76 Comments Add your own

• 1. JohnMu  |  January 20th, 2008 at 6:47 am

  Stock cooler? That’s one of the few things I always buy with a new CPU. Those Intel leaf-blowers are just too loud for me :). Have you checked to make sure that you’re actually using both cores properly? VB.NET is sometimes a bit single-core-focused, unless you work on splitting the workload into multiple threads. On the other hand, if you want to use the system for something while it’s processing, doing the work on a single core might be preferable

  PS Good to see you posting again!

• 2. theDuck  |  January 20th, 2008 at 9:18 am

  Yeah by stock cooler I mean the one that comes with the processor Mu - which is infinitely tiny on the new 45nm cores - about half the size of the latest AMD coolers - but, amazingly, it seems, quite effective. As for the noise - well, suprisingly, again, quite quiet

  I would have thought that smaller pipes (45 nm) would have led to higher temperatures, but it seems intel has also managed to drop the voltage fairly significantly also, so these are not altogether ‘hot’ processors, even when overclocked this much.

  Re whether or not I’m making full use of the cores, no I’m not - the particular app that I’m working on lends itself to sequential processing on one core rather than parallel processing on two, although I’ve been looking into the VB.NET pre-emptive multitasking capabilities (the equivalent of your FORKs etc in C).. hmm.. we’ll see.

  BUT - for testing purposes on this overclock I’ve used CPUBurn and Orthos to load both CPU’s to the max. Yes I have found some (recoverable) instability at this high overclock, and that may be able to be avoided with a better cooler.

  Certainly increasing the VCore doesn’t seem to help much - but reducing the clock to around 3.7Ghz does the trick even for Orthos - although I haven’t been able to find any ‘real world’ application that loads the CPU in quite the same way as those testing apps do, so for all intensive purposes the 3.92 O/C is stable.

  Re: Posting more - I’m moving house to a place that only has dial-up (yikes! and only 20km out of the city - welcome to Brisbane) over the next few weeks so I’m getting my internetting out of the way now

• 3. Joseph Glover  |  February 9th, 2008 at 12:59 pm

  I’m trying to O’C my E8200 and have been following your guide but i appear to be missing somthing….

  I’m running E8200 CPU 4gb Crucial Balistix 8500C5 RAM on a Gigabyte Ga-p35-DQ6.

  so i thought your guide would be ideal! very similar specs.

  if i try and overclock my CPU even to 3ghz to cant boot into windows… but i know its me….

  my ram timings as stock 5.5.5.15 2.2v
  my cpu is the same as yours.. right down to the stepping.
  our BIOS’s cant be that different. so what am i doing wrong?

  the one thing i cant find on my BIOS is RAM:FSB Ratio???? does to go by a different description in the BIOS?

  i have great ZALMAN copper cooling.. i really think with a little know how it could push 4ghz.

  Any help would be great.

  thanks

  Joe.

• 4. theDuck  |  February 10th, 2008 at 11:58 am

  Hi Jo - the RAM:FSB ratio is your problem.

  Not being able to post BIOS is a really common symptom of pushing your RAM way too fast.

  You need to pull the clock rate of the RAM back a little or you’ll be driving it heaps quicker than it can take (by default those boards overclock the RAM relative to the FSB.)

  I’d check out your moby manual - I’d suspect it will be there, maybe just called something different.

  You basically want your ram running at twice the speed of the base clock - so with a base clock of 450MHz (450*8=3.6GhZ actual CPU clock) you want to see your ram running at a multiplier of 2 - ie 900MHz

  M

• 5. Joseph Glover  |  February 10th, 2008 at 10:08 pm

  Great! thanks for that.

  it was called the system Memory Multiplier, and your right, it was set to auto and so pushing my speed passed 2ghz! Hope it hasn’t done any damage.

  I’m doing my first O’C now. I’ll keep you dated.

  cheers

  J.

• 6. Joseph Glover  |  February 10th, 2008 at 10:58 pm

  First O’C and i think it has alot more head room!
  450×8 = 3.60ghz
  RAM Multiplier at 2.00 = 900mhz (can go to 1066 as the ram is DDR8500C5)
  V core @ 1.28 (will try and reduce that)

  Thanks for your help! 1ghz O’C not too shabby for the first crack.

  J.

• 7. theDuck  |  February 11th, 2008 at 11:28 am

  Good show!

  I’d suggest ~3.7GHz @1.27 v is a realistic target for a solid, stable system without going overboard with water coolers etc.

  Cheers,

  M

• 8. Nick  |  February 22nd, 2008 at 1:40 pm

  Hey, thanks for the article. I’m wondering why you opted for the DS3R over the DS3?

  Cheers
  Nick

• 9. Greg K.  |  March 5th, 2008 at 9:32 am

  DS3R has built-in RAID capabilities, if I recall correctly.

  I enjoyed your OC review, and I’d like to contribute my $0.02-worth from my own experience. I currently have a somewhat inferior E4600 on the P35-DS3, but I run it easily at 3.2Ghz with PC2-6400 memory (800Mhz) from Crucial running at 1000Mhz 4-4-4-12 timings. System is 95% stable with some very rare very obscure crashes, that happen so rarely that they do not bother me. For all I know it could be Vista x64 related (I have 4Gb of RAM). The crashes do seem to go away at stock settings, but that just may mean that they’d still happen, only will take more extreme conditions.

  I process HD video and play some DX9 games on my PC, as well as use Photoshop and Premiere filters A LOT. so I’d say I put moderate to heavy use on the CPU. Even more so on the memory.

  Comments:

  #1 good cooling does not necessarily equate to “heaps of money on expensive cooling gear”. A decent Thermalright or Zalman or (my choice) Scythe will do the job very nicely at sub $50. To spend upward of a $1000 on great mobo, memory, and $250 CPU, and to skimp $50 on the lifeline of your whole system - the cooling is plain bad PC design in my mind. My 65nm E4600 never exceeds 53C at highest of loads. Considering that I like to stay warm and my ambient temps run at 70-75F, that’s NOT a bad number. All included, I have 11 fans in my system: 1 120mm Power Supply fan, 4 120mm case fans (strategically placed to blow in on HDDs, in on the video card, out the back,and out of the top), 2 120mm fans on Scythe Infinity, 3 micro fans on the Corsair memory cooler, 1 stock fan on the overclocked ATI 3850. I wouldn’t say that my system is the HTPC dream machine, farfrom it on may counts, BUT it’s pretty darn quiet for every day work. Quietest PC I have in my house, including 2 laptops. Total cost of cooling in my machine (without the PSU and video fans): $90. Unlike liquid systems, these fans will outlive every other component in my machine except for the tower case.

