+ Reply to Thread
Page 1 of 2 1 2 LastLast
Results 1 to 10 of 13

Thread: RAM Speed VS Latency

  1. #1
    ROG Renaissance Man Florre +10 Florre's Avatar
    Join Date
    Aug 2012
    Posts
    38

    RAM Speed VS Latency

    What is more important ram speed or latency?

    2400mhz with latency 10-11 or 1866mhz with latency 9?
    CPU Intel i7 980x @ 4.4Ghz
    RAM Kingston 1600Mhz 12GB (KHX1600C9D3K3/12GX)
    GPU 2 X MSI 580 GTX in SLI
    SSD NONE
    HDD WD Velociraptor 600GB and 2 X WD RE2
    MOBO Asus Rampage III Extreme
    PSU Thermaltake Toughpower 1200W
    CASE Corsair 700D

  2. #2
    For what ?
    Gaming with a PCI-Express VGA ? None.
    Desktop usage ? None.
    A few applications ( WinRAR, 7-zip, WinZIP, KGB, SAP, high traffic SQL Server ) ? The best of both worlds, and also depends on the platform ( chipset + CPU )

    If you're talking about your system, and your primary usage ( which is gaming most likely with your GTX 580 SLI ) then... get a decent DDR3-1600 memory kit with the capacity of your desire ( anything from 6GB and more is the top you can get for gaming performance ).

    Data and proof to back it up ? Sure, go ahead and read my extended testing article

    http://www.benchzone.com/page.php?al...formance_part1

  3. #3
    ROG Guru: Diamond Belt Zka17 +150 Zka17 +150 Zka17's Avatar
    Join Date
    Mar 2012
    Location
    Mount Pleasant, SC, USA
    Posts
    7,979

    Regarding the system memory the most important is to have it running stable... a bad memory can really mess up things... - whatever settings you choose, run memtest86+ for at least 4 cycles, if it passes, then most likely you're fine...

    If you're in benching, then 2400MHz CL10 is better than 1866MHz CL9... - just be aware that not all systems are the same, and there is no guarantee that you can actually reach those memory speeds... (it's about the chip's lottery)

  4. #4
    ROG Bodhi: Orion's belt Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm's Avatar
    Join Date
    Apr 2012
    Location
    Downwind and in range
    Posts
    8,481

    There is a formula that can help you in deciding if one kit is faster than the other when the timings are taken into account.

    (CAS/frequency) x 2000 gives you a number in nanoseconds for comparing speeds.

    Say you see a 1600 kit at 7-8-7-20 and a 1866 kit at 9-10-9-28 which is faster?
    Well 7/1600 x 2000 = 8.75 nanoseconds; 9/1866 x 2000 = 9.646 so the 1600 kit actually looks quicker.

    If these numbers are close or the same choose the higher frequency it will be more efficient.

    Phobya Bench Table / Rampage IV Black Edition / i7 4930K Stock - 4.9GHz / 16GB Corsair Dominator Platinum 2133 CMD16GX3M4A2133C9
    ASUS GTX Titan Black or ASUS Matrix Platinum R9 290X / ASUS Xonar Essence STX
    1x OCZ Vector 150 240GB, 2x OCZ Vertex 3 240GB, 2x WD Caviar Black 500GB
    Custom loop: EK Supreme HF Full EN /EK-FC BLOCK for TI/TITAN / 2x EK Coolstream XTX 360
    Dual D5 Varios with EK D5 X-Tops / EK- RES X3 400
    Dell U2711



  5. #5
    ROG Guru: Diamond Belt Zka17 +150 Zka17 +150 Zka17's Avatar
    Join Date
    Mar 2012
    Location
    Mount Pleasant, SC, USA
    Posts
    7,979

    Arne, what you calculated there isn't the memory latency? If so, higher latency doesn't mean that your RAM is actually slower?

    So, 9.646 ns is actually worse than 8.75 ns... the first takes more time...

  6. #6
    ROG Bodhi: Orion's belt Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm's Avatar
    Join Date
    Apr 2012
    Location
    Downwind and in range
    Posts
    8,481

    Correct! the latency is the CAS no..... the formula uses latency not calculates it......it lets you know which kit is faster when comparing different speed RAM with different latencies.... I thought Florre might have been wanting to know that (maybe my mistake).

    1866 cas 9.. 9.6462 slow
    2400 cas 11.. 9.1666 faster
    2400 cas 10.. 8.3332 fastest
    Last edited by Arne Saknussemm; 03-29-2013 at 06:57 PM.

