Need Help with Ram Timings...

Alright first off I will list the specs of my computer so it will be easier to answer my question.

Intel Core2 Duo E6600 @ 2.4GHZ
Asus P5W DH-Deluxe Motherboard
Cosair XMS2 2X512 1GB PC2-6400 @ 800MHZ

the default timings of this rams is rated at 5-5-5-12, but i am running it at 4-4-4-4 and its stable. Now i found out this new program called Memset and it has these few options under subtimings I was wondering what would be the best configration for them as they are right now set to...

Page Close Idle Timer...................... 16T
Refresh Cycle Time (tRFC)................ 42
Read DElay (tRD)............................. 6
Write to Precharge Command............ 13
Write to Read Command................... 10
Read-Write Turnaround clocks........... 8
Write-Read Turnaround clocks........... 6
Read to Precharge............................ 5
Write-Write Turnaround clocks........... 6

Now I was also wondering if I should even bother with those. If i don't really have to, then is there anyone here who have tested these rams stable with lower latency timings?
15 answers Last reply
More about need timings
  1. Quote:
    Alright first off I will list the specs of my computer so it will be easier to answer my question.

    Intel Core2 Duo E6600 @ 2.4GHZ
    Asus P5W DH-Deluxe Motherboard
    Cosair XMS2 2X512 1GB PC2-6400 @ 800MHZ

    the default timings of this rams is rated at 5-5-5-12, but i am running it at 4-4-4-4 and its stable. Now i found out this new program called Memset and it has these few options under subtimings I was wondering what would be the best configration for them as they are right now set to...

    Page Close Idle Timer...................... 16T
    Refresh Cycle Time (tRFC)................ 42
    Read DElay (tRD)............................. 6
    Write to Precharge Command............ 13
    Write to Read Command................... 10
    Read-Write Turnaround clocks........... 8
    Write-Read Turnaround clocks........... 6
    Read to Precharge............................ 5
    Write-Write Turnaround clocks........... 6

    Now I was also wondering if I should even bother with those. If i don't really have to, then is there anyone here who have tested these rams stable with lower latency timings?
    Better to just leave those other timings where they are. The main 4 timings(4-4-4-12) are the most important and most adjusted timings.
  2. so u saying 4-4-4-12 will work better than 4-4-4-4?
  3. Quote:
    so u saying 4-4-4-12 will work better than 4-4-4-4?
    I doubt you can even set it to 4-4-4-4. 4-4-4-12 is pretty much the standard for DDR2, with some fast modules running @ 3-3-3- 8, and slower(or very high speed) ones running 5-5-5-15.
  4. Cycle Time (Tras) is shown in a memory module's Serial Presence Detect (SPD) EPROM as a value of 12, so as to afford an adequate time interval for row and column cycles to complete. I'm surprised that it's stable at 4-4-4-4. If you benchmark it, I'm sure you'll find that 4-4-4-12 is faster. In any case, memory priorities are as follows:

    (1) Quantity (1Gb or 2Gb)

    (2) Bandwidth (Single or Dual Channel)

    (3) Frequency (667Mhz or 800Mhz)

    (4) Ratio ( 1:1 or X:X)

    (5) Latency (5-5-5-15 or 4-4-4-12)

    Unlike the older DDR platforms, DDR2 timings have little impact on overall system performance.
  5. Quote:
    ...
    Unlike the older DDR platforms, DDR2 timings have little impact on overall system performance.

    It really depends on the chipset, CPU, and MB design. Timings can easily have 5-10% effects on performance for C2D systems, which is of similar magnitude to changing DDR2 speed by 133MHz (effective).
  6. Quote:
    ...
    Cosair XMS2 2X512 1GB PC2-6400 @ 800MHZ
    Does that mean you are running the memory at DDR2-800 speed?
    Quote:

    the default timings of this rams is rated at 5-5-5-12,

    Do you mean the SPD value, or do you mean the rating on the spec sheet from Corsair (presumably at DDR2-800 speed)?
    Quote:

    but i am running it at 4-4-4-4 and its stable.

    How do you know it is stable? Just booting and running isn't a good test -- you need to pass memtest86+ and a good run of Orthos (or pair of PRIME95).

    This post may help get you started:
    http://forumz.tomshardware.com/hardware/modules.php?name=Forums&file=viewtopic&p=1249881#1249881
  7. Quote:
    It really depends on the chipset, CPU, and MB design. Timings can easily have 5-10% effects on performance for C2D systems, which is of similar magnitude to changing DDR2 speed by 133MHz (effective).


    Mondoman, you are correct, in terms of the hardware spectrum. However, the OP's question is confined to 5-5-5-12 vs 4-4-4-12 on his PC. Within the constrants of the OP's scenario, benchmarks will show those DDR2 timings increase memory performance by less than 2%, which again, has "little impact on overall system performance".

    Check out the following THG article regarding memory frequency and timings:

    http://www.tomshardware.com/2006/03/31/tight_timings_vs_high_clock_frequencies/

    Edit: I ran memory benchmarks on my rig with PCMark04, 3 tests at 4-4-4-12, and 3 at 5-5-5-12. The average of the first 3 results divided by the average of the second 3 results are 12695 / 12550 = 1.16% It's still better than 0%, so we'll take whatever we can get. :D
  8. Yeah, I suppose I was missing some information there eh. Well, its Dual-Channel and default is 800Mhz. Then like in Memtest86+ manual say run it until its 100 percent I did that it was stable, then I ran it for few hours it was stable, and then I left it over night it was stable no errors found and computer didn't froze or rebooted. And I thought for Intel's CPUs its is good to have low latency timings since they don't have integrated latency thing. Aright now I don't have good cooler with me so I can't overclock my processor though i want to and make it 1:1 ratio but I gotta wait a month for that. Its running at 2:3 ratio right now.
  9. That's amazing it ran stable at 4-4-4-4. Of course, it's always preferable to run lower latencies if possible, however, 4-4-4-12 is appropriate, and benchmarks faster. You'll find that if you tighten Cycle Time (Tras) too much, it reduces performance, and can potentially cause crashes.
  10. So I should change it to 12 from 4 and see if that works better.
  11. No, 4-4-4-12 is correct.
  12. I see ur spec it says Zalman 9500 - AS5 as ur cooler how much temperature u get on that thing running it at 3.7Ghz and how much it did cost?
  13. Quote:
    I see ur spec it says Zalman 9500 - AS5 as ur cooler how much temperature u get on that thing running it at 3.7Ghz and how much it did cost?


    38c idle, 52c full load @ 3.7Ghz, 1.5125 vCore BIOS, 1.480 ~ 1.504 vCore (24mV fluctuation) measured SpeedFan 4.31.

    Zalman 9500 $50.00.

    AS5 = Arctic Silver 5 Thermal Compound.
  14. 50 dollars that seems a little too expensive, though I have seen some cooler at 30 dollars doing the same job.
  15. There are considerations beyond cost and CPU cooling alone. The Zalman's radial design has an airflow disbursal pattern that cools components adjacent to the CPU socket, such as the northbridge and voltage regulators.

    http://www.frostytech.com/articleview.cfm?articleid=2045&page=4

    http://www.madshrimps.be/?action=getarticle&number=11&artpage=2191&articID=519
Ask a new question

Read More

Memory