  #2 Of all brand name memories, I find Kingston has the worst OC. go for PQI or Crucial if you can. Most of the time PQI is cheaper than big-K anyway. If you want to go crazy, get some Mushkin memory. It’s over-the-top great!

  #3 Lower multi+higher FSB = faster overall system speed at the same Ghz. So 500×8 would beat 445×9 any day. You may have to add some juice to your FSB though, which Gigabyte mobos will take without a degree of temp increase. Your review leaves me to wonder that if my target speed is stable 4Ghz, should I spend money on the 8400 ando do 500×8 or should I just spend half the price difference on better cooling and end up with a longer lifespan for my system and better core temps at the same 500×8 on 8200? Any more than 4Ghz, and I’m due for a very expensive memory upgrade. Have to consider if it’s really worth spending extra $60 on CPU and extra $150 or so to get 1066 memory with CAS 5 or even 6 timings just to push it another 10% to 4.4Ghz or so.

  #4 To Joseph Glover: memory multiplier is not all that can break things. Make sure your PCI-E is not set on Auto. you can render your system dead very quickly at these speeds. PCI-E overclocks very poorly. I set my system to 110 fixed with my eyes closed, and I don’t ever try to push it. Why risk losing a $200 video card and other perepherals? 10% OC on the bus lets me OC my video card a little more. I get sick frame rates in HL2 engine games @ 1920×1200 (I love Team Fortress 2).

  #5 This advice isn’t suitable for most people, but it certainly is for you - opt for a Quad core, and optimize your routines to use them all. You’ll get more performance at 2.6Ghz than you can out of 8200 at 4Ghz. There’s a new AMD 9900 on the market that seems to beat at stock speed (200FSBx13=2.6Ghz) an OC’ed E8400 @ 4.4Ghz. Makes you wonder. This only applies to computational benchmarks and only when software can recognize all 4 cores.

  Hope this helps you guys get more out of your PCs for the same money. I love high speed at a reasonable budget!

• 10. theDuck  |  March 5th, 2008 at 9:50 am

  NICK - DS3R does have raid, but it also has six memory sockets - two of which are for DDR3 - the next gen memory. I just bought it because it makes the board a bit more future proof.

  GREG - Aha! Now we’re talking - some meaty content from someone who’s been there, done that - Thanks Greg.

  And - by the way Greg, thanks so much for pointing out how essential it is to take the PCI off Auto - I’ve blown a motherboard soundcard before by forgetting to turn that off

  Re E8200 - In your circumstance I’d be going the e8400 as, like you say, if you’re reusing your memory sticks you’re going to peak out pretty quick at ~ 3.9GHz with the E8200. The E8400 should see you well in excess of 4.

  Cheers,

  M

• 11. Joseph Glover  |  March 5th, 2008 at 10:19 am

  Thanks Greg… well… HA! i’ve been running my 8200 at 3.66ghz since my initial post with PCIX on AUTO!!! bollox!!! but right now (new system build) I’m running at stock.. 2.66ghz

  If i turn auto off? whats the norm? 110 is an OC right?

  My RAM is crucial Ballistix 8500 CAS5 (5.5.5.15) yet my MOBO reads it as 800mhz so im thinking i have alot of head room.. Hosw much joice should i give the FSB if i’m wanting to hit a stable 4ghz?

  a quick breakdown would be great.

  so for now.. i’m back at the stock..

  thanks for you help guys.

  J.

• 12. theDuck  |  March 5th, 2008 at 11:04 am

  Joseph - that’s a ~50% overclock - for future reference the default is 100, and you’ll prob find things will be stable at 110 - so set to that.

  Re how much joice - I’d do this -

  1. Set your PCI to 110.
  2. Set Ram Multiplier back to ~ 2 instead of auto.
  3. Start the CPU clock at 470×8.

  From there, you just need to up the clock gradually, checking for stability each time. You may need to up the CPU / FSb Voltage too (in fact to get to 4GHz, that’s a dead cert) - possibly you’ll need to upvolt your RAM by +0.2V or so also.

  Just a matter of trial and error.

  M

• 13. cosmico89  |  March 6th, 2008 at 6:30 pm

  hi to all … I have this processor e8200 and created a preview at least I hope hihi .. I have arrived safely to 4ghz only then that the ram is not supported more then went over .. not for the driving I do not want to do but spam is on http://www.deuxexmachina.it
  thx

• 14. doddie  |  March 7th, 2008 at 11:34 am

  Hi Folks,

  1st time OverClocking, bear with me
  I have a E8200 wanting to OC to 3ghz for now.

  I have figured out most of the settings and applied them in bios
  and its reading 3ghz ok. not saved them tho, im not sure about
  the PCI. I know to change from auto, but do i need to set @ 110
  if im only going to 3ghz? would 103 be enough? dont want to fry my board, would it be unstable if its set too low?

  Gigabyte P35 DS3 Rev2, 2 GBCrucial Ball tracer (8500)

  375×8 = 3ghz
  RAM Multiplier at 2.00 = 750 mhz (can go to 1066 as the ram is DDR8500)
  Timing at 5.5.5.15 2.20v

  Do I need to change any other voltages, vcore etc..for 3Ghz?
  Current stock readings in CPUID Monitor now….

  Voltage sensor 0 1.20 Volts [0×4B] (CPU VCORE)
  Voltage sensor 1 2.18 Volts [0×88] (DDR)
  Voltage sensor 2 3.31 Volts [0xCF] (+3.3V)
  Voltage sensor 3 4.89 Volts [0xB6] (+5V)
  Voltage sensor 7 11.84 Volts [0xB9] (+12V)
  Voltage sensor 8 3.18 Volts [0xC7] (VBAT)
  Temperature sensor 0 30°C (85°F) [0×1E] (TMPIN0)
  Temperature sensor 1 22°C (71°F) [0×16] (TMPIN1)
  Fan sensor 0 1744 RPM [0×183] (FANIN0)
  Fan sensor 2 1786 RPM [0×17A] (FANIN2)

  Temperature sensor 0 68°C (154°F) [0×25] (core #0)
  Temperature sensor 1 68°C (154°F) [0×25] (core #1)

  Thanks for any help

• 15. theDuck  |  March 7th, 2008 at 12:49 pm

  Hi Doddie -

  I think you should consider the PCI as seperate to the core clock - there is no rule that says you have to overclock the PCI as well as the CPU, although some folks do (especially gamers). You could even just leave it at 100.