    Phobya Bench Table / Rampage IV Black Edition / i7 4930K Stock - 4.9GHz / 16GB Corsair Dominator Platinum 2133 CMD16GX3M4A2133C9
    ASUS GTX Titan Black or ASUS Matrix Platinum R9 290X / ASUS Xonar Essence STX
    1x OCZ Vector 150 240GB, 2x OCZ Vertex 3 240GB, 2x WD Caviar Black 500GB
    Custom loop: EK Supreme HF Full EN /EK-FC BLOCK for TI/TITAN / 2x EK Coolstream XTX 360
    Dual D5 Varios with EK D5 X-Tops / EK- RES X3 400
    Dell U2711



  7. #7
    ROG Renaissance Man Florre +10 Florre's Avatar
    Join Date
    Aug 2012
    Posts
    38

    Quote Originally Posted by Arne Saknussemm View Post
    Correct! the latency is the CAS no..... the formula uses latency not calculates it......it lets you know which kit is faster when comparing different speed RAM with different latencies.... I thought Florre might have been wanting to know that (maybe my mistake).

    1866 cas 9.. 9.6462 slow
    2400 cas 11.. 9.1666 faster
    2400 cas 10.. 8.3332 fastest


    Thanks for the formula. Its proving to be damn useful!

    Thx.
    CPU Intel i7 980x @ 4.4Ghz
    RAM Kingston 1600Mhz 12GB (KHX1600C9D3K3/12GX)
    GPU 2 X MSI 580 GTX in SLI
    SSD NONE
    HDD WD Velociraptor 600GB and 2 X WD RE2
    MOBO Asus Rampage III Extreme
    PSU Thermaltake Toughpower 1200W
    CASE Corsair 700D

  8. #8
    ROG Bodhi: Orion's belt Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm's Avatar
    Join Date
    Apr 2012
    Location
    Downwind and in range
    Posts
    8,481

    Good stuff!

    Phobya Bench Table / Rampage IV Black Edition / i7 4930K Stock - 4.9GHz / 16GB Corsair Dominator Platinum 2133 CMD16GX3M4A2133C9
    ASUS GTX Titan Black or ASUS Matrix Platinum R9 290X / ASUS Xonar Essence STX
    1x OCZ Vector 150 240GB, 2x OCZ Vertex 3 240GB, 2x WD Caviar Black 500GB
    Custom loop: EK Supreme HF Full EN /EK-FC BLOCK for TI/TITAN / 2x EK Coolstream XTX 360
    Dual D5 Varios with EK D5 X-Tops / EK- RES X3 400
    Dell U2711



  9. #9
    ‎‎‎‎‎‎‎‎‎‎ Gorman +75 Gorman's Avatar
    Join Date
    Apr 2012
    Posts
    628

    Quote Originally Posted by Arne Saknussemm View Post
    (CAS/frequency) x 2000 gives you a number in nanoseconds for comparing speeds.
    Could you explain this formula?

    Looking at the units I see CAS is in clock cycles, frequency is mega clock cycles per second, so the formula yields an answer in micro seconds.
    So where does the 2 come from? I'm guessing this is an approximate formula, and the real one uses a second point of data; (CAS+???)/F * 1000 ?

  10. #10
    ROG Bodhi: Orion's belt Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm +300 Arne Saknussemm's Avatar
    Join Date
    Apr 2012
    Location
    Downwind and in range
    Posts
    8,481

    Not thought about it much... might hide a multitude of simplification but works as a comparative for kits I believe.

    Hmm but If CAS is measured in nano seconds divide by x thousand times y thousand still nanoseconds no?

    http://rog.asus.com/forum/showthread...l=1#post195744

    Phobya Bench Table / Rampage IV Black Edition / i7 4930K Stock - 4.9GHz / 16GB Corsair Dominator Platinum 2133 CMD16GX3M4A2133C9
    ASUS GTX Titan Black or ASUS Matrix Platinum R9 290X / ASUS Xonar Essence STX
    1x OCZ Vector 150 240GB, 2x OCZ Vertex 3 240GB, 2x WD Caviar Black 500GB
    Custom loop: EK Supreme HF Full EN /EK-FC BLOCK for TI/TITAN / 2x EK Coolstream XTX 360
    Dual D5 Varios with EK D5 X-Tops / EK- RES X3 400
    Dell U2711



+ Reply to Thread
Page 1 of 2 1 2 LastLast

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts