Advertisement
Boards Golf Society are looking for new members for 2022...read about the society and their planned outings here!
How to add spoiler tags, edit posts, add images etc. How to - a user's guide to the new version of Boards

[OC log] (\/)0n0L17(-) II

2

Comments



  • CPU Multiplier: 6
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1600
    RAM FSB Mhz: 800
    FSB Memory Ratio: Sync mode

    Timings
    TCL: 4
    TRCD: 4
    TRP: 3
    TRAS: 10
    CMD: 2
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.29
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 9.5
    Memtest86: PASS


    Ok i'm really starting to hate this motherboard :mad: I passed at these settings last night. So I wanted to see if having the CPU divider down at 6 could be loosening the latencies. So I tried these settings

    CPU Multiplier: 8
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1200
    RAM FSB Mhz: 800
    FSB Memory Ratio: 3:2

    Wouldn't BOOT, after a few automatic retries the BIOS went into safemode. So I tried this

    CPU Multiplier: 10
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 800
    RAM FSB Mhz: 800
    FSB Memory Ratio: 1:1

    Wouldn't BOOT again... WHY??? I have no idea. Yet running

    CPU Multiplier: 10
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1000
    RAM FSB Mhz: 800
    FSB Memory Ratio: 5:4

    boots just fine??? I can't get my head around it, how a CPU FSB of 800 is problem, 1000 isn't a problem and 1200 is a problem :confused:




  • CPU Multiplier: 10
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1000
    RAM FSB Mhz: 800
    FSB Memory Ratio: 5:4


    Timings
    TCL: 4
    TRCD: 3
    TRP: 3
    TRAS: 10
    CMD: 2
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.3
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 8.2
    Memtest86: PASS


    Ok, so i'm pretty happy with this. The sticks are running stable at 4-3-3-10 2T, plus the CPU multi doesn't seem to be having an effect, yet.

    Going to try 4-3-3-9 2T next and hope for the best.




  • Ok have a look at these EVEREST screens (who am I talking to? :rolleyes:)

    They show that even at timings 4-3-3-10 i'm seeing a VERY small performance increase over 4-4-4-12. Whats also weird is that 4-4-4-12 @ 419Mhz is actually better. :confused:

    mem_44412_400.jpg

    mem_44412_419.jpg

    mem_43310_400.jpg




  • Frequency is latency for intel, forget eveything you ever learned bout Amd. You'll always get better performance from intel by pushing the fsb, even at looser timings. You've just prooved that.




  • Ok I couldn't get 4-3-3-9 1T stable at 2.2v. So I wanted to see what kind of performance increase just changing the CR would have. I remember on my AMD board the CR gives a bit of a jump in performance at 1T

    CPU Multiplier: 10
    FSB – Memory Clock Mode: unlinked
    CPU FSB Mhz: 1066
    RAM FSB Mhz: 800
    FSB Memory Ratio: Auto


    Timings
    TCL: 4
    TRCD: 4
    TRP: 4
    TRAS: 12
    CMD: 1
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.3
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 8.2
    Memtest86: PASS


    Results:

    Thats a pretty nice jump in performance. I'm going to see if I can lower any of the other timings seperately to get the maximum stable performance.

    mem_44412_1T_400.jpg


  • Advertisement


  • Ok well I tried to boot at

    4-4-3-12 1T, I succeeded, but started to fail memtest86 after only a minute

    I then tried, 4-3-4-12 1T, I couldn't even boot at these timings. I just got the dreaded one long POST beep then the motherboard automatically switched the BIOS into safemode (a pretty nice feature imo, means I don't need to keep manually resetting the CMOS)

    So instead I decided to just try 4-3-3-9 2T, if that works i'm going to try 4-4-4-9 1T next.

    Although i'm starting to lose hope with these sticks, nothing i'm trying is really giving me much, if any, performance increase. My best bet will probably be to just leave it at DDR800 at 1T and start overclocking the CPU.

    Out of interest, those who are reading this, what performance are yous getting out of your sticks and at what timings and clocks?