  Certainly, as Greg points out, peripherals on the PCI bus can be a bit prone to failure if you overclock it too much. My suggestion would be to avoid going over 110, and if you want to be safe, 103 should be absolutely fine.

  I’d be upping your ram multiplier - try 2.5 - this will take you to ~ 940MHz Ram speed, which the PC8500 RAM should handle just fine at your clock speed of 3GhZ - I’ve gone a bit more extreme (3.9 GhZ) and was having problems with the RAM wigging out (which, by the way, usually causes the Gigabyte Dual Bios boards to take a very long time to POST, and resets all overclocking values to default) - that’s why I dropped the multiplier back, but you shouldn’t need to at your desired speed of 3GHz.

  One thing I’m a little concerned about are your core temperatures - they seem a little high for such a moderate overclock - the maximum recommended temperature for the e8200 is 72.4 degrees celcius and you’re temps are a bit close for comfort.

  I often find that using a third party heatsink paste (like arctic silver or something similar) can reduce your temps quite significantly, or you may want to purchase a different cooler (although I think that should be unnecessary)

  If you decide to use a different heatsink compound, make sure that you clean the existing heatsink compound from your processor and the base of your heatsink first - use isopropyl alcohol (called isocol here in Australia) and a clean lint free cloth - remember to let them dry thoroughly prior to adding the new compound.

  You only need a tiny bit of compound - about the size of a grain of rice - check out the arctic silver website for more instructions.

  Cheers,

  M

• 16. theDuck  |  March 7th, 2008 at 1:08 pm

  By the way folks - here in Australia the supply of the new E8XX (e8200, e8400 and e8500) CPU’s seems to have dried up. I’ve checked 6 suppliers, and all bar one (umart) are totally sold out and not expecting new stock until at least the end of March.

  In Umart’s case, they’ve upped the price of their E8200 CPU’s from around $250 to $300 - I guess it’s a case of supply and demand.

  Seems these are popular little chips - I’m currently doing a build for a client, and have ordered a Quad Core (Q6600) on the old 65nm core instead - not sure how well it will overclock but will keep you posted.

  M

• 17. doddie  |  March 7th, 2008 at 1:32 pm

  Thanks M for the reply

  I will leave the PCI at 100 see how it goes, I do play games,
  using a factory overclocked asus 8800GT TOP.
  Would this make much difference?

  I’l also up the mem to 2.5 as suggested.

  Regarding the temp I thought it was high myself, until I googled
  http://www.google.co.uk/search?hl=en&q=e8400+temperature+problems&meta=
  Whats your thoughts on this?
  Im presuming I have a faulty sensor.

  im using stock cooler. my normal CPU temp seems to read ok Temperature sensor 1 22°C (71°F) [0×16] (TMPIN1) idle
  this reading will rise to 39c benchmarking. the 2 core readings never change @ idle or under load, so im near certain I have the
  faulty sensor.

  I will play about with the settings and see how it goes.

  Phil

• 18. theDuck  |  March 7th, 2008 at 1:41 pm

  I don’t think the PCI is going to make much of a difference.

  Those temperatures ~ 30 degrees look like the case temp, not the core temp - I think you’re reading the wrong temps. You should be reading these ones:-

  Temperature sensor 0 68°C (154°F) [0×25] (core #0)
  Temperature sensor 1 68°C (154°F) [0×25] (core #1)

  The temp sensors for the cores are on the die itself - one faulty, possible, two faulty - highly unlikely - but it’s possible that your bios may not support the e8xxx cpu’s properly, so definitely upgrade to the latest BIOS.

  The temps seem on the highish side - use a cpu testing software such as cpuburn or something similar and see what they’re reading under full load (using cpu temp, a free utility).

  If the temps are as reported, it’s unlikely you’ll damage the processor with high temps, but your system may become unstable.

  M

• 19. doddie  |  March 7th, 2008 at 11:44 pm

  I cleaned off old paste and applied artic silver, made sure fan was seated properly.

  I done a benchmark with everest, as you can see the results are the same, with core 1 + 2 always sitting at 68c and cpu temp at 36c case/mb at 30c

  http://img512.imageshack.us/my.php?image=formfu6.png

  the bios is the latest F12 that came with the MB dated jan 08

  btw: my pc is in my back store in house (UK), below average temps

  any ideas thx

• 20. theDuck  |  March 8th, 2008 at 12:00 am

  Got me stumped Doddie! You’re prob right - faulty sensor or a motherboard incompatibility. Did you try using the cpu temp utility?

  At the end of the day I guess all that is really important is that the processor runs stable (try using orthos or prime95)

  Cheers,

  M

• 21. Prospero  |  March 9th, 2008 at 2:57 am

  Hey guys,

  got my E8200 @ 3.81Ghz Stable (Orthos & repeated 3d2006 runs)

  Mem is @ 475 (2,1v and 1:1) and vcore is @ 1.216 so there is really a lot of headroom on this processor.

  Sadly, I have pc6400 Corsairs and they’re not too great, so I think i’ll switching over to some pc8500’s from Crucial. I hope to see a stable +4ghz oc after that

  Coretemps (using realtemp) are +/- 31c @ idle and +/- 58 @ full load (using load-line calibration, this might also explain the somewhat low vcore (but disabling didnt seem to make any difference to overall system stability))

• 22. Joseph Glover  |  March 9th, 2008 at 3:27 am

  Prospero nice work…

  I have a the same CPU and Crucial 8500 1066mhz RAM.. I’ve not been able to push mine further then stable 3.81ghz.

  Truth be told i’ve not pushed the FSB voltages or RAM voltages..

  My new rig is almost ready to start the OC fun.. just waiting on some MOSFET heatsinks.

  J.

• 23. theDuck  |  March 9th, 2008 at 10:40 am

  They’re good figures everyone - congrats!

  Yowsers Prospero!! Those kinds of speeds with PC6400? That’s FANTASTIC work..

  Joseph - if you achieved those speeds without overvolting, that’s a very excellent result. What model Zalman cooling are you using?

  I’d be really interested to know how you go testing for stability with ORTHOS (download it here) - it’s by far the most punishing CPU stress tester that I know of.

  I’d also love to know both your times to calculate pi to a million decimal places (Use a program called superPi - download it here). Do you crack 10 seconds? I think you’d both get damn close.

