1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Can RAM speed make a difference in gaming?

Discussion in 'Motherboards & Memory' started by hat, Mar 17, 2008.

  1. hat

    hat Enthusiast

    Joined:
    Nov 20, 2006
    Messages:
    17,176 (5.53/day)
    Thanks Received:
    2,191
    Location:
    Ohio
    Can RAM speed make a difference in frames per second when gaming?
     
    Crunching for Team TPU More than 25k PPD
  2. WarEagleAU

    WarEagleAU Bird of Prey

    Joined:
    Jul 9, 2006
    Messages:
    10,812 (3.34/day)
    Thanks Received:
    547
    Location:
    Gurley, AL
    I believe it can yes. Not a huge difference, but enough if its being accessed and what not.
     
  3. Pinchy New Member

    Joined:
    Apr 29, 2006
    Messages:
    5,109 (1.54/day)
    Thanks Received:
    284
    Location:
    Sydney, Australia
    It can, but it depends on a lot of things.

    Like from 800Mhz to 1000Mhz, your not going to see much of a difference, if anything. From something like 266Mhz to 1000Mhz, you will obviously see a bigger difference but again, it wont be huge.
     
  4. hat

    hat Enthusiast

    Joined:
    Nov 20, 2006
    Messages:
    17,176 (5.53/day)
    Thanks Received:
    2,191
    Location:
    Ohio
    so 550 to 600 won't be a noticible differnce then?
     
    Crunching for Team TPU More than 25k PPD
  5. EastCoasthandle

    EastCoasthandle New Member

    Joined:
    Apr 21, 2005
    Messages:
    6,889 (1.87/day)
    Thanks Received:
    1,506
    I was asking myself the same question. For me, it's not just going from 800MHz @ 4-4-4-12 to 1066MHz @ 5-5-5-15. But it's going to 1066MHz @ 5-5-4-12 that seems to make some difference with min frame rates. Base on my system setup it's still subjective and, it also depends on the game you play.
     
  6. erocker

    erocker Super Moderator Staff Member

    Joined:
    Jul 19, 2006
    Messages:
    40,460 (12.53/day)
    Thanks Received:
    15,145
    I think that a FarCry or Crysis benchmark will let you know for sure.
     
  7. jbunch07

    jbunch07 New Member

    Joined:
    Feb 22, 2008
    Messages:
    5,261 (1.99/day)
    Thanks Received:
    614
    Location:
    Chattanooga,TN
    i had my ram at 800 @ 4-3-3-10 and then oced to 1000 at 5-5-5-15, didnt see much difference in gaming or benchmarking if any so prob going to go back to 800
     
  8. phanbuey

    phanbuey

    Joined:
    Nov 13, 2007
    Messages:
    5,225 (1.90/day)
    Thanks Received:
    991
    Location:
    Miami
    IMO every system/motherboard has a different sweet spot... Like my gigabyte loves its 2.4:1 (yes 2.4... weird i know), and refuses to run 890 @ 5-5-5-15 on a 2:1 divider. But it runs at 977 @ 4-5-5-15 all day @ 'extreme' access mose (tighter MCH timings?)...

    Also it depends on FSB speed... at an fsb of 333 youre very FSB limited, which means even if your ram is running at 1000 5-5-5-15, your still gonna get about the same (maybe a 100-200 MB/s increase) transfer rate as 800 4-4-4-12.

    Now bump the FSB up to 400 and youre gonna see a huge difference... the 1000 kit will give a substantial increase in FPS. I dont know ab AMD or how the HT link works, you guys already have insane memory performance.

    EDIT: to answer your question, yes ram can make a 'big' difference, but it depends on the system setup. Sorry to give such a lame-o answer, but thats the most accurate... check out some ram reviews, alot of them usually compare FPS in games, and some systems will get up to a 20% boost, while others gain 3-5%.
     
    Last edited: Mar 17, 2008
  9. lemonadesoda

    lemonadesoda

    Joined:
    Aug 30, 2006
    Messages:
    6,315 (1.98/day)
    Thanks Received:
    976
    Increasing the memory clocks makes a noticeable difference, e.g. 600 to 800 to 1066. Depending on the application, it can gain 10-20%.

    Increasing the timing of the memory makes very little difference, e.g. 5-5-5-12 to 4-4-4-12.

    The reason we LIKE 4-4-4-12 is that its got more overclock headroom. But at the SAME clocks, it makes very very little difference. <0.5%
     
  10. hat

    hat Enthusiast

    Joined:
    Nov 20, 2006
    Messages:
    17,176 (5.53/day)
    Thanks Received:
    2,191
    Location:
    Ohio
    So would it be worth it picking up a 2x1GB DDR2 800 4-4-4-12 kit? I was thinking of doing that for my birthday if everything pans out the way I expect (hope) it to along with a better video card. Or should I just continue running these at ddr600?
     
    Crunching for Team TPU More than 25k PPD
  11. phanbuey

    phanbuey

    Joined:
    Nov 13, 2007
    Messages:
    5,225 (1.90/day)
    Thanks Received:
    991
    Location:
    Miami
    since you would be paying like 40 - 50 bucks, and sine that kit could *probably* make it to 1000 or at least 950, then i would say so... but then again, it really depends what youre putting it into.
     
  12. Pinchy New Member

    Joined:
    Apr 29, 2006
    Messages:
    5,109 (1.54/day)
    Thanks Received:
    284
    Location:
    Sydney, Australia
    Well with a real world example: I OC'd my 800 to 1050Mhz and got no performance hit in anything except the everest read benchmark, lol.

    That was with my FSB @ around 380-400 or something BTW.
     