  • CPU Multiplier: 10
    FSB – Memory Clock Mode: unlinked
    CPU FSB Mhz: 1066
    RAM FSB Mhz: 800
    FSB Memory Ratio: Auto

    Timings
    TCL: 4
    TRCD: 3
    TRP: 3
    TRAS: 9
    CMD: 2
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.3
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 8.0
    Memtest86: PASS


    mem_4339_400.jpg

    hmm... looks like i'm better off with 1T :rolleyes:

    Right, i'm done with this RAM. I wanted to get it above 8000MB/s read but that's not going to happen unless I overclock it from within windows which will force the strap not to change. But i'm not prepared to take that risk.

    Onwards and upwards. I'm going to start testing the CPU now.




  • I don't try to tighten timings, i loosen them & push for higher fsb, more gains that way. But i didn't gain much on these ocz sticks, hopefully new gskills (in my sig) i'll have when my mam returns from Usa will oc a bit better.




  • PogMoThoin wrote: »
    I don't try to tighten timings, i loosen them & push for higher fsb, more gains that way. But i didn't gain much on these ocz sticks, hopefully new gskills (in my sig) i'll have when my mam returns from Usa will oc a bit better.

    I think I agree with you. I tried testing 4-4-4-9 CR1 last night and i'm getting a worse score than at 4-4-4-12 CR1. I asked about this on the Mushkin forums and they said this is normal as tightening the timings also has the affect of loosening the NB latencies (which can't be controlled :()

    It seems a bit misleading for RAM suppliers to charge more for tighter timings when they give near to no increase in performance.

    My other problem is that I can't increase the FSB either, as, due to strap changes, it also has the affect of reducing performance rather than increasing it.

    I'll give overclocking the CPU a shot now and hope for the best. My only hope is that because my CPU has an unlocked multiplier I won't have to increase the FSB too much.




  • CPU Multiplier: 12
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1000
    RAM FSB Mhz: 800
    FSB Memory Ratio: 5:4

    Timings
    TCL: 5
    TRCD: 5
    TRP: 5
    TRAS: 12
    CMD: 2
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.3
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 9.3
    OCCT: PASS


    Ok pushed the clock up to 3Ghz straight away... :rolleyes: luckily enough, its stable at 3Ghz with a multi of 12. I'll be stepping up now 100Mhz at a time.


  • Advertisement


  • Nice you can change your multiplier.I dont seem to have that option.
    See youve loosened your timings also.
    Thanks for the link on the other thread.




  • You can lower the multiplier, but not increase it unless you have an extreme edition chip like L31mrod




  • Nice you can change your multiplier.I dont seem to have that option.
    See youve loosened your timings also.
    Thanks for the link on the other thread.

    Yeah, any Intel CPU with an "X" in the name usually means it has an unlocked multiplier. (i.e. yours is a Q6600 so it doesn't, mines a QX6700) Like Pog said, it doesn't mean it can't be lowered though.

    Here are some bench screens, from that link I gave you with bench programs I found this german PI program for benching all 4 cores called MaxPI Multicore

    @ 2667Mhz

    cpu_bench_2666.jpg

    @ 3100Mhz

    cpu_bench_3100.jpg




  • BTW, Prime 95 is now multi core.

    Push up the fsb now m8 :p




  • PogMoThoin wrote: »
    BTW, Prime 95 is now multi core.

    I can't find any information about prime 95's support for quads. What version supports it?




  • CPU Multiplier: 12
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1066
    RAM FSB Mhz: ~<800
    FSB Memory Ratio: 4:3

    Timings
    TCL: 5
    TRCD: 5
    TRP: 5
    TRAS: 15
    CMD: 2
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.3
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 6
    OCCT: FAIL


    hmm... ok, woke up to a message in Vista saying the system had recovered from a system error (BSOD) From the event logs it looks like OCCT lasted 5 hours at 3.2Ghz before restarting. Going to try 1.35v now and test again.

    BTW, my load temps in EVEREST are around 55'c per core under load with the side panel off, i'll be happy pushing it up to 70'c if it can handle it.