  Cheers,

  M

• 24. Prospero  |  March 9th, 2008 at 1:28 pm

  Hey duck,

  Thanx, I guess these Corsair sticks arent too bad afterall

  As per your request, my SuperPI time is 12,469s (im using SuperPi Mod, so I can also see the milliseconds). So not quite breaking the 10s barrier, but close indeed

• 25. theDuck  |  March 9th, 2008 at 1:39 pm

  That’s pretty good Prospero - I managed 12.062 secs at my 3.92GHz overclock - so we’re right on the money

• 26. Prospero  |  March 9th, 2008 at 1:43 pm

  Yup, I think I could push these sticks even further, but right now im still running the stock Heatsink… I think I’ll wait untill my Scythe Infinity gets in. Then we’ll see what these sticks can do with 2.2v

• 27. theDuck  |  March 9th, 2008 at 1:46 pm

  Out of interest for myself and other readers, what timings are you using on those corsair sticks Prospero?

• 28. Prospero  |  March 9th, 2008 at 1:55 pm

  The sticks run @ 5-5-5-18 right now. For some reason my motherboard wont run these sticks @ 4-4-4-12 whatever I do.. These sticks are supposed to be able to run @ 4-4-4-12 with the vdimm @ 2,1 but it wont even post at these settings.. But oddly enough they oc very well (FSB wise)

• 29. Prospero  |  March 9th, 2008 at 2:01 pm

  And ofcourse I couldnt wait, so I lowered the Multi to 6x and upped the FSB to 500 so effectively the corsairs are running @ 1Ghz now.. Im running Orthos over-night to see how they hold up. Maybe upping the vcore for the proc to 1,25/1,3 might just do the trick for a stable 4Ghz.

  (ps, Orthos is stable for 10 minutes now with the 1Ghz clock for the mem. I would’ve expected an error by now, but so far so good ;))

• 30. Prospero  |  March 9th, 2008 at 3:03 pm

  http://valid.x86-secret.com/show_oc.php?id=326685

  Orthos and 3dmark stable (for now :P) Will run stress-test for at least 12 hours now.

• 31. Greg K.  |  March 13th, 2008 at 12:54 pm

  Quick note on all of the comments from above (sorry that I’ve started this big confusion on the PCI issue):

  PCI speed is not related to your FSB(Front Side Bus) speed *AT ALL*.

  PCI speed: 33Mhz
  AGP speed: 66Mhz
  PCI-E speed: 100Mhz (technically 99Mhz)

  While PCI and AGP devices are both EXTREMELY sensitive to an overclock and are designed to function only at their given speed rating, they always account for power and speed fluctuations to some degree. Because of this, you rarely ever need to give any additional voltage to the PCI bus, nor do you need to up the frequency. However, having said that, there are some cases when it is advantageous, and one must understand what these changes do to decide if to use them. So here goes a brief Layman’s description:

  POWER: Some power supplies (most, actually) do not deliver the juice that they promise. So a 3.3V output could easily be 3.1 or 3.25 or 3.4 even. The error is even worse with higher voltages - 5V and 12V lines. Because the CPU needs very precise voltage regulation, the transformer there is usually much better, but even then: I have my mobo set to 1.5V on the CPU (I *REALLY* do not advice that ANYONE sets their voltage that high without testing equipment or knowing what they are doing). The actual Core voltage is read a 1.39-1.42. Good range for what I need and the fluctuation does not cause any crashes. actual 1.5 would have probably caused the temps to go up WAAAY higher.

  PCI SPEED: speed of data reads/writes does not affect the amount of data transfered at once, but it does increase the number of itterations per second (Hertz), so you effectively increase or decrease bandwidth and overall amount of data that is exchanged with the devices. In PC terms, that’s data bandwidth, but in reality it’s electrical signals. So too fast = not enough time for the PCI device circuitry to process the signals (at best) OR burn out and imminent death (at worst). You’ll never damage the PC, but a good sound card costs enough to regret the mistake, and a good video card costs enough to temporarily lose your sanity over the stupid thing. So to be safe, fix the PCI-E speed to 100Mhz and leave it. If you only have a video card on your PCI-E, and MAYBE a high-end sound card, you’re pretty safe with a 10% bandwidth overclock - set things to 110. If you use cheap stuff, it won’t get damaged but may refuse to work at a 10% overclock. Or it may give errors. Try 5% or go back to 100. Biggest thing to test is your USB and FireWire cards that are connected to your PCI. PCI-E will be fine - I promise. PCI speed is automatically calculated from PCI-E or AGP. It’s AGP/2 or PCI-E/3. Simple.

  OVERVOLTAGE: Assuming you know WHY you even need to and you’re careful enough to do it in small increments, you must have some tools to do this right. Cheapest and fastest way to test is with software and built-in sensors. Do not EVER trust the front panel sensors as they sit on top of your devices, and not inside, and all new CPUs have heatsinks built-in on top, so you’ll never get an accurate readout.

  SOFTWARE: You must at least have Everest and CPU-Z. Some variation of auto fan control software is nice too, but unnecessary if your fan noise is not bothering you at full speed at all times. I just set my PC to alert me LOUDLY if any of the fans fail. Beyond that, they run at full FPS at all times. Hey, fans are cheap.

