem to include Bank Interleave, Burst Rate and maybe even Auto-precharge. For optimal performance, tRAS (as measured in clock cycles) should equal the sum of burst length, plus the finite time it takes the RAM to conduct a number of clock independent operations involed with closing a bank (~40 ns) minus one clock if Auto-precharge is enabled (this factor can be slightly effected by CAS, but should not play a role in optimal tRAS). To complicate things even more, one bank cannot precharge a row while the other specifies a column. This brings tRCD into the mix.

Higher isn't always better, but the reason everthing is so weird with tRAS and the Nforce2 is simply because the chipset doesn't make the internal optimizations to accomadate your inputed tRAS value like most other chipsets

**sanford1** · 07-25-2003, 12:39 AM

I also note better hard drive scores with a higher Tras...
This is 6/2/2/2 @190mHz FSB:

This is 9/2/2/2 @ 190mHz FSB:

This is on a Maxtor 6Y080P0 80 gig platter drive with the 3.44 IDE driver.

**alpha_03** · 07-25-2003, 07:42 PM

TheOtherDude,

I tried to explain this very same thing some time ago and it got me no where. What many people do not realize- the NForce mem.controller is in it's infantcy and desperately needs new drivers. I personally have spoken with NVidia about this very issue and Nvidia has confirmed my suspisions. they are to be releasing new drivers mid august or so.

PC Perspective

Thread: nForce2 - High Tras = Better?

Thread Tools

Search Thread

Display

Posting Permissions