Results 1 to 11 of 11

Thread: Htt v's Multiplier

  1. #1
    Xtreme Member
    Join Date
    Nov 2009
    Location
    Gosport, Hampshire UK
    Posts
    220

    Htt v's Multiplier

    I know this has been covered a lot just lately,have been playing with the CH4 formula and 1090T,i know my clocks are not xtreme as yet,but getting there,lol!!.

    I ran 06,vantage,cpu only tests,best of 3 each time,with first up CPU-NB and HT link both @3.0ghz,HTT @250,with my ram as close as possible frequency wise,@6-6-6-18-24.Then i ran Everest memory benchmark,and 32m superPI.

    My personal conclusion is that keeping the CPU-NB and HT link on par gives far better memory latency and also helps somewhat in 3d and 2d benches,sorry to cover this subject again.

    CPU-NB and HT link @3ghz HTT @250





    CPU-NB @3ghz HTlink stock,HTT@250



    Last edited by moorhen2; 05-31-2010 at 01:31 PM.
    ASUS RAMPAGE V EXTREME, I7 5960X, 16Gig GSKILL DDR4 2666, CUSTOM WATERCOOLING[/SIZE]

  2. #2
    Xtreme 3D Team
    Join Date
    Jan 2009
    Location
    Ohio
    Posts
    8,499
    You mean CPUNB clock

    HTT is the Bus Speed, Base Frequency, Base Clock, HT Ref. Clock...whatever you want to call it.
    Smile

  3. #3
    Xtreme Member
    Join Date
    Nov 2009
    Location
    Gosport, Hampshire UK
    Posts
    220
    Quote Originally Posted by BeepBeep2 View Post
    You mean CPUNB clock

    HTT is the Bus Speed, Base Frequency, Base Clock, HT Ref. Clock...whatever you want to call it.
    Yeh,sorry,my mistake,corrected now.I am old school,FSB,but AMD now reffer to it as HTT.
    ASUS RAMPAGE V EXTREME, I7 5960X, 16Gig GSKILL DDR4 2666, CUSTOM WATERCOOLING[/SIZE]

  4. #4
    I am Xtreme
    Join Date
    Feb 2003
    Location
    Netherlands
    Posts
    6,421
    Thanks for showing. The difference in Everest latency is quite noticable to say the least!
    Asus Crosshair V Formula-Z | FX 8350 | 2x4GB Trident-X 2600 C10 | 2x ATI HD5870 Crossfire | Enermax Revo 1050watt | OCZ Vertex 3 60GB | Samsung F1 1TB

    Watercooling: XSPC Raystorm | EK 5870 Delrin fullcover | TFC X-changer 480 w/ 4x Gentle Typhoon | DDC2+ Delrin top | EK 200mm res | Primochill LRT 3/8 tubing

    Case: Murdermodded TJ-07

    sub 9 sec. SPi1M 940BE 955BE 965BE 1090T

  5. #5
    Xtreme Addict
    Join Date
    Dec 2002
    Location
    Sweden
    Posts
    1,261
    indeed, zeus.. 32 ns mums
    Vishera 8320@ 5ghz | Gigabyte UD3 | 8gb TridentX 2400 c10| Powercolor 6850 | Thermalight Silver Arrow (bench Super KAZE 3k) | Samsung 830 128gbx2 Raid 0| Fractal case

  6. #6
    Xtreme Addict
    Join Date
    Jun 2007
    Posts
    2,064
    what cooling methods used to get this result?

  7. #7
    Xtreme Member
    Join Date
    Nov 2009
    Location
    Gosport, Hampshire UK
    Posts
    220
    Quote Originally Posted by Zeus View Post
    Thanks for showing. The difference in Everest latency is quite noticable to say the least!
    Thanks Zeus,i ran each test 3 times and Everest was much the same,with the NB and HT link the same,around the 32ns mark each time,36ns with HT link @ stock,the only test that was slower was superPI 32M,which was slightly faster with HT link at stock.
    ASUS RAMPAGE V EXTREME, I7 5960X, 16Gig GSKILL DDR4 2666, CUSTOM WATERCOOLING[/SIZE]

  8. #8
    Xtreme Member
    Join Date
    Nov 2009
    Location
    Gosport, Hampshire UK
    Posts
    220
    Quote Originally Posted by Serpentarius View Post
    what cooling methods used to get this result?
    I am under water,ram is just under the Corsair fancooler.
    ASUS RAMPAGE V EXTREME, I7 5960X, 16Gig GSKILL DDR4 2666, CUSTOM WATERCOOLING[/SIZE]

  9. #9
    Xtreme Cruncher
    Join Date
    Nov 2002
    Location
    Belgium
    Posts
    605
    Thanks for the tests , now i want to bump my HTT again . I have one problem , when i set cpu NB multiplier to 15 (3075mhz) , i can only go up to 13x (2665mhz) on my HT multiplier ..

    any idea why this could be ?


    Main rig 1: Corsair Carbide 400R 4x120mm Papst 4412GL - 1x120mm Noctua NF-12P -!- PC Power&Cooling Silencer MK III 750W Semi-Passive PSU -!- Gigabyte Z97X-UD5H -!- Intel i7 4790K -!- Swiftech H220 pull 2x Papst 4412 F/2GP -!- 4x4gb Crucial Ballistix Tactical 1866Mhz CAS9 1.5V (D9PFJ) -!- 1Tb Samsung 840 EVO SSD -!- AMD RX 480 to come -!- Windows 10 pro x64 -!- Samsung S27A850D 27" + Samsung 2443BW 24" -!- Sennheiser HD590 -!- Logitech G19 -!- Microsoft Sidewinder Mouse -!- Fragpedal -!- Eaton Ellipse MAX 1500 UPS .





  10. #10
    Xtreme Member
    Join Date
    Nov 2009
    Location
    Gosport, Hampshire UK
    Posts
    220
    Quote Originally Posted by CrimInalA View Post
    Thanks for the tests , now i want to bump my HTT again . I have one problem , when i set cpu NB multiplier to 15 (3075mhz) , i can only go up to 13x (2665mhz) on my HT multiplier ..

    any idea why this could be ?
    Thanks,you need to use HTT/FSB,i ran @250,this gives NB and HT 3010mhz,and drop the ram divider one notch.
    ASUS RAMPAGE V EXTREME, I7 5960X, 16Gig GSKILL DDR4 2666, CUSTOM WATERCOOLING[/SIZE]

  11. #11
    Xtreme Cruncher
    Join Date
    Nov 2002
    Location
    Belgium
    Posts
    605
    you're right , thanks for explaining in words

    but i have to loosen up my ram timings then , since im at 1700+mhz .
    And since RAM timings are critical compared to HTlink speed , i guess i'll just stay where i'm at now


    Main rig 1: Corsair Carbide 400R 4x120mm Papst 4412GL - 1x120mm Noctua NF-12P -!- PC Power&Cooling Silencer MK III 750W Semi-Passive PSU -!- Gigabyte Z97X-UD5H -!- Intel i7 4790K -!- Swiftech H220 pull 2x Papst 4412 F/2GP -!- 4x4gb Crucial Ballistix Tactical 1866Mhz CAS9 1.5V (D9PFJ) -!- 1Tb Samsung 840 EVO SSD -!- AMD RX 480 to come -!- Windows 10 pro x64 -!- Samsung S27A850D 27" + Samsung 2443BW 24" -!- Sennheiser HD590 -!- Logitech G19 -!- Microsoft Sidewinder Mouse -!- Fragpedal -!- Eaton Ellipse MAX 1500 UPS .





Bookmarks

Bookmarks

Posting Permissions

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