  MEMORY: ALL fast memory is actually read by motherboards as 800Mhz. That’s not a mistake, that’s just the top of the JEDEC ratings. The rest is written in the manufacturer’s “best use” specification, called the EPP Profile. Some sticks contain no such profiles, some contain as many as 4. Your memory is DDR (double data rate), which means it always runs by default at twice your FSB. FSB = 333Mhz –> DDR2-667Mhz. When overclocking, for speed considerations (unless you know what you’re doing better than my simple description), try to pick a profile that will work at your FSBx2 (memory multiplier of 1) or FSBx4 (memory multiplier of 2). If you have 1024Mhz memory, that means that FSB can be 512Mhz at 1:1 or 256Mhz at 1:2. However, if your CPU’s multiplier is, say, 8 then you end up with 512Mhzx8 = 4.1Ghz. That’s probably more than what anyone without knowledge of OCing whould do with either of the E8×00 CPUs. Lower the multi to 7, and you get 3.58Ghz - too low to go through such trouble of OCing (goal is 3.95-4Ghz zone). FSB of 256Mhz is too low to bother also because you end up with a slower speed than your CPU is designed for. So this tells you one of two things: (a) we lower the FSB and not use the memory’s full potential, while still benefiting from it’s low latency at such high speeds) or (b) you picked the wrong memory for this CPU or (c) you need to set a different multiplier for both CPU and memory and you won’t get 1:1, maybe 5:6 will do, but you’ll utilize the best that both chips have to offer. I prefer to pick the (a) option: Set FSB to 500, CPU multi to 8, memory divider to 1:1, and you will have a 4Ghz E8×00 with 1Ghz CAS4-5 DDR2 memory. The whole setup should not only fly, but also be pretty stable, so long as each core’s temp is below 60C at full long-term load, and below 50C on daily average. junction temp (combined temp of both cores) is not always the best indicator, as cores can fail seperately, but if that’s all you’re looking at, just do your best to keep that below 55C at all times. Keep in mind that most EPP profiles require a significant voltage boost for high speeds. My motherboard flashes red in the bios when I set the memory voltage to 2.2V. It thinks I’m insane, since the standard voltage is 1.8V. It’s in the frying range for most memory sticks. But it’s what my Ballistix is asking for and it won’t work at less. A LOT of heat is put out by memory with such voltages, and memory rarely has a temp sensor. So just put a fan in the case somehow (even if you have to duck tape it or wire-twisty it to the case) to blow on your memory. small airflow across the sticks is infinitely better than no airflow. If you have an extra $20, but the tri-fan memory cooler from Corsair. The advice above ONLY applies to E8×00 chips with aftermarket coolers and GOOD mobo’s. Do not try this with a stock cooler, cheap mobo, and lower grade chips. Most CPUs won’t survive the voltage requirements they’ll need and the heat generated to run FSB 500 at any multiplier. E8×00 puts out less heat, and does higher FSB on less voltage (it’s stock FSB is 333, not 200 or 266 of previous chips).

  THERMAL GREASE: Arctic Silver 5 is really great, but as long as your paste is not graphite-based, just about anything else will do. We’re not building a rocket ship here, but attempting reasonable mild overclocks. At $1-2 difference - who cares. And a tube of this stuff lasts for numerous builds, so unless you already have something sitting in your parts closet, get AS5 and shush.

  APPLYING GREASE: biggest mistake here is “more is better”. Thermal paste has either ceramic or silver dust in its base. What this means is that a thin layer is highly conductive and clogs the air pockets that normally create between a CPU surface and heatsink surface. This does wonders for heat transfer. But if you put too much, the properties reverse, and you actually get an insulator, rather than a conductor, slow-roasting your precious E8×00. Do you like yours with a “port reduction”? (pun intended) Follow the rules for application of your specific paste to the tee, but normally the way you do it is a straight line through the center of your CPU, about 1/2mm in thickness and about 1/2 of your CPU’s length, and you set your heat sink on it and twist it both ways slightly, causing trapped air to get out. Do not plop a puddle in the middle of a CPU as you will usually put too much, and air traps underneath. Also whether ceramic or silver, the paste needs time to “set”. This does not mean “actual length of time”. It means “heat cycles”. Turn your computer on and put it under mild load for an hour, let the CPU get hot. Turn it off for an hour and let the paste cool down (CPU will get cold within milliseconds, but the heatsink above it will keep the paste warm for a while). Do this about 5-6 times. To an average person it means “turn the PC off at night, and do not stress the CPU to the max for the first week”. I noticed a 2C drop with paste versus without after immediately applying it, and I’ve noticed a total drop of 10C!!!!!! after about 5-6 heat cycles and letting silver dust align (”set”) in the AS5 paste.

• 32. theDuck  |  March 13th, 2008 at 1:13 pm

  Greg - the advice you give just keeps getting better. If you’d ever like to write a piece on utheguru about overclocking, let me know and I’ll give you a login.

  Thanks so much for all the effort!

  Cheers,

  M

• 33. Greg K.  |  March 13th, 2008 at 1:14 pm

  By the way… I ran the SuperPI on my PC. Specs: E4500@3.2Ghz, 4Gb DDR2-800Mhz 4-4-4-10 1:1 (usually I run 1000Mhz at 4-4-4-12 5:6, but today I felt like I’d try this, and it’s a little slower, but not much). Vista Ultimate x64. Gigabyte P35-DS3 mobo. 1M PI calculation took just shy of 20 seconds.

  So, anyone wondering the need to upgrade: the E8200 @ 3.95Ghz for only $96 more ($204-$198 on pricewatch), provides a 67% computational time advantage. Probably more, if OC’ed a little better. And it has an improved instruction set. AND IT STAYS COOLER!

  3.95/3.2 = 23% speed difference. The time should not be 67% faster, but it is. The faster bus means more data bandwidth, so the CPU can do more (less idle time) within the same time.

  (P.S.: E4600 OC’s to the exactly same point as E4500, since multi needs to be lowered for such high FSB anyway, so I am not looking at E4600 price for this compare)

• 34. Greg K.  |  March 13th, 2008 at 1:24 pm

  Thanks, theDuck, but I’d say if anyone wants to OC an E8200 rig, this post already offers a ton, and those who want to OCs everything, can probably find much better advice than mine in the same places where I started learning. Without naming names, there’s a ton of overclockers’ forums out there. What really drew me to this forum is interest in this specific chip, and now with everyone’s efforts, I think we’ve all decided just what it can do and how far it can OC within reason for everyday Joes. I humbly and gratefully thank everyone for this valuable input. I will be updating my rig with an E8400 soon and probably use the E4600 I have for my HTPC. It does just fine playing my HD media on my 73″ Mitsubishi TV @ 1080p. Even Team Fortress 2 runs at ridiculous fps (>120) at 1920×1200 with all settings at highest quality and AAx4 AFx16. But for my video post production, I really want the faster chip. It still takes hours to render video clips.

• 35. theDuck  |  March 13th, 2008 at 1:28 pm

  Greg - do you have any experience with the Q6600? Preferences? Is it speedier than the e8200? I know the clock speed is lower and you have to compare apples with apples and it’s a Quad core jobby - but imagine you’re running a single threaded app - do the other optimisations in the quad core architecture speed things up processing power wise?

  Cheers,

  M

• 36. Greg K.  |  March 13th, 2008 at 2:01 pm

  single thread in software development does not necesarily run on a single CPU. A single thread might still instantiate several classes whose actions can be computed in parallel. Example: Flash player is a single-thread app that runs a little faster on a dual core CPU. It’s the single *CRITICAL SECTION* that says “no other things can start in the cpu until I’m done” that only runs on a single core. However, the performance gain without specific multi-core optimization is minimal. If you worry about a single thread app that crunches numbers (read “uses memory”) then get the fastest data bandwith CPU you can (fastest FSB combined with lowest memory divider - 1:1 at highest memory speed). Latest AMD quad core beats Q6600, and even Intel offers better quads at comparable prices. So I’d think twice before I’d put a Q6600 in my system. But I cannot advise either way about specifics of Q6600 vs. E8200 for your specific tasks. You will probably win on FSB advantages alone with the E8200. But that’s my uneducated guess at best, not to be mistaken for advice at any angle.

