Mem timings...

[sHape oF gReY]

Here's the deal, I try to tighten my mem timings from (see sig) to 6 2 2 2, but when I do i get a blue screen of death before i get a chance to get into windows, says something like "page_fault_in_nopage_area" and some other code(garbage) and says it would not boot into windows because it may cause problems, and "If this is the first time you have seen this message , plase restart your computer"


Any ideas? or am i stuck with my current timings?

PS tried to bump Vdimm to 2.8v but still no dice, same thing happens.

 

[hitechjb1]

From your sig, you have very nice overclocking on the CPU (2525 @ 210x12 on 1.825Vcore).

If both the CPU and the system are stable, even if you cannot get any higher, it can be considered as very good.

As far as memory timing, 5-2-2-2 or 6-2-2-2 is better than 5-3-2-2 is better than 6-3-3-2. But their difference is not much on system performance (say 1-2%). Higher FSB is more important.

I would suggest keep whatever memory timing as if first for stable system, probably 6-3-3-2 (from your sig). Work on higher FSB first (of course adjust multiplier lower accordingly).

If your motherboard is NF7-S rev 2.0, there is a good chance it should do FSB 220 - 230 MHz, or sometimes above. You may need to raise the Vmem to 2.9 V (at least for testing).

My experience with a NF7-S rev 2.0 board is that there might be a stopper for FSB 133 CPU at around 215 MHz FSB, and it might require a L12 mod to set the CPU default FSB to 166. Some found the L12 mod worked, some might not.
FSB_Sense mod to set Tbred B default FSB from 133 to 166, or to 200

 

[sHape oF gReY]

Ok, I will try to fiddle with FSB some more and see how much I can squeeze out of it. Last time I uppped the FSB above 210 had some stability issues so i backed it down a bit.

Will give it a shot after i burn in it tonight.

Thanks for the help hitechjb1

 

[hitechjb1]

FSB is highly related to memory. So to get FSB higher, you may need to increase Vmem, so memory is not holding the FSB back.

 

[sHape oF gReY]

Hmm that makes sense... because I did not increase the Vdimm when I bumped up the FSB.

 

[sHape oF gReY]

well I tried 220x11.5 on 1.875 vcore and it froze while p95 was running and 6 3 3 2 on 2.9v ram.

Then ran it at 220x11 on 1.9v || 6 3 3 2 on 2.9v and it failed one of the first 10 tests on p95.


So either i dont have enough voltage going to something or my ram just cant handle it.
gona try 215x11.5 on 1.875 with same mem timing/voltage.

Not sure what else to do at the moment...how bad of an idea would it be to bump the Vcore to 1.925 or 1.95 on my cooling?

 

[hitechjb1]

Don't mix up FSB overclocking and CPU overclocking, if both variables are changing, it is hard to determine which is the limiting factor.

Since you already have an idea that the CPU can do 2.5+ GHz at 1.825V (your earlier post said). When ever you try to increase the FSB, keep the multiplier lower accordingly so the CPU speed = FSB x multiplier is around 2.5 GHz and Vcore 1.825V, so you know the CPU speed is not causing problem. Also keep the Vmem at 2.9 V to make sure memory does not cause problem.

In my case, the NF7-S rev 2.0 was not able to go over 215 MHz stable, until I did a L12 mod for setting the FSB default to 166 MHz.
FSB_Sense mod to set Tbred B default FSB from 133 to 166, or to 200

After that mod, the FSB is able to go to 232 MHz, memory 6-3-3-2 at 2.9 V.

 

[sHape oF gReY]

well as it turns out the 2.5 wasnt at 1.825....it was at 1.875..BIG differance.

Right now im trying 215x11.5 on 1.875 Vcore... its about as close to 2.5Ghz as i can get at that FSB (if i understand your post correctly, if not set me straight ). So far seems to be runnin stable. P95 still going.




PS Vdimm is on 2.9v now.

 

[hitechjb1]

That is fine, 2.5 GHz at 1.875 V (it is about right), at 1.825 V would be too good.

When changing FSB, especially increasing FSB, keep CPU speed roughly constant or little lower is OK, as well as keeping the Vcore constant. It does not have to be exactly the same. Just to make sure the CPU is not making trouble when increasing the FSB.

Actually for testing FSB stability, you can use 3Dmark01/03 instead of Prim395. A single round of 3Dmark can quickly indicate the stability of FSB.

When you feel FSB is stable at 215 MHz, then increase it by 1 or 2 MHz at a time using 3DMark 01/03 to test it, until it starts crashing, .... Keep Vmem at 2.9 V.

Your overclocking of the whole system is looking good, just have to squeeze the last 10-15 MHz FSB out, hopefully.

 

[sHape oF gReY]

well I'll keep running it in p95 for a while longer...while the 3dmark is downloading.... yeah sucks when your cable is out (stupid lightning) and youre on 56k. 1hr 54 min remaining...

 

[sHape oF gReY]

Got it running stable in 3dMark2001 SE @218x11.5 on 1.9 Vcore. Running P95 now.



Edit

...sigh...failed the very first test in p95, back to the drawing board. Gona try bump the voltage to 1.925 see if i can stable it out.

Edit

bumped the core to 1.93 still failed just as fas, 1.95 failed...the numbers were REALLY off so imthinging its not liking the voltage that muc. Looks like its back to the good ol 215x11.5 for now. May be doing the L12 mod to see if it helps to stabilize the system.

 

[RedDawg41]

S-O-G, have you tried changing your mem timings to 2.5-3-3-8 and see if that works @225FSB... if it does, then try it up to 230FSB... if it doesn't boot up to windows, (kicks out back to post screen) then change up to 2.5.4-4-8 when you get to 230mhz... remember to adjust your multi down to 10.5 or 10 just to get into windows destop stable.

This however brings up the point of max memory through-put... Run a bench mark (sisoft) with your timings set 2-2-2-6 @ 210 or 212 (which ever was your max with your mem) then when you hit up the barrier to 230mhz, with 2.5-3-3-8 (where I'm at) do a mem benchmark again... you'll see how things drastically changes.
Question to ask yourself if its worth the drop in performance when running this high on the FSB to take a drop in memory performce? Your call.

 

[RedDawg41]

CPU-Z Report
--------------------------------------------------------------------------------
CPU-Z version 1.18.
Number of CPUs 1

Cpu Name AMD Athlon XP (Thoroughbred)
Specification AMD Athlon(tm)
Family / Model / Stepping 6 8 1
Extended Family / Model 7 8
Package Socket A
Core Stepping `B0
Technology 0.13µ
Supported Instructions Sets MMX, Extended MMX, 3DNow!, Extended 3DNow!, SSE
CPU Clock Speed 2535.7 MHz
Clock multiplier x 11.0
Front Side Bus Frequency 230.5 MHz
Bus Speed 461.0 MHz
L1 Data Cache 64 KBytes, 2-way set associative, 64 Bytes line size
L1 Instruction Cache 64 KBytes, 2-way set associative, 64 Bytes line size
L2 Cache 256 KBytes, 16-way set associative, 64 Bytes line size
L2 Speed 2535.7 MHz (Full)
L2 Location On Chip
L2 Data Prefetch Logic yes
L2 Bus Width 64 bits

Motherboard manufacturer http://www.abit.com.tw/
Motherboard model NF7-S/NF7 (nVidia-nForce2), 2.X
BIOS vendor Phoenix Technologies, LTD
BIOS revision 6.00 PG
BIOS release date 05/14/2003
Chipset nVidia nForce2 rev. C1
Southbridge nVidia MCP-T rev. A3
Sensor chip Winbond W83627HF

RAM Type DDR-SDRAM
RAM Size 512 MBytes
CAS# Latency 2.0 clocks
RAS# to CAS# 3 clocks
RAS# Precharge 3 clocks
Cycle Time (TRAS) 8 clocks

AGP Status enabled, rev. 2.0
AGP Data Transfert Rate 4x
AGP Side Band Addressing supported, enabled
AGP Aperture Size 128 MBytes

Running Dual Channel with Corsair XMS LL 2x256mb PC3200

 

[Hurk]

Good little post to read before you go further.

I don't know if you guys are aware of this but there's currently a debate on what the best Tras (row active time) setting should be (from a performance standpoint). A lot of people have been saying that 11 is the optimum value (sounds too high to me), while others have suggested that following the formula Tras = Trp + Trcd + 2 produces the optimal result (haven't seen any proof behind that equation). So I decided to test it out for myself. The following is an excerpt from a thread I started over in the AMDMB forums.

Original thread at AMDMB forums: http://www.amdforums.com/showthread.php?s=&threadid=240560

{BEGIN EXCERPT}

Some people have claimed that a Tras of 11 works better than other Tras values. Perhaps that's true on some systems, but it's definitely not true 100% of the time as some would lead you to believe. Below is my test report.