  • http://extremeoverclock.org/category/download/tweak-overclock/stressing/

    Its version 25.4 thats multicore, not sure if its out of beta stage but it works fine. I used occt for a quick 30min test while pushing for max fsb & prime 95 for an overnight torture test.




  • PogMoThoin wrote: »
    http://extremeoverclock.org/category/download/tweak-overclock/stressing/

    Its version 25.4 thats multicore, not sure if its out of beta stage but it works fine. I used occt for a quick 30min test while pushing for max fsb & prime 95 for an overnight torture test.

    cheers... i'll give it a shot when I get home (working xmas eve :()

    Something i've noticed with my temps, since putting the 8800GTX into my P180 i've found my CPU temps have gone up from 28'c idle to 40'c after having the machine on for a few hours. I put this down to the fact that the 8800GTX is so long it basically partitions the main compartment in 2, stopping air from the middle intake fan getting to the RAM and CPU. Theres also the added heat being spewed out by the 8800GTX. I think this is what killed my last set of RAM. I had had the PC on for a few days so I guess the heat just built up on the RAM and eventually killed them. I have no way of getting an air intake around the RAM so i'm just going to leave the side off, the noise levels are also a lot lower as I don't have to have all my fans running at high RPMs to get the same level of cooling. I have a can-o-air to deal with the inevitable dust that will be caused by this.




  • would you not up the voltage on the CPU?
    How high can you go anyways?




  • would you not up the voltage on the CPU?
    How high can you go anyways?

    I did, I said i was going to increase the volts to 1.35v, which worked out good :D i'm now stable at 3.2Ghz

    CPU Multiplier: 12
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1066
    RAM FSB Mhz: ~<800
    FSB Memory Ratio: 4:3

    Timings
    TCL: 5
    TRCD: 5
    TRP: 5
    TRAS: 15
    CMD: 2
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.35
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 8
    OCCT: PASS


    EDIT:

    cpu_bench_3200.jpg


  • Advertisement


  • Stable at 3.3Ghz, loading around 62'c on each core now.

    CPU Multiplier: 12
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1100
    RAM FSB Mhz: 734
    FSB Memory Ratio: 4:3

    Timings
    TCL: 5
    TRCD: 5
    TRP: 5
    TRAS: 15
    CMD: 2
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.35
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 8
    OCCT: PASS


    EDIT:

    cpu_bench_3300.jpg




  • Your safe till 70'c Imo. Nice overclock, but 3.6 should be easily achievable




  • damn! failed at 1.4v. Ok going to try 1.45v now and test again

    CPU Multiplier: 12
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1132
    RAM FSB Mhz: ~750
    FSB Memory Ratio: 4:3

    Timings
    TCL: 5
    TRCD: 5
    TRP: 5
    TRAS: 15
    CMD: 2
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.4
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 8
    OCCT: FAIL




  • i'm stable now at 3.4Ghz, but it's coming at a price... my load temps are hitting 80'c as my volts are up to 1.45v now. Don't know if i'd be happy pushing them much further, so I think i'm going to drop back down to 3.3Ghz @ 1.4v for everyday use. I'm going to investigate also if there are any other settings I can change, besides CPU volts, that can stabilize 3.4Ghz.

    CPU Multiplier: 12
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1132
    RAM FSB Mhz: ~750
    FSB Memory Ratio: 4:3

    Timings
    TCL: 5
    TRCD: 5
    TRP: 5
    TRAS: 15
    CMD: 2
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.45
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 8
    OCCT: PASS


    cpu_bench_3400.jpg




  • ever condsider going WC ?




  • Just goes to show ya why extreme edition intel chips are not worth it (unless they come at a knockdown price), they're only for extreme enthusiasts to put under phase or ln2. I got my q6600 to 3.6 with similar voltage & temps and i can run at 3.3Ghz 24/7 with the same cooler, only difference is i lapped mine.