• 37. Matt  |  March 19th, 2008 at 6:19 am

  Do you think that with a 500mHz FSB you could get it to 4gHz and stable using only an Arctic Cooling Freezer 7 Pro, with a well ventilated Antec P190 and keeping the CPU Voltage at 1.25

• 38. Nikos Chrisoulis  |  March 26th, 2008 at 10:08 am

  I did 4.0 and i didnt even have to worry about it. I set the system up today but i couldnt wait to overclock…..

  Didnt stress the machine, but i guess i can post as hiagh as 4.5 or something with the right settings.
  I appear quite sure that 4.0 can be very stable - great o/c!!!

  MoBo: Gigabyte P35 DS4 - this sh@t is ill!!!!!

• 39. HooT  |  March 29th, 2008 at 8:28 pm

  My stock E8200 runs at 49deg C.

  Is this normal?

• 40. ChrisH  |  March 30th, 2008 at 5:52 pm

  Bit of a beginner to the whole overclocking thing, but I’ve got my E8200 up to 3600mhz on air using Arctic Cooling Freezer 7 Pro and Crucial pc c6400 memory.

  Multiplier is 8 x
  FSB 450

  I realise if I had faster memory, I could probably imporve on these figures (so 2GB 8500 is my next purchase)

• 41. angelofdev  |  March 31st, 2008 at 2:16 am

  hey all iv just over clocked my e8200 to a stable 3.70ghz on an ASUS p5kc with PNY PC6400 2048mb sticks x2 with timings of 5-5-5-18

  iv got fsb at 463*8=3.7ghz
  my dram at 463*2= 926
  DRAM:FSB 1:1

  core voltage is 1.27v
  under prime95 torture test it reaches 59*C
  idle temp is 30*C on stock heatsink and fan.

  i got a 1M pi score of 12.683

  its my first over clock so im stoked about results altho im worried about temperatures

• 42. Blackiice  |  April 2nd, 2008 at 6:26 am

  i want to buy a E8200 Tomorow .
  I will use it on a P35-DS3 mainboard + 4gb of A-Data Vitesta Extreme 800Mhz Cl4
  I heard that are 2 revision of this cpu, the latest is C1 and fixed some of the bugs, including the temperature senzor .
  Hopefully i`l run it @ at 3600 or 3700 ,i`ll be back with details . U guys have C0 revision ?

• 43. Prospero  |  April 2nd, 2008 at 7:10 am

  Hey Blackiice.. Yea, I run a C0, but luckily I havent got the stuck sensor problem. 3,6Ghz shouldnt pose any problems, I think ure mem-sticks should do 450 easily, maybe even @ cl4

  Good luck and let us know how you did ?

• 44. angelofdev  |  April 2nd, 2008 at 11:48 am

  how do you know if you have the problematic C0 revision of the E8200?

  cheers
  -angel

• 45. Prospero  |  April 2nd, 2008 at 11:51 am

  Use CPU-Z (cpuid) to identify your revision.

  http://www.cpuid.com

• 46. angelofdev  |  April 2nd, 2008 at 12:06 pm

  yeah I’ve already got cpuZ and I’ve got the C0 revision but that doesn’t tell me if I have the C0 revision with the problem.
  Or doesn’t the stuck sensor tell the revision number in cpuZ?
  sorry im a bit lost.

• 47. Starbuck  |  April 3rd, 2008 at 8:56 am

  What software are you guys using to measure you’re CPU temps?

  As I’m using core temp and when using prime 95 it says that the cpu temp is 70 degree, and it Crysis it peaks at 6065 degrees??

  By the way at the moment my E8200 is running at:
  3.2GHz, 400×8 at 1.25v
  Ratio is 1:1 with RAM at 4-4-4-12 and 2.2v

  Have I done something wrong?

• 48. Starbuck  |  April 3rd, 2008 at 8:58 am

  Sorry its meant to read 60 to 65 degrees when playing Crysis

• 49. Prospero  |  April 3rd, 2008 at 8:59 am

  You have 2 possibilities:

  1> U have a proc with a faulty temp-sensor. In that case you could use RealTemps to see if your tempsensor is stuck.

  2> Your heatsink is curved, meaning it doesnt cover the whole surface of your processor. Only solution is to buy a new heatsink. Or the Heatsink isnt properly seated, in that case reseating it should do the trick.

• 50. theDuck  |  April 3rd, 2008 at 9:22 am

  Prospero - I’ve been monitoring this thread but not actively participating because I’m writing a PhD proposal :-D.. many thanks for your great answers.

  M

• 51. Starbuck  |  April 3rd, 2008 at 11:02 am

  Prospero, I’ve already tried reseating the Heatsink, and it had made some difference with temps. But what do you mean by curved? I don’t quite get what you’re getting at.

• 52. Prospero  |  April 3rd, 2008 at 11:08 am

  Its possible that during the production process of the heatsink, the bottom surface (which “touches” the heatspreader of the processor) isnt made completely smooth and flat. If thats the case, then the heatspreader on the processor isnt able to transfer heat from the surface of the heatspreader to the surface of the heatsink. Even with the use of thermal-paste this would lead to insulation which is bad. Cause we want conduction and not insulation. Im not saying that this is the case with your heatsink, but Ive heard many stories of people who changed their heatsinks and had amazing gains in cooling-capacity.

• 53. Starbuck  |  April 3rd, 2008 at 11:32 am

  Well, I’ve just e-mailed intel about getting a replacement heatsink(it cant hurt to try another intel HSF can it!). Otherwise I’ll just get a Thermalright ultra 120 as I’ve heard lots of good things about them.

  So I guess I’ll just have to see. Thanks for the help

  Oh, what about lapping the just Heatsink would that make a difference?

• 54. Prospero  |  April 3rd, 2008 at 11:37 am

  The iNtel heatsink is actually pretty nice if you have a good one. So emailing them is a good idea, probably falls under warranty. Ive heard alot of good thing about “Scythe Ninja’s” but those Thermalright are pretty decent aswell.

  Lapping the heatsink is always a good idea, since this smoothens the surface which is good for heat dissipation.