First of all, system specs:

- Abit NF7-S Rev 2.0
- Athlon XP 2500+ @ 2.2GHz (11 x 200) @ 1.7V
- OCZ PC3500 EL (512MB, single channel) @ 200MHz @ 2.6V

I decided to test the difference between 11-3-3-2 and 5-3-3-2. After I had finished, I also decided to test out 3-3-3-2 to see if that made any difference. My testing procedure was as follows:

1) go into BIOS and select desired timings
2) boot into Windows (XP Professional)
3) start up SiSoftware Sandra Standard (version 2003.7.9.73)
4) run the Memory Bandwidth Benchmark portion of Sandra
5) record the value of integer buffered memory bandwidth
6) repeat Steps 4 and 5 four times

Below are the results, expressed in megabytes per second (MB/s):

11-3-3-2
--------
2834
2839
2858
2860
2859

Average: 2850.0

5-3-3-2
-------
2833
2842
2858
2861
2860

Average: 2850.8

3-3-3-2
-------
2833
2841
2857
2860
2858

Average: 2849.8

UPDATE #1
8-3-3-2
-------
2835
2845
2857
2861
2861

Average: 2851.8


As you can see, the difference between the highest and the lowest average memory bandwidth is 1.0MB/s, which is about one thirtieth (1/30) of one percent and easily within the margin of error. The difference between 11-3-3-2 and 5-3-3-2 would be more appropriately expressed in terms of kilobytes per second than megabytes per second. Clearly then, there's effectively no difference between Tras settings of 11 and 5.

Bottom line: choose the Tras value that results in the most stable system. It sure isn't making any difference in performance.

UPDATE #1
As requested below, I ran my test procedure with a Tras set according to the following formula: Tras = Trp + Trcd + 2. Since I'm using Trp=3 and Trcd=3, that means a Tras of 8. Above is my result. As you can see, it is indeed slightly higher than the other Tras values. However, the difference is tiny: just 1.0MB/s between it and the next highest result, and just 2.0MB/s between it and the worst result. 2.0MB/s represents a difference of less than one tenth (1/10) of one percent. So this is hardly earth-shattering news, and it doesn't change the previous conclusion at all.

---------------------------

I've decided to do some additional testing, except that this time I varied CAS latency instead of Tras. Here are my results:

5-3-3-2
-------
2833
2842
2858
2861
2860

Average: 2850.8

5-3-3-2.5
--------
2780
2791
2804
2806
2806

Average: 2797.4

5-3-3-3
-------
2782
2787
2799
2806
2807

Average: 2796.2

The results are clear. There's about a 2% difference going from CL2 to CL2.5, but going from CL2.5 to CL3 produces almost no difference (less than one twentieth (1/20) of one percent). Therefore, pushing up the RAM to 204MHz at CL2.5 would provide essentially the same performance as 200MHz at CL2. The conclusion is clear: increasing RAM speed does more for performance than tightening the RAM timings. Don't be afraid to use CL2.5 or even CL3 (regardless of what the "hardcore" overclockers say about "loose" timings) if it means that you can increase the RAM speed by at least a few MHz.

---------------------------


Below are the results, as you requested. The RAM was set to 220MHz @ 2.8V, and everything else was untouched.

11-3-3-2
--------
3097
3105
3123
3124
3125

Average: 3114.8

5-3-3-2
-------
3092
3103
3121
3121
3121

Average: 3111.6

The difference is 3.2MB/s, or slightly over one tenth (1/10) of one percent. Again, the difference is so tiny that it's easily within the margin of error. Thus we can conclude that even at high RAM speeds, my previous conclusion stands: there's essentially no performance difference between a Tras of 11 and a Tras of 5, so you should pick based on which value gives you the most stability.

{END EXCERPT}

 

[hitechjb1]

Interesting test results, it seems to be consistent with that described in these links.

RAS, CAS Timing and BIOS Memory Setting

Single Channel vs Dual Channel DDR Memory

 

[hitechjb1]

Got it running stable in 3dMark2001 SE @218x11.5 on 1.9 Vcore. Running P95 now.



Edit

...sigh...failed the very first test in p95, back to the drawing board. Gona try bump the voltage to 1.925 see if i can stable it out.

Edit

bumped the core to 1.93 still failed just as fas, 1.95 failed...the numbers were REALLY off so imthinging its not liking the voltage that muc. Looks like its back to the good ol 215x11.5 for now. May be doing the L12 mod to see if it helps to stabilize the system.

Since you are keeping CPU clock frequency roughly constant at 2.5 GHz, and aiming at pushing FSB higher, Vcore increase won't help much. Keep Vcore back to where is were.

So it looks like the FSB has problem getting over 215-218 MHz.

You may try:

- changing memory timing to 7-3-3-2.5 or 7-3-3-3, if it helps that means memory is limiting at 6-3-3-2 (but I don't think this is the case for TwinMOS winbond CH-5 chip at 2.9 V)

- consider doing the FSB_Sense mod to change the CPU default FSB from 133 to 166, which help me to get to 234 MHz from 215 MHz

PS: Make sure the chipset Vdd is set to 1.7 V (bios max)