    Hopefully you'll tweak your voltages a bit & get 3.4 at lower temps. Good luck m8




  • PogMoThoin wrote: »
    Just goes to show ya why extreme edition intel chips are not worth it (unless they come at a knockdown price), they're only for extreme enthusiasts to put under phase or ln2. I got my q6600 to 3.6 with similar voltage & temps and i can run at 3.3Ghz 24/7 with the same cooler, only difference is i lapped mine.

    Hopefully you'll tweak your voltages a bit & get 3.4 at lower temps. Good luck m8

    got mine fairly cheap though, it was around the time a lot of people where shedding their QX6700's fearing the cheaper Q6600's. Can't believe some people paid near a grand for these things when they first came out. The main reason I got this chip so cheap was that the seller knew it had a heat problem. Still I like a challenge :D I've also lapped my quad and Tuniq to a mirror finish and installed a higher CFM fan. The problem is the CPU is generating heat faster than the metal (or the liquid in the copper pipes) can conduct it. Aside from using a TEC, water or ln2 I doubt i'm going to get much better. But i'll still try over the next week or so to see if increasing the other voltages, or lowering the HT multiplier will have any affect. I might also lower the multiplier and increase the FSB and see if that has an affect.

    here's hoping




  • Ok, well after hitting a thermal wall at 3.4Ghz, i've decided to, for now, get a working everyday overclock, which means seeing if my RAM can run at its rated speeds with a CPU overclock, and if possible have a CR of 1T (should be very possible) At this level i've already tested what is the stable voltage for both my CPU and RAM so if I run into any issues now, another variable will be at fault, more than likely the NB temperatures or voltages.

    I had to increase the multiplier to 13 as I tried 300*11 (3.3Ghz) but this gave me a QDR of 1200 which, for some odd reason, my system can't boot at. This would of been a perfect option as it gives me the stable 3.3Ghz and a DDR of 800 for the RAM, but c'est la vie :rolleyes: At a multi of 13, my clock is 3.25Ghz with a DDR of 800. I'm going to stress test this setting, then lower the timings to rated, then try 1T. Once I get a stable everyday overclock i'm to see about pushing past 3.4Ghz. I've noticed that increasing the multi to 13 has risen my idle/load temps higher than what they where at 12 @ 3.3Ghz, so maybe the high multiplier is the cause of the horrible temperatures.

    I'm going to drop it down to 9 and start upping the FSB instead and see if there is any change in temps at 3.4Ghz.

    CPU Multiplier: 13
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1000
    RAM FSB Mhz: 800
    FSB Memory Ratio: 5:4

    Timings
    TCL: 5
    TRCD: 5
    TRP: 5
    TRAS: 15
    CMD: 2
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.4
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 9
    OCCT: PASS




  • Ok so this will probably be my stable overclock for now. I have the timings back to 4-4-4-12 1T @ 812Mhz and my overclock at 3302Mhz. Weird thing is my RAM performance has dropped off considerable to what it was previously when I had the clock mode set to unlinked at these exact timings and this exact DDR clock. Plus the FSB is currently LOWER than stock so it shouldn't be affecting latencies negatively. I don't think i'm ever going to understand this mobo, its like shooting in the dark.

    3.3Ghz

    CPU Multiplier: 13
    FSB – Memory Clock Mode: linked
    CPU FSB Mhz: 1016
    RAM FSB Mhz: 812
    FSB Memory Ratio: 5:4

    Timings
    TCL: 4
    TRCD: 4
    TRP: 4
    TRAS: 12
    CMD: 1
    TRRD: 3
    TRC: 24
    TWR: 5
    TWTR: 14
    TREF: Auto

    Volts
    CPU Core: 1.4
    CPU FSB: 1.2
    Memory: 2.2
    Nforce SPP: 1.25
    Nforce MCP: 1.5
    HT nForce SPP <-> MCP: 1.2

    Time Ran (hrs): 9.30
    OCCT: PASS


  • Advertisement


  • any recomendations on benchmarking software?
    not just for the GPU but more for the system itself.

    I can see the improvements on my overclock but would like to get the stats also.


Advertisement