  13. exodusprime1337

    exodusprime1337

    Joined:
    Nov 27, 2007
    Messages:
    2,188 (0.80/day)
    Thanks Received:
    343
    what about say ddr2 800 to ddr2 1180. currently testing out my clocks at 3.55Ghz on my 5000 so i'm runing 295x12 and my memory being 1-1 is giving me 1180 exact, i've noticed a pretty sizeable increase in some applications. also my brute memory thoroughput has gone from the 8k range all the way to low 11k, that's gotta have an effect on something?? right?? i hope or buying new memory was a waste lol
     
  14. jbunch07

    jbunch07 New Member

    Joined:
    Feb 22, 2008
    Messages:
    5,261 (1.99/day)
    Thanks Received:
    614
    Location:
    Chattanooga,TN
    thats a huge oc what voltage is your ram set at and what kind is it?
    also you should try some benching before and after your oc to see the actual difference :toast:

    Edit: never mind i read your specs, always forget about that!
     
  15. phanbuey

    phanbuey

    Joined:
    Nov 13, 2007
    Messages:
    5,225 (1.90/day)
    Thanks Received:
    991
    Location:
    Miami
    something other than raw memory performance was going on... every MB has its 'sweet spot'... i get better performance @ 977 @ 5-5-5-15 than at 1001 5-5-5-15 because my MB automatically adjusts specific MCH timings, which are not available in the bios, for stability.
     
  16. Pinchy New Member

    Joined:
    Apr 29, 2006
    Messages:
    5,109 (1.54/day)
    Thanks Received:
    284
    Location:
    Sydney, Australia
    No but what im saying is...with every Mhz increase my memory read benchmark always went up, so the performance was still going up, but it still didnt do anything for me in applications and games.
     
    phanbuey says thanks.
  17. jbunch07

    jbunch07 New Member

    Joined:
    Feb 22, 2008
    Messages:
    5,261 (1.99/day)
    Thanks Received:
    614
    Location:
    Chattanooga,TN
    i wish i could manually control the cpu/ divider i cant seem to find it in my bios, but if i set all my memory setting on auto my divider changes to cpu/7 but when i go and set my own timings it changes to cpu/8 ? anyone know why this is?
    my guess is its just not a great bios, seeing as how i have to press ctrl+f1 to get to the advanced chipset features...
     
  18. Darknova

    Darknova

    Joined:
    Nov 8, 2006
    Messages:
    5,037 (1.62/day)
    Thanks Received:
    535
    Location:
    Manchester, United Kingdom
    Not a major difference no. but I do see a difference going from 800 4-4-4-12 to 1000 5-5-5-15.

    You're biggest difference will be from decreasing tRD, but this only really applies to Intel systems.
     
  19. jbunch07

    jbunch07 New Member

    Joined:
    Feb 22, 2008
    Messages:
    5,261 (1.99/day)
    Thanks Received:
    614
    Location:
    Chattanooga,TN
    ok so what would be better for an amd cpu, better clocks or better timings? for ex 800@4-4-4-12 or 1000@5-5-5-15? dont really see any diff between the 2, if there is one its prob not noticeable?
     
  20. AsRock

    AsRock TPU addict

    Joined:
    Jun 23, 2007
    Messages:
    11,740 (4.06/day)
    Thanks Received:
    2,046
    Location:
    US
    Amount yes speed of it never noticed any difference in games.. Thats why i am getting rid of my 4GB of Transend and going buy 8GIG of 1066. And yes 8GIG is some what over kill but i hit 3gig as it is all ready with some things i do.
     
  21. ntdouglas

    ntdouglas New Member

    Joined:
    Jan 19, 2007
    Messages:
    1,026 (0.34/day)
    Thanks Received:
    44
    Location:
    Chicago
    Games will load a little faster also.
     
  22. Darknova

    Darknova

    Joined:
    Nov 8, 2006
    Messages:
    5,037 (1.62/day)
    Thanks Received:
    535
    Location:
    Manchester, United Kingdom
    Not as much as a brand new fast harddrive.
     
  23. phanbuey

    phanbuey

    Joined:
    Nov 13, 2007
    Messages:
    5,225 (1.90/day)
    Thanks Received:
    991
    Location:
    Miami
    Ahh i see what youre saying. I read it totally wrong the first time...:slap.

    Now that i think about it, the maximum benefit (assuming the same timings for both) from OCing to 1000MHz from 800MHz can only be 25%... thats only if you have an infinitely fast CPU, at an infinitely fast FSB with a graphics card to match... so the ram would HAVE to be a 100% bottleneck to see a 25% boost.

    This can never happen, so i totally see how your real world example is right-on since games tend to be bottlenecked by the GPU and CPU. Most systems (including mine) would see higher benchmark scores but not else much if going from 800 4-4-4-12 to 1000 5-5-5-15...
     
  24. JrRacinFan

    JrRacinFan Served 5k and counting ...

    Joined:
    Mar 17, 2007
    Messages:
    19,418 (6.50/day)
    Thanks Received:
    4,495
    Location:
    Youngstown, OH
    If it helps any hat, my rig gets unstable if I try to use a divider at anything above a 333fsb to raise my ram higher than 667mhz. So far my highest 100% stable fsb setting is 350mhz fsb with 700mhz ram. I run mine on a 24/7 266fsb with a 2:3 divider. That gives me 2.7Ghz cpu and 800Mhz ram.


    EDIT: Also, my gaming benches only rise in FPS if I overclock the video card and run @ 1440x900 res.
     
  25. ntdouglas

    ntdouglas New Member

    Joined:
    Jan 19, 2007
    Messages:
    1,026 (0.34/day)
    Thanks Received:
    44
    Location:
    Chicago

    Or a raid set up like mine.:D
     

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page