  And you’re welcome

• 55. Starbuck  |  April 3rd, 2008 at 11:46 am

  I’ll be sure get hold of various grains of sand paper and give it a try, what kind of temperature changes should I expect?

• 56. theDuck  |  April 3rd, 2008 at 12:00 pm

  Starbuck - with respect to Prospero, I’m tending to be a little sceptical that your heatsink is warped.. others may differ but I think it’s pretty rare to have such a major warp in a heatsink that it would cause these kind of issues and I think Intel will give you a negative response.

  I think it would be wise (so long as you have excluded the possibility that the temperature reading is fallacious) to try cleaning your heatsink as described above and apply new compound - arctic silver is my favorite. If you go to their website they’ll give you very good instructions about just how to apply the compound properly.

  Other causes may be:-

  1. You haven’t clipped the heatsink in properly on all four corners (very common).
  2. You’ve used too much heatsink compound.

  Cheers,

  M

• 57. doddie  |  April 6th, 2008 at 1:58 am

  Hi,

  I have since overclocked my wolfdale to 3.4 Ghz (see previous post) using a new MB, I RMA the last board as there was a bit of play in 1 of the memory dimms and I was getting reboots.

  Specs:
  Gigabyte P35 DS3 Rev2 Bios F13a
  Wolfdale E8200 @ 3.4Ghz stock cooler
  2x 1GB Crucial Ball tracer (8500) dual running @ 4-4-4-12

  I think I have all the settings correct, (see screenshot below )
  Im new to this overclocking stuff
  The system has been running rock stable, there’s a few points i’d like to mention.

  1.My wolfdale must have the stuck temp sensor, as both core temps, never goes below 68c. (see previous post) so im ignoring these.

  2.The core voltage seems to drop when under load
  1.22v > 1.18v is this normal?
  also overall it seems to be below average users voltage 1.24 etc…
  best to leave this alone?

  3. Is CPU temp 54c a bit high, if it’s correct? tho other software shows the same readings.

  4 The memory’s voltage seems below the crucial spec of 2.20,
  tho the system is stable. should i change this?

  4. if theres anything I could tweak/change plz let me know.

  Cheers
  Dodds

  http://img357.imageshack.us/my.php?image=mixe820034ghzloadle1.jpg

• 58. Gooralski  |  April 7th, 2008 at 6:13 pm

  I have a small question. My rig is almost the same as theDuck’s rig, except for memory which is Patriot LLK 6400 (4-4-4-12 @ 800 MHz, 2,2V). MoBo seems to be quite good for overclocking, but only one setup is running fine for me - FSB @ 400MHz, mem @ 960 MHz 5-5-5-15 @ 2.2V. When i tried to set FSB to 425 MHz and FSB:DRAM to 1:1 (850 MHz should be ok after stable work @ 960 MHz) MoBo starts twice and set back CPU to 2.66 GHz. Not sure what’s wrong, i’ve tried a lot of sets (including VCORE ~ 1.3V) and nothing. Any ideas? Thanks in advice.

• 59. Jacob  |  April 8th, 2008 at 7:00 am

  Hi Doc,

  Thanks for the information here. Long story short: Need to overclock my new E8200 to say 3.5 or to whatever speed is more stable. I hope you or someone here can help; this is my setup:

  GIGABYTE GA-EP35C-DS3R
  ARCTIC COOLING Freezer 7 Pro
  4×1 G.SKILL DDR2-800 (F2-6400PHUT-2GBHZ)
  EVGA 9800GTX (G92)
  CORSAIR-650 WidnowsV

• 60. ed  |  April 9th, 2008 at 11:15 am

  I’am deciding e8200 or e8400 in search of higher overclock mhz and still not too hot system.Any help

• 61. Mika75  |  April 11th, 2008 at 11:56 am

  Hey Doc, Office 200 keygen? hahah ……nice overclock mate !

• 62. GK  |  April 17th, 2008 at 9:34 pm

  Hey guys,
  just bought an:
  E8200
  Asus P5n-E SLI
  G.Skill DDR2800 PC-6400 (CL4-4-4-12)

  I thought I might OC to 3~3.5GHz (maybe more) and was wondering if I could get some pointers.

  I’ve read a bit and you guys really know what your talking about.
  This will be my first OC and I wanna make sure I get it right!!

  Any comments appreciated…

• 63. angelofdev  |  April 17th, 2008 at 11:54 pm

  you can easily get it overclocked to a stable 3.7ghz

  8 x 463= 3704= 3.7ghz
  DRAM 925mhz or 930mhz

  i find it that 1:1 FSB:DRAM ratio is the best for a stable system

• 64. Joseph Glover  |  April 18th, 2008 at 12:03 am

  Set your cpu core to 1.25 when your hitting the 3.6+ mark.

  I dropped my 8200 into my Lanparty gig and its really capped my OC. Went from 3.7ghz on the GA-P35-DQ6 to only 3.0ghz on my GA-G33M-DS3R…

  I think its all in the voltages… any ideas?

  cheers guys.

  J.

• 65. Joseph Glover  |  April 18th, 2008 at 12:21 am

  TheDuck, Sorry mate i hadnt read your post about wanting me to show my SuperPi and ORTHOS scores… I’ve downloaded the Apps and i’ll have a play tonight… I’ve moved my CPU into a DS2R which has capped my OC like a bitch.. now why i hear you ask would i go from the DQ6 to the DS2R?? i had to use a M-ATX mobo in my HackPro Mod.

  check it out.

  http://a363.ac-images.myspacecdn.com/images01/106/l_b0f564f224c8dc8b8e7a0cc6b0e35c8a.jpg

  My specs and cooling are as follows:

  Intel Core 2 Duo E8200 @ 3.0 Ghz (Zalman CNPS9700-LED)
  Gigabyte GA-G33M-DS2R (*North Bridge* Noctua NC-U6 Dual Heatpipe, *South Bridge* Thermalright HR-05 IFX)
  4GB Crucial Ballistix Tracer 8500C5 (Antec SpotCool Fan)
  Nvidia BFG 9800GTX OC 512MB
  Mac Pro Case ( 4 x Noctua NF-P12 120mm)

  check out the mod build:

  http://aquamac.proboards106.com/index.cgi?board=hack1&action=display&thread=460&page=7

  I’d really like to push my CPU back upto a stable 3.66!

  lets make it happen.

  Joe.

• 66. Kijaszek  |  April 18th, 2008 at 5:01 pm

  http://zuo.kijaszek.net/images/kijaszek_oc.jpg

  4GHz!! This is brillinat CPU

• 67. Phil  |  April 20th, 2008 at 6:53 pm

  @Starbuck

  Mate, the current version of CoreTemp assumes a value incorrectly that makes it report all temps for the e8×00 series cpus incorrectly…

  What you want is realtemp, which as I understand it estimates said value a little closer to the mark, and thus reports the temperatures more accurately.

  http://www.xtremesystems.org/forums/showthread.php?p=2809778

• 68. Cogman  |  April 27th, 2008 at 12:27 am

  Hey, just wanted to say a big thank you for your article. I was desperate for a new processor (as Visual Studio compiling time is driving me crazy!) and I was looking at either the e8200 or the q9300, but after reading about how easily the e8200 overclocked I went straight out and bought one.

  I got 3.6Ghz from 450×8 with stock voltages. It was so easy and heat is at 40/60. Everything seems stable no matter what I throw at it, Orthos or Prime95. I’m running super pi in 12.9 seconds, so VERY impressed, beating my old e4300 by 20 seconds. I had so much trouble overclocking my old e4300 at all, so this thing has made my day.

  Thanks again

• 69. theDuck  |  April 27th, 2008 at 7:45 am

  Thanks Mark - small world - I came across your site too a couple months back.

  Cheers,

  M

• 70. mix  |  May 1st, 2008 at 5:08 am

  I did a stress test with Orthos on my stock 8200 (stock cooling) and got the following temps: both cores stay at 39C idle, after about a minute of stressing one core sits at 52C and the other slowly starts to climb to about 45-46C.

  Is this normal to have such a difference in temperature between the cores?Is the second cores temp sensor faulty?Would overclocking to 3.2ghz (400fsb) using 2×1GB Kingmax pc6400 ram be a good idea using the box intel cooler (my mobo is a Gigabyte EP35-DS3R) ?

• 71. theDuck  |  May 1st, 2008 at 6:35 pm

  @Mix - I’d be suggesting you check out arcticsilver’s website and have a look at their instructions for that core - they recommend laying two lines of heatsink, as the two cores are seperated by around 7mm on the die (according to the cut-out view on that website)

  So.. my most likely interpretation of your ‘issues’ (although both temps are quite reasonable) would be that one core isn’t getting as effective contact with the heatsink as the other.

  As for overclocking with the stock heatsink - 3.2 should be absolutely well and truly achievable. With the stock heatsink I obtained stable O/C of 3.6+.. your temperatures seem reasonable so go for it. As for using PC6400, depending upon the quality of the RAM you may find that the capabilities of your RAM might restrict your O/C to around that 3.2 level - if you find that your motherboard has a particularly long reboot / post after your overclock and clears all your settings back to default the most likely culprit is the RAM.

  Cheers,

  Matt

• 72. absy  |  May 2nd, 2008 at 6:58 pm

  E8200 @ 3,4Ghz 1,2V bios, 1,18v in charge (watch in OCCT, hwmonitor etc.). My crucial pc8500 is at stock 1066 5-5-5-15, but for stability, I need to push the fsb and mch voltage +0,05V with the ga-p35-dds4.
  For those that it can help.

  temp. core 63° full, ~50° idle.

• 73. mix  |  May 4th, 2008 at 12:40 am

  I changed the thermal compound with the AS5 but the temps stayed the same (39 idle on both cores, 55 one and 46-47 the other one in full load).

  So now I’m running the E8200 at 3,2Ghz, stock voltage, the memory at 800Mhz, stock timings and stock voltage (Kingmax pc6400) and the temps are 39 idle and 58-59 one core, 51-52 the second core in full load. Running Orthos for 30 minutes now and all is well. Thanks for the help

  By the way which program do you use to monitor the temps? I use Real Temp and I hope it’s right cuz Core Temp is giving me about +10deg above real temp. Also Easy Tune 5 says 52deg CPU Temp in full load. Is that a good reading or should I stick to real temp?

• 74. OCnewbie  |  May 7th, 2008 at 2:30 am

  hi nice article! im thinking of building a pc, buying parts as i can afford them, the basic spec will be:

  Gigabyte GA-N650SLI-DS4L
  Intel Core 2 Duo E8200
  Kingston 2GB DDR2 PC2-6400

  and because i am cheap (uk prices are really expensive) i would like to beable to overclock the E8200 even if it is just up to 3.2Ghz i will be happy 3.6Ghz would be amazing though.

  if i was to overclock the cup to either of these speeds what bios settings would you recomend? keep in mind that i am new to this so if you could keep it as simple as possible the steps i would have to go through i would be very greatful.

  p.s i know i would need a better cpu cooler so im thinking of getting a Arctic Cooling Freezer 7. i know its only a small overclock but it will help with overall system temp aswell.

• 75. theDuck  |  May 7th, 2008 at 3:22 pm

  OCNewbie - if I were you I’d ditch the idea of buying a cpu cooler and buy better RAM instead - as mentioned above, PC8500. That way you won’t have to muck around with overclocking the RAM too much (or at all), which can be very tricky for beginners as these gigabyte motherboards tend to inexplicably reset all your changes if the RAM is going to0 quickly.

  All your going to need to really do is drop the RAM:FSB ratio to around 1:1 or 1.25 to 1, slightly up the CPU voltage (as mentioned in my article, above) and up the CPU clock to whatever it will handle - 400 will get you 3.2GHz and you will find that the stock CPU cooler will handle that no problems whatsoever.

  With PC6400, and a RAM:FSB ratio of 1:1 you’ll get to 3.2GHz without any problems. With PC8400 and a ratio of 1:1 you’ll get in excess of 3.6GHz without any problems, with the stock cooler.

  Unfortunately I can’t really offer any more specific advice as motherboards differ.

  M

• 76. OCnewbie  |  May 8th, 2008 at 2:31 am

  ok. thank you for the help, ill see if i can get better ram even though im still happy with 3.2GHz it would be nice to have better ram incase i want to go higher than that in the future.

  right so i have looked around a bit and i have found this

  http://www.dabs.com/productview.aspx?Quicklinx=4KZK&SearchType=1&SearchTerms=OCZ+Technology+2×1GB+240DIMM+PC2+8500+Reaper&PageMode=3&SearchKey=All&SearchMode=All&NavigationKey=0

  its only a little more expensive than what i was going to use before anyway.

  what do you think?

Leave a Comment

Some HTML allowed:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>

Trackback this post  |  Subscribe to the comments via RSS Feed